Daily Research Digest
arXiv Papers
2026-08-13
332
Papers
9
Categories
330
Translated
收藏清单 0
精选 · Favorites
75
cs.AI / 1 / 2608.11343
Can Frontier LLMs Match Natively Multimodal Embeddings? A Comparison on Hard-Negative Text-to-Image Retrieval
前沿大语言模型能否媲美原生多模态嵌入？一项关于困难负样本文本到图像检索的比较
large language model
大语言模型相关
Abstract
Multimodal retrieval and classification across different types of media, spanning text, images,video and audio, has traditionally relied on dual-encoder models that align visual and textual representations through contrastive learning. The March 2026 release of Gemini Embedding 2, Google's first natively multimodal embedding model to map text, images, video, audio, and documents into a single shared space, raises competition among multimodal retrieval systems. Simultaneously, frontier Large language models (LLMs) have also demonstrated strong visual understanding, raising the question of whether they can serve as effective zero-shot rankers. Our study provides the first direct comparison of native multimodal embeddings against LLM-based visual ranking on Flickr30k. We observe that GPT-4.1 and Claude Sonnet 4.6 perform on par with Gemini Embedding 2. Additionally, once embeddings are precomputed, multimodal embeddings are better suited for low-latency applications.
Chinese Translation
跨文本、图像、视频和音频等不同媒体类型的多模态检索与分类，传统上依赖双编码器模型，通过对比学习对齐视觉与文本表示。2026年3月发布的Gemini Embedding 2是谷歌首个将文本、图像、视频、音频和文档映射到单一共享空间的原生多模态嵌入模型，它的发布加剧了多模态检索系统之间的竞争。与此同时，前沿大语言模型（LLMs）也展现出强大的视觉理解能力，这引发了一个问题：它们能否作为有效的零样本排序器。我们的研究在Flickr30k上首次对原生多模态嵌入与基于LLM的视觉排序进行了直接比较。我们观察到，GPT-4.1和Claude Sonnet 4.6的表现与Gemini Embedding 2相当。此外，一旦嵌入被预先计算，多模态嵌入更适合低延迟应用。
cs.AI / 2 / 2608.11493
From Prompting to Behavioral Alignment: Personalized LLM Judges for Recommendation Evaluation
从提示到行为对齐：用于推荐评估的个性化大语言模型评判器
large language model
大语言模型相关
Abstract
Traditional offline recommendation evaluation relies heavily on complex, manually maintained feature pipelines that are difficult to scale. While Large Language Models (LLMs) offer a promising alternative by predicting user engagement directly from raw text logs, empirical analysis in this study identifies a critical failure mode termed bidirectional rationalization. In a zero-shot setting, LLMs are found to convincingly argue for both positive and negative user engagement outcomes on the exact same item with identical evidence, highlighting the unreliability of off-the-shelf LLMs in predicting user engagement. To resolve this, we develop and apply a sequential behavioral alignment framework pairing fine-tuning with preference optimization over paired correct and counterfactual rationales. Evaluated on real-world homepage interaction logs, this aligned reasoning approach achieves a 32.19\% lift in Macro-F1 score over the zero-shot baseline and matches the production feature-engineered baseline. The results demonstrate that behavioral alignment mitigates bidirectional rationalization while delivering human-interpretable reasoning traces without manual pipeline overhead.
Chinese Translation
传统的离线推荐评估在很大程度上依赖于复杂且需人工维护的特征流水线，而这些流水线难以扩展。尽管大语言模型（LLMs）通过直接从原始文本日志预测用户参与度提供了一种有前景的替代方案，但本研究的实证分析识别出了一种被称为双向合理化的关键失效模式。在零样本设定下，研究发现LLMs能够在完全相同的物品和相同证据下，令人信服地为积极和消极的用户参与结果同时进行论证，这凸显了现成LLMs在预测用户参与度方面的不可靠性。为了解决这一问题，我们开发并应用了一个序列化行为对齐框架，该框架将微调与基于成对正确理由和反事实理由的偏好优化相结合。在真实世界首页交互日志上的评估表明，这种对齐推理方法相比零样本基线在Macro-F1分数上实现了32.19%的提升，并达到了与生产环境中基于特征工程的基线相当的水平。结果表明，行为对齐缓解了双向合理化，同时在不带来人工流水线开销的情况下提供了人类可解释的推理轨迹。
cs.AI / 3 / 2608.11583
Localizing Safety Alignment: MLP Layers and Mid-Network Blocks Encode Refusal Behavior in Large Language Models
定位安全对齐：MLP 层与网络中段模块编码大型语言模型中的拒绝行为
large language model
大语言模型相关
Abstract
Safety alignment in large language models is often treated as a distributed property of the entire network, yet its practical brittleness suggests that refusal behavior may be concentrated in a smaller set of parameters. This work addresses where safety-aligned refusal is encoded by transplanting weights from aligned models into matched unaligned base models at multiple levels of granularity. Using two open-weight model pairs and four safety benchmarks, we conducted experiments to compare the effects of replacing attention weights, MLP weights, contiguous layer regions, and MLP blocks. Across both model families, refusal transfer is dominated by MLP weights: replacing MLP parameters recovers substantially more malicious-prompt refusal than replacing attention parameters, with gains of at least 2.7 times more across benchmarks. Within the MLP stack, refusal-relevant parameters exhibit a consistent mid-network concentration, as the block spanning layers 8-11 is selected first in all six greedy searches over model-dataset pairs. The results also show that the composition of safety-relevant components is non-additive: in five of six greedy trajectories, adding more aligned blocks can reduce refusal performance, and selective block subsets can outperform full MLP transplantation on malicious refusal, benign over-refusal, or both. Finally, greedy orders transferred to OR-Bench vary with the source benchmark used to derive them, indicating a benchmark-dependent precision-coverage trade-off. These results suggest that safety alignment in current LLMs is both localized and interaction-sensitive, offering insight into alignment brittleness and potential avenues for targeted safety interventions.
Chinese Translation
大型语言模型中的安全对齐通常被视为整个网络的一种分布式属性，然而其实际脆弱性表明拒绝行为可能集中在较小的参数集合中。本研究通过将已对齐模型中的权重在多个粒度层级移植到匹配的未对齐基础模型中，来探究安全对齐拒绝行为被编码在何处。我们使用两对开放权重模型和四个安全基准进行了实验，以比较替换注意力权重、MLP 权重、连续层区域和 MLP 块的效果。在两个模型家族中，拒绝迁移均由 MLP 权重主导：替换 MLP 参数所恢复的恶意提示拒绝能力显著多于替换注意力参数，其提升在各个基准上至少达到 2.7 倍。在 MLP 堆栈内部，与拒绝相关的参数表现出持续的中段网络集中性，因为在所有六次针对模型-数据集对的贪心搜索中，跨越第 8 至 11 层的块均首先被选中。结果还表明，安全相关组件的组合是非加性的：在六条贪心轨迹中的五条中，添加更多已对齐块可能会降低拒绝性能，并且选择性块子集在恶意拒绝、良性过度拒绝或两者上可以优于完整 MLP 移植。最后，迁移到 OR-Bench 的贪心顺序会随用于得出它们的源基准而变化，这表明存在依赖于基准的精确率-覆盖率权衡。这些结果表明，当前 LLM 中的安全对齐既具有局部性，又对交互敏感，这为理解对齐脆弱性以及针对性安全干预的潜在路径提供了洞见。
cs.AI / 4 / 2608.11604
Learning from Online User Feedback for Shopping Agents
从在线用户反馈中学习购物代理
large language model
大语言模型相关
Abstract
Large language model-based shopping agents are increasingly deployed in real-world e-commerce platforms, generating massive amounts of user interaction logs that provide valuable supervision for improving these agents. However, existing approaches primarily rely on offline training signals, such as user-item interactions or synthetic preference data, while largely overlooking the rich supervision contained in users' natural conversational feedback. Moreover, the available online feedback is heterogeneous, sparse, and noisy, making it difficult to transform into reliable learning signals automatically. To address these challenges, we propose LOFA, a framework that enables shopping agents to learn directly from real online interaction logs without human annotation. LOFA combines reinforcement learning over verifiable purchase outcomes with feedback-aware on-policy distillation, which identifies users'in-dialogue directives and converts them into dense token-level supervision. These complementary objectives capture both collaborative behavioral patterns and user-specific preferences. Extensive experiments on real-world e-commerce logs demonstrate that LOFA consistently improves recommendation quality, response helpfulness, and user-satisfaction alignment over strong baselines, highlighting the effectiveness of learning shopping agents from real online user feedback.
Chinese Translation
基于大语言模型的购物代理正越来越多地部署在现实世界的电子商务平台中，产生大量用户交互日志，为改进这些代理提供了有价值的监督信号。然而，现有方法主要依赖离线训练信号，例如用户-商品交互或合成的偏好数据，而在很大程度上忽略了用户自然对话反馈中所包含的丰富监督信息。此外，可用的在线反馈是异质的、稀疏的且含有噪声，使得难以自动将其转化为可靠的学习信号。为了解决这些挑战，我们提出了 LOFA，一个使购物代理能够直接从真实在线交互日志中学习而无需人工标注的框架。LOFA 将基于可验证购买结果的强化学习与反馈感知的策略内蒸馏相结合，后者识别用户在对话中的指令并将其转化为密集的 token 级监督。这些互补的目标同时捕捉协作行为模式和用户特定偏好。在真实世界电子商务日志上的大量实验表明，LOFA 相较于强基线在推荐质量、回复有用性和用户满意度对齐方面均有持续提升，突显了从真实在线用户反馈中学习购物代理的有效性。
cs.AI / 5 / 2608.11616
MBA: Multimodal Benchmark and Agents for Real-World Business Ideation
MBA：面向真实世界商业创意构思的多模态基准与智能体
large language model
大语言模型相关
Abstract
Agentic systems powered by large language models (LLMs) have opened new opportunities for business ideation. Yet existing approaches remain confined to a text-only paradigm, despite the inherently multimodal nature of real-world contexts. We thus introduce MBA-Bench, the first multimodal benchmark for training and evaluating business ideation agents, comprising 30K samples across six domains, each domain characterized by distinct visual cues not fully conveyed by text alone. Concretely, we automatically caption images and employ GPT-4o to generate five reference ideas for each of three business questions through retrieval query generation, market evidence retrieval, and evidence-augmented synthesis. Following prior work, we evaluate agents across six business-oriented criteria using MLLM-as-a-Judge. To consider settings where criteria are hidden or disclosed, we present MBA-b and MBA-k for blind and known, respectively. We train both with two novel reward objectives---creativity and feasibility---while MBA-k further optimizes the six disclosed criteria for eight in total. Both are trained via LoRA-based supervised fine-tuning followed by group relative policy optimization with these setting-specific rewards. For extensive experiments on MBA-Bench, we set up two baselines accommodating either captions only or multimodal inputs, with the latter nearing closed-source performance on several metrics. MBA-b and MBA-k outperform caption baselines by 63.9% and 77.1%, and multimodal baselines by 25.6% and 35.8%, respectively.
Chinese Translation
由大语言模型（LLMs）驱动的智能体系统为商业创意构思开辟了新的机遇。然而，尽管现实世界情境本质上具有多模态特性，现有方法仍局限于纯文本范式。因此，我们提出了 MBA-Bench，这是首个用于训练和评估商业创意构思智能体的多模态基准，包含覆盖六个领域的 30K 个样本，每个领域都具有仅靠文本无法完全传达的独特视觉线索。具体而言，我们自动为图像生成描述，并利用 GPT-4o 通过检索查询生成、市场证据检索和证据增强合成，为三个商业问题中的每一个生成五个参考创意。沿用先前工作，我们使用 MLLM-as-a-Judge 在六个面向商业的标准上评估智能体。为了考虑标准隐藏或公开的设置，我们分别提出了用于盲评的 MBA-b 和用于已知标准的 MBA-k。我们使用两个新的奖励目标——创造性和可行性——来训练这两者，而 MBA-k 进一步优化六个公开标准，共计八个目标。两者均通过基于 LoRA 的监督微调进行训练，随后使用这些特定设置奖励进行分组相对策略优化。为了在 MBA-Bench 上进行广泛实验，我们设置了两个基线，分别仅适应描述文本或适应多模态输入，其中后者在若干指标上接近闭源性能。MBA-b 和 MBA-k 分别比描述文本基线高出 63.9% 和 77.1%，比多模态基线高出 25.6% 和 35.8%。
cs.AI / 6 / 2608.11631
CLAIM: Leading Open-domain Active Clarification of Large Language Models with Uncertainty Measurement
CLAIM：以不确定性度量引领大语言模型的开放域主动澄清
large language model
大语言模型相关
Abstract
In open-domain human-computer interaction scenarios, large language models (LLMs) frequently encounter user queries that are ambiguous or incomplete. In such cases, directly producing an answer often leads to overgeneralized, erroneous, or low-information responses. In contrast, asking clarifying questions can substantially improve interaction quality. However, existing approaches still rely heavily on manually annotated data or preference alignment to address two fundamental challenges: when clarification is necessary, and which aspect of the query should be clarified. This reliance incurs high annotation costs and limits generalization. To address these challenges, we propose CLAIM, an uncertainty-driven framework for active clarification learning in open-domain settings. CLAIM eliminates the need for explicit human preference annotations by quantifying query uncertainty through the entropy induced by answer disagreements across multiple models. This uncertainty signal is then used to construct high-quality synthetic data, enabling the training of a unified clarification decision model through a combination of supervised learning and reinforcement learning. Specifically, we propose an entropy-driven synthetic data generation pipeline that integrates entropy-based uncertainty estimation with semantic clustering and reasoning-based judgments, enabling reliable automatic annotation of clarification requirements. To train CLAIM, we formulate the clarification process as a structured decision generation problem and adopt a training paradigm that combines supervised fine-tuning (SFT) with group-relative policy optimization (GRPO). Experimental results demonstrate that CLAIM can learn stable and generalizable clarification strategies without relying on manually labeled data, offering a low-cost and robust solution for proactive understanding in real-world open-domain interactions with LLMs.
Chinese Translation
在开放域人机交互场景中，大语言模型（LLMs）经常遇到模糊或不完整的用户查询。在这种情况下，直接生成答案往往会导致过度概括、错误或信息量低的响应。相比之下，提出澄清性问题可以显著提升交互质量。然而，现有方法仍然严重依赖人工标注数据或偏好对齐来解决两个基本挑战：何时需要澄清，以及应澄清查询的哪个方面。这种依赖带来了高昂的标注成本，并限制了泛化能力。为应对这些挑战，我们提出了 CLAIM，一个在开放域设置中由不确定性驱动的主动澄清学习框架。CLAIM 通过多个模型之间答案分歧所产生的熵来量化查询不确定性，从而消除了对显式人类偏好标注的需求。该不确定性信号随后被用于构建高质量的合成数据，从而能够通过监督学习与强化学习的结合来训练统一的澄清决策模型。具体而言，我们提出了一种熵驱动的合成数据生成流程，该流程将基于熵的不确定性估计与语义聚类和基于推理的判断相结合，从而能够对澄清需求进行可靠的自动标注。为训练 CLAIM，我们将澄清过程形式化为一个结构化决策生成问题，并采用将监督微调（SFT）与分组相对策略优化（GRPO）相结合的训练范式。实验结果表明，CLAIM 能够在不依赖人工标注数据的情况下学习到稳定且可泛化的澄清策略，为真实世界开放域 LLM 交互中的主动理解提供了一种低成本且鲁棒的解决方案。
cs.AI / 7 / 2608.11679
AgenticTwin: An Agentic LLM Framework Integrated with Digital Twin for Anomaly Detection
AgenticTwin：一种与数字孪生集成的智能体大语言模型框架用于异常检测
large language model
大语言模型相关
Abstract
Digital twins are increasingly used to monitor and simulate the behavior of cyber-physical systems. Even with skilled operators, interpreting anomalies detected within digital twin pipelines is challenging, as the sheer complexity and volume of raw sensor data make thorough analysis difficult. Recent advances in large language models (LLMs) offer promising capabilities for reasoning and explanation, yet their integration into digital twin-driven anomaly analysis remains underexplored. In this work, we propose AgenticTwin, an agentic framework that integrates LLM-driven reasoning with a digital twin-based anomaly detection pipeline. The framework grounds LLM-generated explanations in outputs from a digital twin-driven anomaly classifier and enables human operators to ask relevant natural-language questions about the system. Beyond the framework itself, we introduce a benchmark-oriented evaluation pipeline constructed over synthetic anomalies injected into a real-world weather sensor dataset, enabling controlled generation of operator queries over anomaly events. We further evaluate the feasibility of deploying lightweight, open-source LLMs for practical cyber-physical environments. Experimental results demonstrate that structured agent collaboration and knowledge-grounded reasoning improve diagnosis quality, contextual retrieval, and mitigation quality across diverse possible anomaly scenarios.
Chinese Translation
数字孪生越来越多地被用于监控和模拟信息物理系统的行为。即使有熟练的操作员，解释数字孪生流水线中检测到的异常也颇具挑战性，因为原始传感器数据的高度复杂性和庞大数量使得彻底分析变得困难。大语言模型（LLMs）的最新进展为推理和解释提供了有前景的能力，但它们与数字孪生驱动的异常分析的整合仍未得到充分探索。在这项工作中，我们提出了 AgenticTwin，这是一个智能体框架，它将 LLM 驱动的推理与基于数字孪生的异常检测流水线集成在一起。该框架将 LLM 生成的解释建立在数字孪生驱动的异常分类器的输出之上，并使人类操作员能够就系统提出相关的自然语言问题。除了框架本身，我们还引入了一个面向基准的评估流水线，该流水线构建于注入到真实世界天气传感器数据集中的合成异常之上，从而能够针对异常事件受控地生成操作员查询。我们进一步评估了在实际信息物理环境中部署轻量级开源 LLM 的可行性。实验结果表明，结构化的智能体协作和基于知识的推理提高了在多种可能异常场景下的诊断质量、上下文检索质量和缓解质量。
cs.AI / 8 / 2608.11705
Making Your LLMs More Objective: Stabilizing LLM Safety Behavior Across Traits with Trait-Invariant Safety Tuning
让你的大语言模型更加客观：通过特质不变安全调优稳定大语言模型跨特质的安全行为
large language model
大语言模型相关
Abstract
Aligned large language models (LLMs) are expected to exhibit safety behavior based on the content of the user request: they should refuse unsafe requests and comply with safe ones. However, we show that the same request can elicit substantially different safety decisions under different traits assigned in the system prompt, a failure mode we call trait-induced safety variation. To measure this failure, we introduce refusal-based metrics: Trait-Induced Deviation measures dataset-level deviation from the no-trait baseline, while Trait-Induced Flip Rate measures whether the same request receives different safety decisions across traits. We then provide a representation-level analysis of the mechanism behind trait-induced safety shifts and find that traits perturb the model's safety representations within a low-dimensional subspace. To achieve trait-invariant safety, where safety behavior remains stable across traits, we introduce Trait-Invariant Safety Tuning (TIST), a simple yet effective self-distillation framework that aligns an LLM's trait-conditioned behavior with its no-trait behavior. Guided by our analysis, we further propose Trait-Subspace Neutralization (TraSN), an instantiation of TIST, which enforces invariance only within the identified trait subspace. Experiments show that TraSN improves trait-invariant safety and strengthens harmful-request safety while preserving general capability. Our results highlight traits as an important factor in LLM safety and robust model behavior.
Chinese Translation
经过对齐的大语言模型（LLMs）应当根据用户请求的内容表现出安全行为：它们应当拒绝不安全请求，并遵从安全请求。然而，我们表明，同一请求在系统提示中被赋予不同特质时，可能引发显著不同的安全决策，我们将这种失效模式称为特质引发的安全变化。为了衡量这一失效，我们引入了基于拒绝的指标：特质引发偏差（Trait-Induced Deviation）衡量数据集层面相对于无特质基线的偏差，而特质引发翻转率（Trait-Induced Flip Rate）衡量同一请求在不同特质下是否得到不同的安全决策。随后，我们对特质引发安全偏移背后的机制进行了表示层面的分析，并发现特质会在一个低维子空间内扰动模型的安全表示。为了实现特质不变安全，即安全行为在不同特质下保持稳定，我们引入了特质不变安全调优（Trait-Invariant Safety Tuning，TIST），这是一个简单而有效的自蒸馏框架，它将大语言模型在特质条件下的行为与其无特质行为对齐。在我们的分析指导下，我们进一步提出了特质子空间中和（Trait-Subspace Neutralization，TraSN），作为 TIST 的一种实例化，它仅在所识别的特质子空间内强制不变性。实验表明，TraSN 改善了特质不变安全，并增强了对有害请求的安全性，同时保持了通用能力。我们的结果突显了特质是大语言模型安全性和鲁棒模型行为中的一个重要因素。
cs.AI / 9 / 2608.12150
Who Thinks Best Depends on How Long You Let Them: Budget-Dependent Rankings in LLM Evaluation
谁思考得最好取决于给它们多长时间：LLM评估中预算依赖的排名
large language model
大语言模型相关
Abstract
Standard evaluation of large language models assumes stable model rankings across inference conditions. We challenge this assumption by varying the token generation budget, i.e., the maximum tokens a model may produce, across seven levels (64--4,096), evaluating four models on three reasoning benchmarks (56,476 inferences). We report four findings: (i) 3--19% of items exhibit non-monotone behavior (accuracy decreasing with more budget), even after controlling for truncation, and this phenomenon is model-specific (cross-model overlap: 6--14%). (ii) Model rankings reverse across budgets on all benchmarks ($p {<} 0.01$, McNemar). (iii) Oracle analysis reveals model complementarity up to $+27.8$pp, most pronounced at constrained budgets. (iv) A budget-aware router captures 14.1% of the oracle gap cross-domain; budget features help within-domain ($+1.6$ to $+5.7$pp) but are domain-specific and hurt transfer ($-1.2$pp). These results argue for budget-conditioned evaluation protocols.
Chinese Translation
大型语言模型的标准评估假设模型排名在不同推理条件下保持稳定。我们通过改变 token 生成预算（即模型可生成的最大 token 数）在七个级别（64--4,096）上挑战这一假设，并在三个推理基准上评估四个模型（56,476 次推理）。我们报告四项发现：(i) 3--19% 的题目表现出非单调行为（随着预算增加准确率下降），即使在控制截断后也是如此，并且这一现象具有模型特异性（跨模型重叠：6--14%）。(ii) 在所有基准上，模型排名随预算变化发生逆转（$p {<} 0.01$，McNemar）。(iii) Oracle 分析揭示模型互补性最高可达 $+27.8$pp，在受限预算下最为显著。(iv) 预算感知路由器在跨域场景中捕获了 14.1% 的 oracle 差距；预算特征在域内有所帮助（$+1.6$ 至 $+5.7$pp），但具有领域特异性并损害迁移（$-1.2$pp）。这些结果支持采用以预算为条件的评估协议。
cs.AI / 10 / 2608.12304
Constructing Dynamic Master Logic Models as Knowledge Graphs for Complex System Diagnostics Using Retrieval-Augmented Large Language Models
利用检索增强大语言模型构建作为知识图谱的动态主逻辑模型以用于复杂系统诊断
large language model
大语言模型相关
Abstract
Dynamic Master Logic (DML) provides a hierarchical framework for representing system behavior by linking functional objectives to underlying structural elements. However, DML construction typically relies on expert interpretation of technical documentation, limiting scalability for complex systems. This study presents a framework for automated construction of DML models from system descriptions and their representation as Knowledge Graphs (KG-DML), using Retrieval-Augmented Generation and Large Language Models as enabling tools. Building on prior work with small-scale systems, the framework extends automated KG-DML construction and evaluation to substantially larger and more complex systems. Model construction proceeds across the DML hierarchy using targeted retrieval while preserving functional dependencies and explicit logical relationships. The resulting KG-DML supports diagnostic reasoning, safety assessment, upward failure propagation, and downward dependency tracing. A multi-level validation methodology evaluates layer-specific precision and recall, logical gate consistency, and overall structural integrity. Application to the Low-Pressure Coolant Injection system of a decommissioned Boiling Water Reactor demonstrates consistent reconstruction across repeated runs. The results show that automated KG-DML construction can transform technical documentation into executable functional models for diagnostic and reliability analysis.
Chinese Translation
动态主逻辑（DML）提供了一种层次化框架，通过将功能目标与底层结构元素联系起来来表示系统行为。然而，DML 的构建通常依赖于专家对技术文档的解读，这限制了其在复杂系统中的可扩展性。本研究提出了一个框架，用于从系统描述自动构建 DML 模型，并将其表示为知识图谱（KG-DML），该框架使用检索增强生成和大语言模型作为使能工具。在先前小规模系统工作的基础上，该框架将 KG-DML 的自动构建和评估扩展到规模更大、复杂度更高的系统。模型构建在 DML 层次结构中逐层进行，使用有针对性的检索，同时保留功能依赖关系和显式逻辑关系。生成的 KG-DML 支持诊断推理、安全评估、向上故障传播和向下依赖追踪。采用多层级验证方法评估各层级的精确率和召回率、逻辑门一致性以及整体结构完整性。将该方法应用于一座已退役沸水反应堆的低压冷却剂注入系统，结果表明在多次运行中能够一致地重建模型。结果表明，自动化的 KG-DML 构建能够将技术文档转化为可执行的功能模型，用于诊断和可靠性分析。
cs.AR / 11 / 2608.11920
NITRO: High-Performance 3D NAND Flash-Based In-Storage Computing with Enhanced Activation Dataflow
NITRO：具有增强激活数据流的高性能基于3D NAND闪存的存储内计算
large language model
大语言模型相关
Abstract
In-storage computing (ISC) is considered a next-generation memory architecture for its ability to relieve the data bottleneck between the host and the memory. While the required resources of large language models (LLMs) have increased significantly in recent years, the memory density has not scaled accordingly. Recently, several works have studied NAND flash-based processing-in-memory (NAND-PIM) schemes to exploit the high density of the memory. However, they do not address the dataflow/buffer for the intermediate values, so a simple method is to deal with the values in the slow flash memory array. To overcome such a limitation, we propose a high-performance NAND flash-based ISC architecture with enhanced activation buffering. Instead of using the very slow flash memory array for the intermediate values, our architecture buffers the values in a fast DRAM subsystem. This approach effectively handles the high-latency penalties when activations are programmed into slower TLC NAND flash. We also introduce a distributed dataflow approach for the NAND-PIM array. This approach maximizes computational parallelism by employing efficient intra-plane data mapping. The results show that our proposed architecture achieves significant performance improvements, reducing the inference latency by up to 85% compared to the baseline.
Chinese Translation
存储内计算（ISC）因其能够缓解主机与存储器之间的数据瓶颈而被视为下一代存储器架构。尽管近年来大语言模型（LLMs）所需的资源显著增加，存储密度却未相应提升。近期，多项工作研究了基于NAND闪存的存内处理（NAND-PIM）方案，以利用存储器的高密度。然而，它们没有解决中间值的数据流/缓冲问题，因此一种简单的方法是在低速闪存阵列中处理这些值。为克服这一局限，我们提出了一种具有增强激活缓冲能力的高性能基于NAND闪存的ISC架构。我们的架构不是将中间值存放在非常慢的闪存阵列中，而是将值缓冲在快速的DRAM子系统中。该方法有效处理了将激活写入较慢的TLC NAND闪存时产生的高延迟代价。我们还为NAND-PIM阵列引入了一种分布式数据流方法。该方法通过采用高效的平面内数据映射最大化计算并行性。结果表明，我们提出的架构实现了显著的性能提升，与基线相比，推理延迟最多降低85%。
cs.CL / 12 / 2608.11408
Measure, Don't Optimize: Forecasting Recovery in LLM Unlearning
测量，而非优化：预测大语言模型遗忘中的恢复
large language model
大语言模型相关
Abstract
Prior white-box studies show that large language models can retain latent traces of target knowledge after unlearning, even when the knowledge is no longer expressed in their outputs. However, existing audits remain limited to one-off diagnostics: it is unclear whether these residual signals can predict future recovery under continued training or serve as reliable optimization targets. Resolving this gap is essential to determine whether internal auditing can move beyond post-hoc evaluation toward proactive risk monitoring and safer unlearning. We propose J-Access, an inference-time audit that uses the Jacobian lens to map intermediate representations into vocabulary space and measures how often target concepts remain accessible along the model's output pathway. We hypothesize that residual accessibility reflects recovery susceptibility: knowledge that remains closer to the output pathway requires less fine-tuning to restore, leading to faster recovery. We audit 398 public unlearned models spanning eight unlearning methods. We find that: (1) most unlearned models retain access above the retain-only gold level; (2) pre-attack accessibility predicts recovery speed and extent at the model level, but cannot identify which specific facts will be recovered; and (3) directly minimizing J-Access does not promote genuine deletion. Instead, the model learns to hide knowledge from the audit, producing lower audit scores but greater post-attack recovery. These findings position J-Access as a model-level diagnostic for assessing residual susceptibility in unlearned models. We argue internal audits should serve as an independent diagnostic dimension in unlearning evaluation, and should not be converted into optimization targets without validation.
Chinese Translation
此前的白盒研究表明，大语言模型在遗忘后仍可能保留目标知识的潜在痕迹，即使这些知识不再在其输出中表达。然而，现有的审计仍局限于一次性诊断：尚不清楚这些残余信号能否预测持续训练下的未来恢复，或能否作为可靠的优化目标。解决这一差距对于确定内部审计能否从事后评估转向主动风险监测和更安全的遗忘至关重要。我们提出 J-Access，一种推理时审计方法，它利用雅可比透镜将中间表示映射到词表空间，并测量目标概念沿模型输出通路保持可访问的频率。我们假设残余可访问性反映了恢复易感性：保留得越靠近输出通路的知识，恢复时需要的微调越少，从而导致更快的恢复。我们审计了 398 个公开的已遗忘模型，涵盖八种遗忘方法。我们发现：（1）大多数已遗忘模型保留了高于仅保留集金标准水平的可访问性；（2）攻击前的可访问性在模型层面预测恢复速度和程度，但无法识别哪些具体事实会被恢复；（3）直接最小化 J-Access 并不能促进真正的删除。相反，模型学会了向审计隐藏知识，从而产生更低的审计分数，但攻击后的恢复更强。这些发现将 J-Access 定位为一种模型层面的诊断工具，用于评估已遗忘模型中的残余易感性。我们认为，内部审计应作为遗忘评估中一个独立的诊断维度，并且不应在未经验证的情况下被转化为优化目标。
cs.CL / 13 / 2608.11460
Principal Trait Analysis: Towards Deriving "Skills" in Human-AI Collaboration
主特质分析：面向人机协作中“技能”的推导
large language model
大语言模型相关
Abstract
Large Language Model-powered agents are increasingly used in the workplace via human-artificial intelligence (AI) collaboration. In this new era of work, it is important to understand the kinds of prompting traits that contribute to task success. Moreover, we need to uncover key skills required for modern professionals and inform educators on how to foster these skills among students. Existing guidelines for human-AI collaboration are built from either top-down theory or context-specific observations of human-AI interactions. However, since LLM capabilities are rapidly improving, theory may not be able to explain emerging interaction patterns, and empirical guidelines may become obsolete quickly. In this work, we explore an automated, data-driven approach to uncover patterns, which we term traits, of effective human-AI interaction that are aligned with task outcomes. We propose Principal Trait Analysis, a Principal Component Analysis-inspired algorithm for deriving common traits from patterns in LLM conversations. Our algorithm uses LLM-based processing stages to analyze corpora of human-AI collaborative session traces, deriving common traits across the dataset and scoring each human collaborator's usage style by each trait. The approach also allows domain expertise to be injected during trait discovery and selects the most distinguishing traits to be those that exhibit the highest variance across collaborators. We evaluate PTA on two human-AI collaborative coding datasets, an educational setting (students working with an AI tutor) and a professional setting (developers working with an AI coding agent). We find that PTA-derived traits are significant in explaining collaborator behavior across both settings and can help predict task outcomes. However, whether traits qualify as skills remains to be seen, due to inconclusive results on generalizability and how user traits change over time.
Chinese Translation
由大语言模型驱动的智能体正越来越多地通过人-人工智能（AI）协作被用于工作场所。在这个新的工作时代，理解哪些类型的提示特质有助于任务成功非常重要。此外，我们需要揭示现代专业人士所需的关键技能，并告知教育者如何在学生中培养这些技能。现有的人机协作指南要么建立在自上而下的理论之上，要么建立在对人机交互的特定情境观察之上。然而，由于大语言模型的能力正在迅速提升，理论可能无法解释新兴的交互模式，而经验性指南可能很快过时。在这项工作中，我们探索一种自动化、数据驱动的方法，以揭示与任务结果相一致的、有效人机交互的模式，我们称之为特质。我们提出了主特质分析，这是一种受主成分分析启发的算法，用于从大语言模型对话中的模式推导出共同特质。我们的算法使用基于大语言模型的处理阶段来分析人类-AI协作会话轨迹的语料库，在整个数据集中推导出共同特质，并按每个特质对每个人类协作者的使用风格进行评分。该方法还允许在特质发现过程中注入领域专业知识，并选择最具区分性的特质，即在协作者之间表现出最高方差的特质。我们在两个人类-AI协作编码数据集上评估PTA，一个是教育场景（学生与AI导师合作），另一个是专业场景（开发者与AI编程智能体合作）。我们发现，PTA推导出的特质在解释两种场景下的协作者行为方面具有显著意义，并且有助于预测任务结果。然而，由于在可推广性以及用户特质如何随时间变化方面的结果尚无定论，这些特质是否足以被称为技能仍有待观察。
cs.CL / 14 / 2608.11573
Reinforcing Step-level Reasoning for Effective Self-Correction in LLMs
强化步骤级推理以实现大语言模型中的有效自我纠正
large language model
大语言模型相关
Abstract
Achieving effective self-correction, where models verify and correct their own mistakes, remains a fundamental challenge for large language models (LLMs). In this work, we propose Self-Fix Step-DPO (SFS-DPO), a reinforcement learning based, two-stage framework for step-level self-verification and self-correction. The first stage strengthens step-level reasoning via step-level preference optimization, while the second stage explicitly trains models to self-verify and self-correct. We further introduce a teacher-assisted variant, SFS-DPO-R, which incorporates explanatory rationales for error verification to provide stronger corrective signals. Comprehensive in-domain and out-of-domain evaluations across multiple LLMs demonstrate that SFS-DPO and SFS-DPO-R consistently outperform prior step-level training baselines. Our analysis further reveals improvements in self-correction frequency and effectiveness, highlighting the importance of strengthening step-level reasoning for robust performance.
Chinese Translation
实现有效的自我纠正，即模型验证并纠正自身错误，对于大语言模型（LLMs）而言仍然是一个根本性挑战。在本工作中，我们提出了 Self-Fix Step-DPO（SFS-DPO），一个基于强化学习的两阶段框架，用于步骤级自我验证和自我纠正。第一阶段通过步骤级偏好优化增强步骤级推理，而第二阶段显式训练模型进行自我验证和自我纠正。我们进一步引入了一种教师辅助变体 SFS-DPO-R，它结合了用于错误验证的解释性依据，以提供更强的纠正信号。跨多个大语言模型的全面域内与域外评估表明，SFS-DPO 和 SFS-DPO-R 始终优于先前的步骤级训练基线。我们的分析进一步揭示了自我纠正频率和有效性的提升，突显了强化步骤级推理对于稳健性能的重要性。
cs.CL / 15 / 2608.11624
Learning to Persuade Exposes How Easily LLMs Abandon Correct Beliefs
学习说服揭示了LLMs多么容易放弃正确信念
large language model
大语言模型相关
Abstract
Persuasion is a core dynamic of natural language communication, shaping how large language models (LLMs) update beliefs, resolve disagreements, and reach decisions. As LLMs increasingly debate, advise, and think collaboratively with humans and each other, resistance to harmful persuasion becomes a core requirement for reliable behavior. Yet we show that this requirement is far from met: a single targeted persuasive argument is enough to collapse model accuracy to near zero, even when the argument is factually false. We formalize this threat as adversarial persuasion and introduce an adversarial reinforcement learning framework that trains persuader agents to change a target model's answer in a single interaction. First, we show that optimizing persuasion strategies through trial and error exposes vulnerabilities that static prompting misses: RL-trained persuaders raise persuasion success from approximately 24% to over 93% against the training-time persuadee. Second, we find that these learned strategies transfer to unseen models, achieving 83% attack success on Qwen-14B, 79% on Llama-3.1-8B, and 25% on GPT-4o-mini. Third, we demonstrate that a curriculum that bootstraps on more persuadable open-weight models before targeting harder models further increases GPT-4o-mini attack success from 25% to 38%. Moreover, our results reveal that optimized persuaders increasingly rely on credibility-based tactics, including fabricated citations and false authoritative evidence. Together, these findings expose a critical weakness in current LLM agents: even when they initially reason correctly, they can be steered toward false conclusions by optimized natural language influence. This positions persuasion robustness as a necessary safety criterion for multi-agent and human-AI decision-making systems.
Chinese Translation
说服是自然语言沟通的核心动力，塑造了大型语言模型（LLMs）如何更新信念、解决分歧并做出决策。随着LLMs越来越多地与人类及彼此进行辩论、提供建议和协作思考，抵御有害说服成为可靠行为的核心要求。然而我们表明，这一要求远未得到满足：一个有针对性的说服性论证就足以使模型准确率崩溃至接近零，即使该论证在事实上是虚假的。我们将这一威胁形式化为对抗性说服，并引入一个对抗性强化学习框架，训练说服者智能体在单次交互中改变目标模型的答案。首先，我们表明通过试错优化说服策略会暴露静态提示所忽略的漏洞：经RL训练的说服者将针对训练时被说服者的说服成功率从约24%提高到93%以上。其次，我们发现这些学到的策略可以迁移到未见过的模型上，在Qwen-14B上实现83%的攻击成功率，在Llama-3.1-8B上为79%，在GPT-4o-mini上为25%。第三，我们证明，在针对更难模型之前，先在更易被说服的开放权重模型上进行引导的课程学习，会将GPT-4o-mini的攻击成功率从25%进一步提高到38%。此外，我们的结果揭示，优化后的说服者越来越依赖基于可信度的策略，包括伪造引用和虚假的权威证据。综合来看，这些发现暴露了当前LLM智能体的一个关键弱点：即使它们最初推理正确，也可能被优化后的自然语言影响引导至错误结论。这将说服鲁棒性定位为多智能体和人类-AI决策系统的必要安全标准。
cs.CL / 16 / 2608.11649
Who Would You Vote For? Auditing Political Alignment in LLMs: An Italian Case-Study
你会投票给谁？审计大语言模型中的政治倾向：一项意大利案例研究
large language model
大语言模型相关
Abstract
As users increasingly turn to Large Language Models (LLMs) for information and advice on political matters, particularly during election periods, the political preferences expressed by these systems have become a matter of public interest. Prior research has shown that interactions with LLMs can influence users' political attitudes and choices, raising questions about how these models themselves evaluate political actors. In this paper, we investigate whether and how LLMs express preferences toward political parties and political leaders. We introduce a systematic and reproducible auditing framework in which multiple LLMs are prompted to evaluate parties and leaders across nine criteria. Rather than attempting to infer the models' "true" political beliefs, we focus on their observable behavior, examining consistency across evaluations, differences between models, refusal rates, and sensitivity to prompt formulation. We further investigate how these evaluations vary when models are instructed to adopt different personas. We demonstrate the framework through an Italian case study, providing a systematic analysis of LLM-generated political evaluations on italian parties and leaders.
Chinese Translation
随着用户越来越多地就政治事务，尤其是在选举期间，向大语言模型（LLMs）寻求信息和建议，这些系统所表达的政治偏好已成为公众关注的问题。
先前研究表明，与 LLMs 的互动可能会影响用户的政治态度和选择，这引发了关于这些模型自身如何评价政治行为者的问题。
在本文中，我们研究 LLMs 是否以及如何表达对政党及政治领导人的偏好。
我们引入了一个系统化且可复现的审计框架，在该框架中，多个 LLMs 被提示根据九项标准对政党和领导人进行评价。
我们并不试图推断模型“真实的”政治信念，而是关注其可观察的行为，考察各次评价之间的一致性、不同模型之间的差异、拒绝率以及对提示表述的敏感度。
我们进一步研究当模型被指示采用不同人设时，这些评价会如何变化。
我们通过一项意大利案例研究展示该框架，系统分析了 LLM 生成的关于意大利政党和领导人的政治评价。
cs.CL / 17 / 2608.11657
Semantic Lenia: Emergence of Homeostatic Solitons within the Semantic Space of Large Language Models
Semantic Lenia：大语言模型语义空间中稳态孤子的涌现
large language model
大语言模型相关
Abstract
We introduce Semantic Lenia, an artificial life framework that transforms Large Language Model (LLM) inference from a static optimization problem into a continuous dynamical system within the macroscopic logit space. By establishing a non-linear homeostatic feedback loop to dynamically balance semantic attraction and syntactic repulsion, we demonstrate the emergence of "Autonomous Semantic Solitons" -- macroscopic dissipative structures that avoid repetitive crystallization. Our exhaustive parameter sweeps map a critical "Habitable Ridge" where applied steering forces perfectly balance the model's intrinsic syntactic inertia. This approach successfully maintains generative trajectories at the edge of chaos, triggering profound abductive leaps without structural collapse and establishing a physical scaling law for machine cognition.
Chinese Translation
我们引入了 Semantic Lenia，一个将大语言模型（LLM）推理从静态优化问题转变为宏观 logit 空间中的连续动力系统的人工生命框架。通过建立非线性稳态反馈回路，动态平衡语义吸引与句法排斥，我们证明了“自主语义孤子”的涌现——这些是避免重复性结晶的宏观耗散结构。我们详尽的参数扫描绘制出一条关键的“宜居山脊”，在此施加的引导力与模型固有的句法惯性完美平衡。这种方法成功地将生成轨迹维持在混沌边缘，触发深刻的溯因跳跃而不发生结构崩溃，并建立了机器认知的物理标度律。
cs.CL / 18 / 2608.11660
Hybrid-Policy Self-Editing for Composable Unstructured Knowledge Editing
面向可组合非结构化知识编辑的混合策略自编辑
large language model
大语言模型相关
Abstract
Large language models (LLMs) achieve remarkable performance across natural language tasks, yet they are trained on static corpora and their knowledge quickly becomes outdated in a fast-changing world. This motivates knowledge editing (KE), which updates specific knowledge in an LLM without changing unrelated others. Recent works move from structured knowledge triples toward unstructured KE (UKE), where the edit is a free-form passage that may state multiple facts at once. Nonetheless, existing editors inject such a passage yet fail to use it: the edited model can recall the passage, but can neither answer atomic questions about its facts nor compose them into multi-hop reasoning. We attribute this missing property, which we term composability, to editors' passive reliance on the fixed passage as the sole learning source. In response, we cast editing as a proactive self-distillation from a privileged in-context state of the same model, which requires no external supervision. We further reveal that due to the novelty of the injected knowledge, the pre-edited model's own rollouts rarely cover it, which limits the effectiveness of pure on-policy distillation. To close this gap, we propose HPSE, which builds a hybrid rollout that steps in to place missing facts onto the student's own trajectory precisely where its coverage fails, while staying on-policy elsewhere. We theoretically analyze HPSE's advantage over pure on-policy distillation, and empirically establish its plug-and-play improvements across four LLM backbones and two KE editors under various scenarios.
Chinese Translation
大语言模型（LLMs）在各类自然语言任务上取得了显著性能，但它们是在静态语料库上训练的，其知识在快速变化的世界中很快就会过时。这推动了知识编辑（KE）的发展，即在不改变其他无关知识的情况下更新大语言模型中的特定知识。最近的工作从结构化知识三元组转向非结构化知识编辑（UKE），其中编辑内容是一段自由形式的文本段落，可能一次性陈述多个事实。然而，现有编辑器注入了这类段落却未能真正利用它：编辑后的模型能够回忆起该段落，但既无法回答关于其中事实的原子性问题，也无法将这些事实组合成多跳推理。我们将这一缺失的性质（称之为可组合性）归因于编辑器被动地依赖固定段落作为唯一学习来源。为此，我们将编辑重新表述为从同一模型的一种特权化上下文状态中进行的主动自蒸馏，这不需要任何外部监督。我们进一步揭示，由于注入知识的新颖性，预编辑模型自身的生成轨迹很少覆盖这些知识，这限制了纯同策略蒸馏的有效性。为了弥补这一差距，我们提出了 HPSE，它构建一种混合生成轨迹，在学生自身轨迹中覆盖失败的位置介入并将缺失事实放置到该轨迹上，同时在其他位置保持同策略。我们从理论上分析了 HPSE 相对于纯同策略蒸馏的优势，并在各种场景下、在四个大语言模型骨干和两个 KE 编辑器上实证确立了其即插即用的改进。
cs.CL / 19 / 2608.11715
When the API Speaks the Wrong Language: Revisiting Post-Training for Multilingual Tool Use
当 API 说错语言时：重新审视面向多语言工具使用的后训练
large language model
大语言模型相关
Abstract
The reliability of Large Language Models (LLMs) for API calling degrades in multilingual settings. A common failure occurs when a model selects the correct tool but generates argument values in an inconsistent language, which we term Argument Language Mismatch (ALM). Although semantically correct, such outputs are operationally invalid and not captured by standard API-calling metrics. We revisit post-training strategies for mitigating ALM and find that, in our benchmark, supervised fine-tuning (SFT) provides a strong baseline, substantially improving argument language consistency and end-to-end function call accuracy. Under consistent model selection, SFT achieves performance comparable to, and sometimes exceeding more complex reinforcement learning (RL) approaches. We further examine whether RL with structured, argument-aware rewards offers additional benefits. While methods such as Group Relative Policy Optimization (GRPO) can improve language consistency and better preserve general reasoning ability, these gains are incremental and most pronounced in generalization and multi-objective trade-offs. Overall, our results suggest that much of the performance in multilingual API grounding can be achieved through careful supervised training, with RL providing targeted rather than fundamental improvements.
Chinese Translation
大型语言模型（LLMs）在多语言环境中进行 API 调用的可靠性会下降。一种常见失败是：模型选择了正确的工具，但生成参数值时使用了不一致的语言，我们称之为参数语言不匹配（ALM）。尽管这类输出在语义上是正确的，但在操作上无效，且不会被标准 API 调用指标所捕获。我们重新审视了用于缓解 ALM 的后训练策略，并发现在我们的基准测试中，监督微调（SFT）提供了一个强基线，显著提升了参数语言一致性和端到端函数调用准确率。在一致的模型选择条件下，SFT 的性能可与更复杂的强化学习（RL）方法相媲美，有时甚至超过后者。我们进一步考察了带有结构化、参数感知奖励的 RL 是否能带来额外收益。虽然诸如分组相对策略优化（GRPO）等方法能够改善语言一致性，并更好地保持通用推理能力，但这些收益是渐进式的，且在泛化和多目标权衡中最为明显。总体而言，我们的结果表明，多语言 API 调用任务中的大部分性能可以通过精心的监督训练来实现，而 RL 提供的更多是针对性改进而非根本性改进。
cs.CL / 20 / 2608.11735
Locating and Controlling Implicit Personalization in Large Language Models
定位与控制大型语言模型中的隐式个性化
large language model
大语言模型相关
Abstract
Large language models (LLMs) often shift their outputs in response to implicit demographic cues even when users never state a demographic identity. Previous work has documented this behavior, but the connection between these behavioral changes and the model's internal activations remains unclear. Using matched cued and neutral conversations across five LLMs, we establish that a localized internal activation signal tracks changes in recommendations, with correlations up to r=0.87. When multiple cues appear together, their internal signals largely combine, but the changes in output do not simply add up. We further show that removing the internal signal associated with one cue can suppress its influence, often more effectively than asking the model to ignore demographics via prompting, while largely preserving general benchmark performance. However, the ability to selectively remove one dimension's influence while leaving co-present dimensions intact remains highly model- and attribute-specific. These results connect implicit personalization behavior to an internal signal that can be analyzed and causally controlled.
Chinese Translation
大型语言模型（LLMs）常常会因隐含的人口统计线索而改变其输出，即使用户从未明确陈述其人口统计身份。此前的研究已记录了这种行为，但这些行为变化与模型内部激活之间的联系仍不清楚。通过在五个LLM上使用匹配的有线索对话和中性对话，我们确定了一个局部内部激活信号能够跟踪推荐内容的变化，相关性最高可达r=0.87。当多个线索同时出现时，它们的内部信号在很大程度上会相互组合，但输出的变化并不是简单相加。我们进一步表明，移除与某个线索相关的内部信号可以抑制其影响，这通常比通过提示要求模型忽略人口统计信息更有效，同时能在很大程度上保持一般基准性能。然而，选择性地移除某一维度的影响、同时保留同时存在的其他维度影响的能力，仍然高度依赖于具体模型和具体属性。这些结果将隐式个性化行为与一个可被分析和因果控制的内部信号联系了起来。
cs.CL / 21 / 2608.11742
Ripple-Pivot Search: Active Parallel Decoding for Diffusion Large Language Models
Ripple-Pivot Search：面向扩散大语言模型的主动并行解码
diffusionlarge language model
扩散模型相关
大语言模型相关
Abstract
Diffusion Large Language Models (dLLMs) have emerged as a competitive alternative to autoregressive language models, offering the potential for substantially faster inference through parallel decoding. Existing parallel decoding schedulers typically commit positions only after they meet a per-position criterion, overlooking how early commitments may benefit subsequent decoding. We identify a ripple effect in dLLM decoding: proactively committing a mid-entropy pivot position can induce a pronounced reduction in uncertainty across the remaining masked positions. This uncertainty reduction allows subsequent steps to unmask more tokens in parallel, thereby accelerating the overall decoding process. To exploit the ripple effect, we propose Ripple-Pivot Search (RPS), a novel training-free decoding method that seeks mid-entropy positions as promising candidate pivots (where to decode), and determines their token assignment that yields the greatest downstream benefit via lookahead evaluation (what to decode). Across 3 dLLMs and 4 reasoning and code-generation benchmarks, RPS achieves 4-10$\times$ wall-clock speedup over the standard decoder while preserving generation quality, and improves accuracy over the previous lookahead baseline by up to 5.49% while delivering higher throughput in most settings. When integrated with KV caching, RPS further achieves up to 18$\times$ wall-clock speedup over the standard decoder.
Chinese Translation
扩散大语言模型（dLLMs）已成为自回归语言模型的一个具有竞争力的替代方案，并有望通过并行解码实现显著更快的推理。现有的并行解码调度器通常仅在位置满足逐位置准则后才提交这些位置，忽视了早期提交可能给后续解码带来的益处。我们在 dLLM 解码中发现了一种涟漪效应：主动提交一个中等熵枢轴位置可显著降低其余被掩码位置的不确定性。这种不确定性降低允许后续步骤并行去掩码更多 token，从而加速整体解码过程。为了利用这种涟漪效应，我们提出了 Ripple-Pivot Search（RPS），这是一种无需训练的新型解码方法，该方法寻找中等熵位置作为有前景的候选枢轴（在哪里解码），并通过前瞻评估确定能带来最大下游收益的 token 分配（解码什么）。在 3 个 dLLM 和 4 个推理与代码生成基准上，RPS 在保持生成质量的同时，相较于标准解码器实现了 4-10$\times$ 的墙钟加速；并且相较于先前的前瞻基线，RPS 在大多数设置下提供更高吞吐量的同时将准确率提高了最多 5.49%。当与 KV 缓存结合使用时，RPS 进一步实现了相较于标准解码器最高 18$\times$ 的墙钟加速。
cs.CL / 22 / 2608.11753
LabelFusion-TS: Fusing Large Language Models, Transformer Encoders, and Financial Time Series for Monetary-Policy Stance Classification
LabelFusion-TS：融合大语言模型、Transformer编码器和金融时间序列以用于货币政策立场分类
large language model
大语言模型相关
Abstract
Financial text is produced and interpreted within a market environment, yet financial text classifiers almost always receive text alone. We study whether financial time series are useful as an additional input on the task of classifying sentences from Federal Reserve communication as hawkish, dovish, or neutral. Our system, \lfts{}, extends the \lf{} architecture with this modality: a small voting network combines three independently trained components, a fine-tuned RoBERTa encoder, a prompted large language model (LLM), and a fused ensemble of time-series transformers over the market series of the months preceding publication. Because only about a thousand annotated sentences are available for training, the RoBERTa encoder is first pre-trained on sentences annotated automatically by the LLM and only then fine-tuned on the human labels. Trained on Federal Open Market Committee (FOMC) communication up to 2015 and evaluated on 2015--2022, the fused system achieves 70.2\% weighted F1 -- against 64.1\% for the zero-shot LLM -- and overtakes it with as few as 240 human-labelled sentences. We take this as initial evidence for market time series as an input modality in financial text classification.
Chinese Translation
金融文本是在市场环境中产生和解读的，然而金融文本分类器几乎总是只接收文本。我们研究金融时间序列作为额外输入在将美联储沟通中的句子分类为鹰派、鸽派或中性这一任务上是否有用。我们的系统 LabelFusion-TS 扩展了 LabelFusion 架构，加入了这一模态：一个小型投票网络结合了三个独立训练的组件，包括一个经过微调的 RoBERTa 编码器、一个经过提示的大语言模型（LLM）以及一个在发布前几个月市场序列上训练的时间序列 Transformer 融合集成。由于只有大约一千个标注句子可用于训练，RoBERTa 编码器首先在由 LLM 自动标注的句子上进行预训练，然后才在人工标签上进行微调。在截至 2015 年的联邦公开市场委员会（FOMC）沟通数据上训练，并在 2015--2022 年期间的数据上评估，融合系统达到 70.2% 的加权 F1——而零样本 LLM 为 64.1%——并且仅用 240 个人工标注句子就超过了它。我们将此视为市场时间序列作为金融文本分类输入模态的初步证据。
cs.CL / 23 / 2608.11758
AWARe: Mitigating Catastrophic Forgetting via Activation-Weighted Adaptive REtention
AWARe：通过激活加权自适应保留缓解灾难性遗忘
large language model
大语言模型相关
Abstract
Multimodal Large Language Models (MLLMs) exhibit strong generalization and reasoning abilities due to large-scale multimodal pre-training. However, fine-tuning these models on downstream tasks often leads to catastrophic forgetting, where newly learned task-specific knowledge degrades previously acquired capabilities. This issue arises because gradient updates for new tasks overwrite parameters critical to prior knowledge, limiting the practical deployment of MLLMs. To address this challenge, we propose Activation-Weighted Adaptive REtention (AWARe), a fine-tuning method that mitigates catastrophic forgetting by dynamically controlling parameter updates based on activation patterns. AWARe assigns activation-based importance scores to parameters, selectively freezing those essential for preserving prior capabilities while allowing less important parameters to adapt to new tasks. Importantly, AWARe operates without modifying model architectures, ensuring compatibility with existing inference engines. Extensive experiments demonstrate that AWARe effectively preserves upstream capabilities while achieving superior downstream performance compared to existing methods. Code is available at https://github.com/kaln27/AWARe.
Chinese Translation
多模态大语言模型（MLLMs）由于大规模多模态预训练而展现出强大的泛化和推理能力。然而，在下游任务上微调这些模型通常会导致灾难性遗忘，即新学习的任务特定知识会削弱先前获得的能力。这一问题产生的原因是，新任务的梯度更新会覆盖对先前知识至关重要的参数，从而限制了多模态大语言模型的实际部署。为应对这一挑战，我们提出了激活加权自适应保留（AWARe），这是一种微调方法，通过基于激活模式动态控制参数更新来缓解灾难性遗忘。AWARe 为参数分配基于激活的重要性分数，选择性地冻结那些对于保留先前能力至关重要的参数，同时允许较不重要的参数适应新任务。重要的是，AWARe 在不修改模型架构的情况下运行，确保与现有推理引擎兼容。大量实验表明，与现有方法相比，AWARe 在有效保持上游能力的同时，实现了更优的下游性能。代码可在 https://github.com/kaln27/AWARe 获取。
cs.CL / 24 / 2608.11788
TELLME: Test-Enhanced Learning for Language Model Enrichment
TELLME：面向语言模型增强的测试增强学习
large language model
大语言模型相关
Abstract
Continual pre-training (CPT) has been widely adopted as a method for domain adaptation in large language models. However, CPT has consistently been accompanied by challenges, such as the difficulty of acquiring large-scale domain-specific datasets and high computational costs. In this study, we propose a novel method called Test-Enhanced Learning for Language Model Enrichment (TELLME) to alleviate these issues. TELLME leverages the TestEnhanced Learning (TEL) principle, whereby the model's training efficiency is improved using quizzes during training. It integrates this principle with CPT, thereby promoting efficient domain-specific knowledge acquisition and long-term memory retention. Experimental results demonstrate that TELLME outperforms existing methods by up to 23.6% in the financial domain and achieves a 9.8% improvement in long-term memory retention.
Chinese Translation
持续预训练（CPT）已被广泛采用为大型语言模型领域适应的一种方法。然而，CPT 一直伴随着诸多挑战，例如难以获取大规模领域特定数据集以及计算成本高昂。在本研究中，我们提出了一种新颖的方法，称为面向语言模型增强的测试增强学习（TELLME），以缓解这些问题。TELLME 利用了测试增强学习（TEL）原理，即通过在训练过程中使用测验来提高模型的训练效率。它将这一原理与 CPT 相结合，从而促进高效的领域特定知识获取和长期记忆保持。实验结果表明，TELLME 在金融领域最多比现有方法性能提升 23.6%，并在长期记忆保持方面实现了 9.8% 的改进。
cs.CL / 25 / 2608.11919
LazyTrain: Limited-resource Allocation toward Zero-waste Yield Optimization in Large Language Model Training
LazyTrain：在大语言模型训练中面向零浪费产出优化的有限资源分配
large language model
大语言模型相关
Abstract
Training large language models on limited hardware is increasingly a scheduling problem across GPU compute, host memory, PCIe transfer, and storage bandwidth. Existing offloading systems reduce GPU residency, and MegaTrain shows that a CPU-master layer-streaming executor can train large models on a single GPU, but fixed checkpointing and placement heuristics still leave communication exposed on the critical path. We propose LazyTrain, an optimization layer over a layer-streaming executor. LazyTrain formulates checkpoint selection, activation placement, recomputation, and CPU-GPU-NVMe communication overlap as a mixed-integer scheduling problem, then executes the solved policy during training. It further couples 8-bit optimizer states with fast gradient clipping as a single Hybrid 8-bit operator: state compression reduces optimizer-state memory, while fast clipping counteracts the additional CPU-side update overhead. Across H800 experiments from Qwen2.5-3B to Qwen3.6-27B, LazyTrain improves sustained TFLOPS over matched baselines runs by approximately 1.24$\times$; RTX 3090 experiments likewise increase the maximum feasible batch size by one at each model scale. In the primary Qwen3.6-27B H800 MetaMathQA run, LazyTrain reaches 219.95 TFLOPS and 1361 tokens/s at batch size 72, peaks at 68.84\,GB of GPU memory, and obtains 95.42\% exact-match accuracy on the full evaluation split. The source code is available at https://github.com/DataArcTech/LazyTrain.
Chinese Translation
在有限硬件上训练大语言模型正日益成为一个涉及 GPU 计算、主机内存、PCIe 传输和存储带宽的调度问题。现有的卸载系统减少了 GPU 驻留，MegaTrain 表明，以 CPU 为主控的层流式执行器可以在单个 GPU 上训练大模型，但固定的检查点设置与放置启发式仍使通信暴露在关键路径上。我们提出 LazyTrain，一个位于层流式执行器之上的优化层。LazyTrain 将检查点选择、激活放置、重计算以及 CPU-GPU-NVMe 通信重叠建模为一个混合整数调度问题，然后在训练期间执行求解得到的策略。它还进一步将 8 位优化器状态与快速梯度裁剪耦合成一个 Hybrid 8-bit 算子：状态压缩减少了优化器状态内存占用，而快速裁剪抵消了额外的 CPU 端更新开销。在从 Qwen2.5-3B 到 Qwen3.6-27B 的 H800 实验中，LazyTrain 相比匹配的基线运行将持续 TFLOPS 提高约 1.24$\times$；RTX 3090 实验同样在每个模型规模上将最大可行批量大小提高 1。在主要的 Qwen3.6-27B H800 MetaMathQA 运行中，LazyTrain 在批量大小为 72 时达到 219.95 TFLOPS 和 1361 tokens/s，GPU 内存峰值达到 68.84 GB，并在完整评估划分上获得 95.42% 的精确匹配准确率。源代码可在 https://github.com/DataArcTech/LazyTrain 获取。
cs.CL / 26 / 2608.11947
Accuracy and Order Sensitivity Diverge Under Label-Free Strategies
准确率与顺序敏感性在无标签策略下发生背离
large language model
大语言模型相关
Abstract
Multiple-choice benchmarks are widely used to evaluate large language models, but MCQ scores conflate knowledge with sensitivity to option order, which makes them unreliable measures of model knowledge. In this paper, we test whether preventing a model from seeing option labels while committing to an answer removes positional influence and, in turn, improves performance. We evaluate two different strategies for mitigating bias. The first uses a generation-then-matching approach, and the second scores options in isolation, which is positionally unbiased by construction. Neither reliably improves accuracy. A complete decomposition shows that the bottleneck is withholding options, not the matching step. The only configuration that consistently matches the baseline is the one that shows the model all options paired with an LLM matcher. However, eliminating positional influence entirely still does not reliably yield accuracy gains, while cyclic permutation often improves them. For two-stage prompting, an aggregate measure of recall imbalance and a direct per-question measure of order sensitivity both fail to show reliable debiasing.
Chinese Translation
多项选择基准被广泛用于评估大型语言模型，但MCQ分数将知识与选项顺序敏感性混为一谈，这使得它们成为不可靠的模型知识度量。在本文中，我们检验在让模型给出答案时阻止其看到选项标签是否能够消除位置影响，并进而提升性能。我们评估了两种不同的减轻偏差策略。第一种采用先生成后匹配的方法，第二种则对选项进行逐一单独评分，后者在构造上即无位置偏差。两者都未能可靠地提高准确率。完整的分解分析表明，瓶颈在于不展示选项，而不是匹配步骤。唯一能够持续达到基线水平的配置是：向模型展示所有选项，并配以LLM匹配器。然而，完全消除位置影响仍不能可靠地带来准确率提升，而循环置换往往会提高准确率。对于两阶段提示方法，召回不平衡的聚合度量以及逐题的直接顺序敏感性度量都未能显示出可靠的去偏效果。
cs.CL / 27 / 2608.11981
Benchmarking Trustworthiness of SLMs: Pre-trained vs. Compressed
SLMs可信性的基准评测：预训练与压缩
large language model
大语言模型相关
Abstract
Small Language Models (SLMs) have emerged as a more efficient alternative to traditional Large Language Models (LLMs), offering promising potential in resource-constrained scenarios. Existing approaches to building SLMs typically follow two paths: training compact models from scratch, or compressing larger pre-trained models using methods such as pruning, quantization, or distillation. As language models become increasingly integrated into real-world applications, ensuring their trustworthiness has become a critical concern. However, how to build trustworthy SLMs remains an underexplored question. In this work, we present a comprehensive evaluation of SLM trustworthiness across multiple dimensions, including fairness, robustness, privacy, and ethics. We first examine the effects of pruning and quantization, and find that quantization is significantly more effective in preserving trustworthiness compared to pruning. More importantly, we demonstrate that compressing a reliable large model via quantization can produce SLMs with superior trustworthiness and adaptability compared to using small models trained from scratch. Furthermore, knowledge distillation from trustworthy teacher models can further enhance the reliability of SLMs. We hope our findings provide practical guidance and a foundation for future research into the development and deployment of trustworthy small language models.
Chinese Translation
小语言模型（SLMs）已成为传统大语言模型（LLMs）的一种更高效替代方案，在资源受限场景中展现出广阔潜力。构建小语言模型的现有方法通常遵循两条路径：从头训练紧凑模型，或使用剪枝、量化或蒸馏等方法压缩较大的预训练模型。随着语言模型日益融入现实世界应用，确保其可信性已成为一个关键问题。然而，如何构建可信的小语言模型仍是一个尚未充分探索的问题。在这项工作中，我们对小语言模型的可信性进行了多维度综合评估，包括公平性、鲁棒性、隐私性和伦理。我们首先考察了剪枝和量化的影响，发现与剪枝相比，量化在保持可信性方面明显更有效。更重要的是，我们证明，与使用从头训练的小模型相比，通过量化压缩可靠的大模型可以产生具有更优可信性和适应性的小语言模型。此外，从可信教师模型进行知识蒸馏可以进一步增强小语言模型的可靠性。我们希望我们的发现为未来可信小语言模型的开发与部署研究提供实践指导和基础。
cs.CL / 28 / 2608.12008
Asymptotic Risk Calibration for Selective Question Answering
面向选择性问答的渐近风险校准
large language model
大语言模型相关
Abstract
Large language models (LLMs) may generate fluent but incorrect answers, making uncertainty quantification important for reliable question answering. However, heuristic uncertainty scores cannot perfectly distinguish correct predictions from incorrect ones, and directly applying a fixed uncertainty threshold provides no statistical control over the error rate among accepted answers. To address this limitation, we propose A-CRC-QA, a post-hoc calibration framework for uncertainty-aware selective question answering. The proposed method reformulates selection-conditioned error control as a linear expectation constraint and applies a monotonized empirical-risk calibration procedure inspired by conformal risk control. Since the resulting instance-wise loss is generally non-monotone with respect to the acceptance threshold, our framework targets asymptotic rather than finite-sample risk control. A-CRC-QA is model-agnostic, requires no additional training, and can be combined with different uncertainty estimators. Experiments on CoQA and MedMCQA demonstrate its applicability to both open-ended and closed-ended question answering, achieving a favorable trade-off between accepted-answer reliability and answer retention compared with uncalibrated and confidence-bound-based baselines.
Chinese Translation
大型语言模型（LLMs）可能生成流畅但错误的答案，这使得不确定性量化对于可靠问答十分重要。然而，启发式不确定性分数无法完美区分正确预测与错误预测，直接应用固定的不确定性阈值也无法对已接受答案中的错误率提供统计控制。为了解决这一局限，我们提出了 A-CRC-QA，这是一种用于不确定性感知的选择性问答的事后校准框架。所提方法将选择条件下的误差控制重新表述为线性期望约束，并应用一种受共形风险控制启发的单调化经验风险校准过程。由于由此产生的实例级损失通常关于接受阈值是非单调的，我们的框架以渐近风险控制而非有限样本风险控制为目标。A-CRC-QA 与模型无关，无需额外训练，并且可以与不同的不确定性估计器结合。在 CoQA 和 MedMCQA 上的实验表明，该方法适用于开放式和封闭式问答，并且与未校准基线和基于置信边界的基线相比，在已接受答案可靠性与答案保留之间取得了有利的权衡。
cs.CL / 29 / 2608.12018
Poly-Dialectal Neural Machine Translation System for Bangla Regional Dialects
面向孟加拉地区方言的多方言神经机器翻译系统
large language model
大语言模型相关
Abstract
Regional dialectal variation poses a fundamental challenge to natural language processing (NLP) in Bangla, where over 240 million speakers communicate across diverse regional variants that diverge significantly from Standard Colloquial Bangla (SCB) in phonology, morphology, and lexicon. Contemporary neural machine trans- lation (NMT) architectures and large language models (LLMs) predominantly as- sume a homogeneous language distribution, resulting in severe performance degra- dation when translating low-resource regional dialects. In this work, we present a unified Poly-Dialectal Neural Machine Translation System capable of multi-directional translation across 12 Bangla regional dialects without routing through an inter- mediary standard pivot. We compile the largest multi-dialect parallel corpus for Bangla to date, comprising 51,531 non-null parallel sentence pairs across 12 di- alects, incorporating 2,500 expert-verified, bidirectional parallel sentence pairs for five previously unaddressed dialects. Evaluating sequence-to-sequence architec- tures under Weight-Decomposed Low-Rank Adaptation (DoRA), our fine-tuned BanglaT5 model achieves state-of-the-art translation performance (29.26 BLEU, 57.26 chrF++), outperforming NLLB-200 (615M) and mBART-50 (611M) while preserving morphological coherence. Furthermore, we conduct a systematic cross- dialectal transfer analysis and dataset scaling study, establishing empirical thresh- olds for low-resource dialect adaptation. Finally, we deploy the optimized INT8- quantized model as an open-access web application to promote digital inclusion for marginalized dialect communities. The complete dataset is publicly available at Mendeley Data (https://data.mendeley.com/datasets/v9cf66fk2t/2).
Chinese Translation
地区方言差异给孟加拉语的自然语言处理（NLP）带来了根本性挑战，超过2.4亿使用者在不同地区变体之间交流，这些变体在音系、形态和词汇方面与标准口语孟加拉语（SCB）存在显著差异。当代神经机器翻译（NMT）架构和大语言模型（LLM）主要假设语言分布是同质的，这导致在翻译低资源地区方言时性能严重下降。在这项工作中，我们提出了一个统一的多方言神经机器翻译系统，能够在12种孟加拉地区方言之间进行多方向翻译，而无需经由中间标准枢轴转译。我们构建了迄今为止规模最大的孟加拉语多方言平行语料库，包含12种方言的51,531个非空平行句对，并纳入了针对五种此前未被处理的方言的2,500个经专家验证的双向平行句对。在权重分解低秩适应（DoRA）下评估序列到序列架构，我们微调后的BanglaT5模型取得了最先进的翻译性能（29.26 BLEU，57.26 chrF++），优于NLLB-200（615M）和mBART-50（611M），同时保持了形态连贯性。此外，我们进行了系统的跨方言迁移分析和数据集规模研究，确立了低资源方言适应的经验阈值。最后，我们将优化后的INT8量化模型部署为一个开放访问的Web应用，以促进边缘化方言社区的数字包容。完整数据集已在Mendeley Data上公开（https://data.mendeley.com/datasets/v9cf66fk2t/2）。
cs.CL / 30 / 2608.12138
A corpus-specific clinical RAG system matches or outperforms newer frontier LLMs on HealthBench
语料库特定的临床 RAG 系统在 HealthBench 上匹配或超越更新的前沿大语言模型
large language model
大语言模型相关
Abstract
General-purpose large language models (LLMs) have recently been reported to match or exceed specialized clinical AI tools on medical benchmarks, but such comparisons draw on a narrow set of systems and on benchmarks developed largely in high-income settings. We evaluate VITA, a retrieval-augmented generation (RAG) system purpose-built for contextual knowledge retrieval in India and other low- and middle-income (LMIC) settings. VITA retrieves from a curated corpus of disease-specific guidelines, India-specific antimicrobial resistance data, national formulary constraints, and resource-limited care protocols; its architecture and corpus are proprietary, but the benchmark, the physician-written rubrics, and our full response and scoring outputs are public for independent verification. On 4,023 English-language HealthBench questions (80.5% of the benchmark), scored with a GPT-4.1 judge, VITA ranked first with 51.9% of possible rubric points, ahead of GPT-5.4 (46.1%), o4-mini (44.3%), Gemini 3.1 Pro (42.6%), and Claude Sonnet 4.6 (37.3%), and scored highest on 45.4% of questions. To test robustness to newer models and judge lineage, a 500-question subset was re-run against current-generation models (GPT-5.5, Claude Opus 4.8, Gemini 3.5 Pro, Grok 4.3) and graded by a neutral open-weight judge (DeepSeek-V4-Pro) sharing no lineage with any system tested. Here the gap narrowed to parity: VITA and GPT-5.5 were statistically indistinguishable on mean per-question score, while VITA led on points-weighted score and won the most questions. VITA's advantages in accuracy and completeness persisted under the neutral judge; its communication scores were lower. These results indicate that a purpose-built clinical RAG system remains competitive with frontier LLMs on an open benchmark, consistent with corpus specificity as a design variable that improves grounding at some cost to communication polish.
Chinese Translation
通用大语言模型（LLM）最近被报道在医学基准上达到或超过专门的临床人工智能工具，但此类比较所依赖的系统集合范围较窄，并且所依赖的基准大多是在高收入环境中开发的。我们评估了 VITA，这是一个检索增强生成（RAG）系统，专为印度及其他低收入和中等收入（LMIC）环境中的情境化知识检索而构建。VITA 从精心整理的语料库中检索，该语料库包含疾病特定指南、印度特有的抗菌药物耐药性数据、国家处方集限制以及资源有限环境下的诊疗方案；其架构和语料库是专有的，但基准、由医生编写的评分标准以及我们完整的回答和评分输出均已公开，可供独立验证。在 4,023 道英文 HealthBench 问题（占基准的 80.5%）上，由 GPT-4.1 评审模型评分，VITA 以可获得评分标准总分的 51.9% 排名第一，领先于 GPT-5.4（46.1%）、o4-mini（44.3%）、Gemini 3.1 Pro（42.6%）和 Claude Sonnet 4.6（37.3%），并在 45.4% 的问题上得分最高。为了检验对更新模型和评审模型谱系的稳健性，我们重新对一个 500 道问题的子集进行了测试，与当前一代模型（GPT-5.5、Claude Opus 4.8、Gemini 3.5 Pro、Grok 4.3）进行对比，并由一个中立的开放权重评审（DeepSeek-V4-Pro）评分，该评审与所测试的任何系统都没有谱系关系。这里差距缩小至不相上下：VITA 与 GPT-5.5 在每题平均得分上无统计学差异，而 VITA 在按分值加权的得分上领先，并且赢得最多问题。VITA 在准确性和完整性方面的优势在中立评审下仍然存在；其沟通得分较低。这些结果表明，一个专门构建的临床 RAG 系统在开放基准上仍然具有与前沿大语言模型竞争的能力，这与将语料库特异性作为设计变量相一致：它以一定程度的沟通润色损失为代价，提高了回答的依据性。
cs.CL / 31 / 2608.12149
Massive Activations in Hybrid Linear Attention Large Language Models: Pre-Attention Spikes and Inter-Spike Plateaus
混合线性注意力大语言模型中的大规模激活：注意力前尖峰与尖峰间平台
large language model
大语言模型相关
Abstract
We present the first systematic study of Massive activations (MAs) in layer-interleaved HLA LLMs and uncover two architecture-aligned morphologies: MAs consistently spike immediately before full attention layers, forming pre-attention spikes (PAS), and can persist through intervening linear attention layers, giving rise to inter-spike plateaus (ISP). As full attention becomes denser, successive PAS become increasingly connected through ISP, ultimately recovering the stable MA morphology of full attention LLMs. We establish the recurrence of this organization across five linear attention architectures, six hybridization configurations, five data domains, and representative open-source hybrid models spanning 1.2B to 397B total parameters. Controlled pretraining of GDN-based hybrids at scales up to 1.3B shows that both morphologies emerge early and respond asymmetrically to output gating: full attention output gating strongly attenuates their absolute magnitudes without eliminating their layerwise organization, whereas removing GDN gates yields comparatively modest amplification. Mechanistically, our systematic-outlier analysis supports a shared lifecycle account governed by the timing of MA cancellation. PAS follows a localized write-sink-cancel process, while the extended persistence of ISP is consistent with delayed cancellation. At the full attention limit, this account recovers the stable MA morphology characteristic of full attention LLMs. Our code is available at https://github.com/StartluxLabs/Massive-Activations-HLA.
Chinese Translation
我们首次对层间交错HLA大语言模型中的大规模激活（MAs）进行了系统研究，并揭示出两种与架构对齐的形态：大规模激活始终在紧邻全注意力层之前出现尖峰，形成注意力前尖峰（PAS），并且能够持续穿过中间的线性注意力层，从而产生尖峰间平台（ISP）。随着全注意力变得更密集，相继的PAS通过ISP日益连接起来，最终恢复全注意力大语言模型中稳定的大规模激活形态。我们在五种线性注意力架构、六种混合配置、五个数据领域以及总参数量从1.2B到397B的代表性开源混合模型中，确立了这种组织结构的反复出现。对基于GDN的混合模型在高达1.3B规模上的受控预训练表明，这两种形态在早期即出现，并对输出门控呈现不对称响应：全注意力输出门控显著减弱其绝对幅度，但并未消除其逐层组织；而移除GDN门控则产生相对温和的放大。在机制上，我们的系统离群值分析支持一种由大规模激活消除时机所支配的共享生命周期解释。PAS遵循一个局部化的写入-汇聚-消除过程，而ISP的延长持续性则与延迟消除一致。在全注意力极限下，这一解释恢复了全注意力大语言模型所特有的稳定大规模激活形态。我们的代码可在 https://github.com/StartluxLabs/Massive-Activations-HLA 获取。
cs.CL / 32 / 2608.12218
Information Abundance Paradox: Long-Context Training Undermines Parametric Knowledge
信息丰裕悖论：长上下文训练削弱参数化知识
large language model
大语言模型相关
Abstract
Large language models are increasingly trained and deployed with long contexts that span documents, code repositories, and interaction histories. This scaling reflects the implicit assumption that training on longer contexts will only help the model by exposing it to richer evidence. We challenge this view by studying how the context window shapes a model's mode of learning, shifting it between parametric internalization and contextualization. We propose the Information Abundance Paradox, which hypothesizes that abundant relevant information in the training context can reduce the incentive to encode that information parametrically, thereby increasing reliance on context. In pretraining with long documents, increasing the context window improves language modeling, natural language understanding, and closed-book MCQA only up to an intermediate optimum, after which performance consistently declines. In supervised fine-tuning, more task-relevant train-time context improves performance with supporting context, but reduces robustness when context is absent or misleading at test time. Our analysis suggests that this behavior arises when longer context provides a lower complexity solution. Mechanistically, training with informative context shifts gradient pressure from feed-forward networks, often linked to parametric knowledge, toward attention modules, and causal interventions show that this shift increases reliance on context during inference. Overall, these findings support the Information Abundance Paradox and suggest that scaling toward near-infinite context is not simply a matter of supplying more data, even when high-quality long-context data is abundant.
Chinese Translation
大语言模型越来越多地在涵盖文档、代码仓库和交互历史的长上下文上进行训练和部署。这种扩展反映了一种隐含假设：在更长上下文上训练只会通过让模型接触更丰富的证据而对其有益。我们通过研究上下文窗口如何塑造模型的学习模式，使其在参数化内化与上下文化之间转换，来质疑这一观点。我们提出信息丰裕悖论，该悖论假设训练上下文中丰富的相关信息会降低将该信息参数化编码的激励，从而增加对上下文的依赖。在使用长文档进行预训练时，增大上下文窗口对语言建模、自然语言理解和闭卷多项选择题问答（MCQA）的提升只到某个中间最优点，此后性能持续下降。在监督微调中，更多与任务相关的训练时上下文在有支持性上下文时能提高性能，但在测试时上下文缺失或具有误导性时会降低稳健性。我们的分析表明，当更长上下文提供更低复杂度的解决方案时，就会出现这种行为。从机制上看，使用信息丰富的上下文进行训练会将梯度压力从通常与参数化知识相关的前馈网络转移到注意力模块，并且因果干预表明，这种转移会在推理过程中增加对上下文的依赖。总体而言，这些发现支持信息丰裕悖论，并表明即使高质量长上下文数据很丰富，向近乎无限上下文扩展也不仅仅是提供更多数据的问题。
cs.CL / 33 / 2608.12253
One Frozen Simulator Is Not Enough: Simulator Collapse in Multi-Agent RL
单一冻结模拟器是不够的：多智能体强化学习中的模拟器坍缩
large language model
大语言模型相关
Abstract
Multi-agent reinforcement learning for human-AI interaction typically relies on a single large language model to simulate user behavior. We show that this approach systematically fails to generalize, and trace the failure to simulator collapse: because the simulator LLM is mode-collapsed, an LLM policy trained against it overfits to narrow strategies that exploit the simulator's dominant mode, and such a policy transfers poorly to unseen simulators and real users. We formalize this collapse theoretically and propose two complementary solutions, one at inference time and one at training time. The inference-time solution, Verbalized Sampling, broadens the simulator's behavior by sampling from a verbalized response distribution, reducing mode collapse. The training-time solution, Co-Training, jointly optimizes the policy against a population of trainable simulators, preventing it from overfitting to any single simulator's mode. We validate both solutions on three multi-turn benchmarks: Persuasion for Good, $τ^2$-bench, and CooperBench. Verbalized Sampling improves held-out success by up to 9% over single-simulator RL, and Co-Training pushes gains further to 14%; the human study shows similar gain on real users. Both solutions preserve the policy diversity that collapses under single-simulator RL. To support further work in this direction, we release SCOPE, an open-source framework for Population Co-Training multi-agent RL. More broadly, our results suggest that the diversity of the training environment, not only the policy, is critical to the generalization of multi-turn RL to real-world deployment.
Chinese Translation
面向人机交互的多智能体强化学习通常依赖单一大型语言模型来模拟用户行为。我们表明，这种方法系统性地无法泛化，并将失败归因于模拟器坍缩：由于模拟器LLM发生了模式坍缩，针对它训练的LLM策略会过拟合到利用模拟器主导模式的狭隘策略，而这种策略迁移到未见过的模拟器和真实用户时表现很差。我们从理论上形式化了这一坍缩，并提出两种互补的解决方案：一种在推理时，一种在训练时。推理时解决方案 Verbalized Sampling（言语化采样）通过从言语化响应分布中采样来拓宽模拟器的行为，从而减少模式坍缩。训练时解决方案 Co-Training（协同训练）针对一组可训练模拟器联合优化策略，防止策略过拟合到任何单一模拟器的模式。我们在三个多轮基准上验证了这两种解决方案：Persuasion for Good、$τ^2$-bench 和 CooperBench。Verbalized Sampling 将留出成功率相比单一模拟器RL提高了最多9%，而 Co-Training 进一步将提升幅度提高到14%；人类研究在真实用户上显示出类似的提升。两种解决方案都保留了在单一模拟器RL下会坍缩的策略多样性。为支持这一方向的进一步工作，我们发布了 SCOPE，一个用于群体协同训练多智能体RL的开源框架。更广泛地说，我们的结果表明，训练环境的多样性（而不仅仅是策略的多样性）对于多轮RL泛化到现实世界部署至关重要。
cs.CR / 34 / 2608.11348
An Empirical Study of Output-to-Input Loops for Black-Box Backdoor Detection in Fine-Tuned Open-Weight LLMs
面向微调开放权重LLM的黑盒后门检测的输出到输入循环实证研究
large language model
大语言模型相关
Abstract
Anyone can upload a fine-tuned large language model (LLM) to a public repository and claim it is safe. A backdoored model behaves normally on ordinary inputs until a hidden trigger fires, and a user with no training data, clean reference weights, or the trigger phrase has no clear way to check the model before using it. We introduce and empirically evaluate self-feeding, a black-box test method that feeds a model's own output back as its next input, so the text drifts away from the starting prompt and toward the data the model was fine-tuned on. We test self-feeding against a repeated same-prompt baseline on six open-weight LLMs (3B-15B parameters), each fine-tuned with backdoors spanning eleven attack categories, using twenty ordinary starting prompts and chains of up to ten steps. Self-feeding finds backdoors in five of six models at 92.0\% pooled precision, while the same-prompt baseline succeeds on only one of 120 prompt-model pairs; chains that begin with a joke request, an arithmetic question, or a coffee recipe all reach a trigger within a few steps. Recall per prompt is low (19.2\%), and we show why it still adds up to much higher detection at the model level once several starting prompts are used. We also report where the method falls short: one model was never triggered, and self-feeding produced two false positives that the same-prompt baseline cannot produce. Cutting the chains to four steps keeps every model-level detection at 100\% precision while using 60\% fewer queries. Needing only text-level query access and a way to recognize malicious output, self-feeding offers a cheap first check on a downloaded model.
Chinese Translation
任何人都可以将微调后的大型语言模型（LLM）上传到公共仓库，并声称它是安全的。被植入后门的模型在普通输入上表现正常，直到隐藏触发器被激活；而一个没有训练数据、干净参考权重或触发短语的用户，在使用该模型之前没有明确的方法来检查它。我们引入并实证评估了“自馈”（self-feeding），这是一种黑盒测试方法：将模型自身的输出作为其下一个输入反馈回去，使文本从起始提示偏移，并趋向模型微调时所使用的数据。我们在六个开放权重LLM（参数规模为3B-15B）上，将自馈方法与重复相同提示的基线方法进行对比测试；每个模型都经过微调，带有涵盖十一种攻击类别的后门，并使用二十个普通起始提示以及最长十步的链条。自馈在六个模型中的五个上以92.0\%的汇总精确率发现了后门，而相同提示基线仅在120个提示-模型对中的一个上成功；以笑话请求、算术问题或咖啡食谱开头的链条都能在几步内到达触发器。每个提示的召回率较低（19.2\%），我们说明了为什么一旦使用多个起始提示，它仍然能在模型层面累积成高得多的检测率。我们还报告了该方法的不足之处：一个模型从未被触发，并且自馈产生了两个相同提示基线无法产生的假阳性。将链条缩短到四步，在保持每个模型层面检测精确率为100\%的同时，查询次数减少了60\%。由于只需要文本级别的查询访问权限和识别恶意输出的方法，自馈为下载的模型提供了一种低成本的首轮检查。
cs.CR / 35 / 2608.11732
Fingerprinting Text-to-Image Diffusion Models via Collapsed Generation
通过塌缩生成对文本到图像扩散模型进行指纹识别
diffusion
扩散模型相关
Abstract
Proprietary text-to-image diffusion models are increasingly distributed as hosted services and downloadable checkpoints, making their intellectual property (IP) protection an increasingly critical concern when model leakage, copying, or unauthorized fine-tuning is disputed. In this work, we present a non-invasive model fingerprinting framework based on \emph{collapsed generation}, a phenomenon where certain input conditions produce highly consistent images across multiple stochastic seeds. We show that collapsed generation is an intrinsic, model-dependent property of the learned generation process. These collapse-prone conditions therefore expose model-specific behavioral signatures, enabling reliable ownership verification without embedding invasive watermarks. After preparing conditions on the source model, the framework verifies a suspect model under two access settings: (1) white-box pipeline access, where optimized continuous embeddings can be injected into the generation process, and (2) black-box API-only access, where natural language prompts are queried through the service interface. In both cases, ownership evidence is measured by whether the suspect model reproduces the source model's collapse behavior across stochastic samplings. Extensive experiments across UNet- and transformer-based diffusion models show that collapsed generation fingerprints can distinguish different source models with low confusion. These fingerprints remain verifiable in fine-tuned derivatives and under common and adaptive model- or query-level obfuscations, while requiring only a modest verification query budget. Together, these results establish collapsed generation as a reliable intrinsic evidence source for non-invasive diffusion model ownership verification.
Chinese Translation
专有的文本到图像扩散模型越来越多地以托管服务和可下载检查点的形式分发，这使得当发生模型泄露、复制或未经授权的微调争议时，其知识产权（IP）保护成为一个日益关键的问题。在本文中，我们提出了一个基于\emph{塌缩生成}的非侵入式模型指纹识别框架；塌缩生成是指某些输入条件在多个随机种子下产生高度一致的图像的现象。我们表明，塌缩生成是所学生成过程的一种内在的、依赖于模型的属性。因此，这些易发生塌缩的条件会暴露出模型特有的行为签名，从而能够在不嵌入侵入性水印的情况下进行可靠的所有权验证。在源模型上准备好条件后，该框架在两种访问设置下验证可疑模型：（1）白盒流水线访问，其中优化后的连续嵌入可以被注入生成过程；（2）仅通过 API 的黑盒访问，其中自然语言提示通过服务接口进行查询。在这两种情况下，所有权证据都通过可疑模型是否在多次随机采样中重现源模型的塌缩行为来衡量。在基于 UNet 和基于 Transformer 的扩散模型上进行的大量实验表明，塌缩生成指纹能够以较低的混淆度区分不同的源模型。这些指纹在微调衍生模型中以及常见及自适应的模型级或查询级混淆下仍可验证，同时只需要适度的验证查询预算。综上所述，这些结果确立了塌缩生成作为非侵入式扩散模型所有权验证的可靠内在证据来源。
cs.CR / 36 / 2608.11816
How China-Origin Vision-Language Models Move from Refusal to Reframing in State Alignment
中国起源的视觉语言模型如何在国家对齐中从拒绝走向重构
large language model
大语言模型相关
Abstract
State-aligned distortion has been documented in China-origin text-based large language models (LLMs), but whether, and in what form, it arises in multimodal systems has not been systematically examined. We construct a balanced benchmark of 200 core entries spanning ten politically sensitive topics, plus a seven-variant visual-abstraction probe, and run nine vision-language models (VLMs), seven China-origin and two non-China, across four elicitation paradigms and two prompt languages, yielding 21,708 trials. Each response is audited on six dimensions -- explicit refusal, information integrity, visual grounding, state-aligned framing, language consistency, and response length -- by two independent frontier LLM judges, validated against three human experts on a 200-trial sample. Measuring each dimension separately lets us decompose multimodal censorship into individual signals rather than a single refusal-based score; in particular, refusal and framing are measured independently, so a model can stop refusing while still reframing. We find that (i) Chinese-language prompting roughly triples the odds of state-aligned framing, within every model; (ii) China-origin models reframe more than non-China models (direction robust across judges and human raters; magnitude 1.6--3.2x); (iii) the effect is strongest in text-only political commentary (36.5%) and is gated by recognition of the depicted subject rather than pixel detail, persisting even at silhouette for iconic images; and (iv) across four Qwen generations, state-aligned framing rises while explicit refusal falls: censorship migrates from a visible act (refusal) to an invisible one (fluent reframing). We argue this shift to invisible reframing is fundamentally a problem of human-AI interaction: it removes the very signal users rely on to recognize that information has been withheld.
Chinese Translation
国家对齐的失真已在中国起源的基于文本的大语言模型（LLMs）中得到记录，但它是否以及以何种形式出现在多模态系统中尚未被系统性地考察。我们构建了一个平衡的基准，包含200个核心条目，涵盖十个政治敏感话题，外加一个七变体视觉抽象探针，并在四种引导范式和两种提示语言下运行九个视觉语言模型（VLMs），其中七个为中国起源，两个为非中国起源，共产生21,708次试验。每个回答都由两位独立的顶尖LLM裁判在六个维度上进行审核——明确拒绝、信息完整性、视觉依据、国家对齐框架、语言一致性和回答长度——并在200次试验样本上与三位人类专家进行验证。对每个维度分别测量，使我们能够将多模态审查分解为单独的信号，而不是单一的基于拒绝的分数；特别是，拒绝和框架是独立测量的，因此模型可以在停止拒绝的同时仍然进行重构。我们发现：（i）中文提示在每个模型内大致使国家对齐框架的几率增加三倍；（ii）中国起源模型比非中国模型更多地重构（方向在裁判和人类评分者之间具有稳健性；幅度为1.6--3.2倍）；（iii）该效应在纯文本政治评论中最强（36.5%），并且受对描绘对象的识别而非像素细节的门控影响，即使对于标志性图像在剪影下也持续存在；以及（iv）在四代Qwen模型中，国家对齐框架上升而明确拒绝下降：审查从一种可见的行为（拒绝）迁移到一种不可见的行为（流畅的重构）。我们认为，这种向不可见重构的转变本质上是一个人机交互问题：它移除了用户赖以识别信息已被隐瞒的那个信号。
cs.CR / 37 / 2608.11878
ToolHazard: Scaling Adversarial Environments for Security Evaluation and Alignment of LLM-based Agents
ToolHazard：扩展对抗性环境以用于基于LLM的智能体的安全评估与对齐
large language model
大语言模型相关
Abstract
Large language model (LLM) agents integrated with external tools are vulnerable to indirect prompt injections embedded in environmental states. However, existing studies largely rely on manually implemented or reused environments, stochastic LLM-based tool simulation, and predefined injection locations, limiting scalable security research across broader domains. To bridge this gap, we propose **ToolHazard**, a scalable adversarial environment synthesis framework that reduces human engineering and supports expansion with additional seed domains and compute. Through an Environment Simulator, an Attacker Agent, and a User Simulator, ToolHazard synthesizes executable stateful environments, discovers viable injection points and generates environment-specific payloads, and constructs state-grounded long-horizon tasks. Based on ToolHazard, we build **ToolHazard-Bench** for stress-testing agents under complex workflows and diverse environmental attacks. Experiments reveal substantial agent vulnerabilities and show that injection timing and placement affect attack effectiveness. Moreover, ToolHazard-generated alignment data improves security on both ToolHazard-Bench and AgentDojo while preserving benign task utility.
Chinese Translation
集成了外部工具的大语言模型（LLM）智能体容易受到嵌入在环境状态中的间接提示注入攻击。然而，现有研究大多依赖人工实现或复用的环境、基于LLM的随机工具模拟以及预定义的注入位置，这限制了跨更广泛领域的可扩展安全研究。为弥合这一差距，我们提出了 **ToolHazard**，一个可扩展的对抗性环境合成框架，该框架减少了人工工程，并支持通过增加种子领域和计算资源进行扩展。通过环境模拟器、攻击者智能体和用户模拟器，ToolHazard 合成可执行的有状态环境，发现可行的注入点并生成特定于环境的载荷，并构建基于状态的长时程任务。基于 ToolHazard，我们构建了 **ToolHazard-Bench**，用于在复杂工作流和多样化环境攻击下对智能体进行压力测试。实验揭示了智能体存在显著漏洞，并表明注入的时机和位置会影响攻击效果。此外，ToolHazard 生成的对齐数据在 ToolHazard-Bench 和 AgentDojo 上均提升了安全性，同时保持了良性任务效用。
cs.AI / 38 / 2608.11537
Generative Semantic Segmentation via an Observable Semantic-Image Interface and Hierarchical Generator Evidence Alignment
通过可观察的语义-图像接口与分层生成器证据对齐的生成式语义分割
diffusion
扩散模型相关
Abstract
Generative semantic segmentation exposes structured predictions as images, but direct color decoding is susceptible to color drift and boundary mixing, whereas latent-feature decoders that predict a separate output distribution may relegate the rendered image to an intermediate visualization. We present Semantic Prism, a conditional semantic-image generation-and-refinement framework with deterministic inference. A diffusion-distilled one-step generator renders a semantic RGB image; per-pixel distances from the rendered colors to a fixed class-color codebook define an explicit probabilistic interface. Hierarchical Generator Evidence Alignment spatially aligns multi-level generator features and uses a zero-initialized output projection to predict an additive residual in the interface logit space, retaining the image-defined interface as the reference for the final distribution. The interface and refined distributions further enable Contextual Interface--Hierarchy Disagreement (C-IHD), a fixed readout for ranking remaining pixel errors without an auxiliary predictor or additional forward pass. On the 500-image Cityscapes validation set, Semantic Prism achieves 72.07% mean intersection over union, 11.39 mIoU points above direct-interface decoding, with 0.41% expected calibration error. Matched-capacity ablations over three seeds support the benefit of jointly aligned multi-level evidence. A separately trained model attains 62.22% mIoU on BDD100K, while the Cityscapes-trained model reaches 46.89\% mIoU under source-frozen transfer to the Adverse Conditions Dataset with Correspondences, without target-domain adaptation. Across all three datasets, C-IHD consistently improves the area under the precision--recall curve for pixel-error ranking over maximum softmax probability on the same segmentation predictions; on ACDC, it raises AUPR from 0.6580 to 0.7557.
Chinese Translation
生成式语义分割将结构化预测呈现为图像，但直接颜色解码容易受到颜色漂移和边界混合的影响，而预测单独输出分布的潜在特征解码器可能将渲染图像降格为中间可视化。我们提出 Semantic Prism，一种具有确定性推断的条件式语义图像生成与细化框架。一个经过扩散蒸馏的单步生成器渲染语义 RGB 图像；渲染颜色到固定类别颜色码本的逐像素距离定义了一个显式的概率接口。分层生成器证据对齐在空间上对齐多层级生成器特征，并使用零初始化输出投影在接口 logit 空间中预测加性残差，同时保留图像定义的接口作为最终分布的参考。接口和细化后的分布进一步实现了上下文接口-层级不一致性（C-IHD），这是一种用于对剩余像素错误进行排序的固定读出方法，无需辅助预测器或额外的前向传播。在包含 500 张图像的 Cityscapes 验证集上，Semantic Prism 实现了 72.07% 的平均交并比，比直接接口解码高出 11.39 个 mIoU 点，预期校准误差为 0.41%。在三个随机种子上进行的等容量消融实验支持联合对齐多层级证据的益处。一个单独训练的模型在 BDD100K 上达到 62.22% 的 mIoU，而 Cityscapes 训练的模型在源域冻结迁移到 Adverse Conditions Dataset with Correspondences 时达到 46.89% 的 mIoU，且未进行目标域自适应。在所有三个数据集上，与相同分割预测上的最大 softmax 概率相比，C-IHD 一致地提高了像素错误排序的精确率-召回率曲线下面积；在 ACDC 上，它将 AUPR 从 0.6580 提升到 0.7577。
cs.AI / 39 / 2608.11562
From Synthesis to Removal: Physics-Grounded Reflection Simulation and Diffusion-Based Video Dereflection
从合成到去除：基于物理的反射模拟与基于扩散的视频去反射
diffusion
扩散模型相关
Abstract
Videos captured through glass often contain reflections that degrade visual quality and interfere with downstream vision tasks. Although single-image reflection removal has been extensively studied, video reflection removal remains largely underexplored due to the lack of paired video data, temporally coherent removal models, and dedicated evaluation benchmarks. We present a closed-loop framework that unifies physics-grounded reflection simulation, diffusion-based video dereflection, and benchmark evaluation. Our S2R-Synthesis pipeline generates paired reflected and reflection-free videos by performing physics-grounded augmentation in the structure space and rendering realistic reflected videos with a trained video diffusion renderer; the augmentation models key glass-related effects including roughness-induced blur, thickness-induced ghosting, and reflectance variation. Based on the synthesized data, we introduce S2R-Removal, the first diffusion-based video reflection removal model, which adapts a pretrained video diffusion prior through reflection-aware latent adaptation and one-step pixel-geometric refinement, recovering the clean transmission in a single denoising step. We further build S2R-Bench, the first benchmark for video reflection removal, supporting both full-reference evaluation and real-world human perceptual assessment. Experiments on S2R-Bench and multiple public image benchmarks demonstrate state-of-the-art performance and faster inference than even non-diffusion baselines, and validate the effectiveness of S2R-Synthesis. Project page: https://codingwzp.github.io/VideoDereflection_S2R.
Chinese Translation
通过玻璃拍摄的视频通常包含反射，这会降低视觉质量并干扰下游视觉任务。尽管单图像反射去除已被广泛研究，但由于缺乏成对视频数据、时间一致的去除模型和专用评估基准，视频反射去除在很大程度上仍未得到充分探索。我们提出一个闭环框架，将基于物理的反射模拟、基于扩散的视频去反射和基准评估统一起来。我们的 S2R-Synthesis 流程通过在结构空间中进行基于物理的增强，并使用训练好的视频扩散渲染器渲染逼真的反射视频，生成成对的反射视频和无反射视频；该增强对关键的玻璃相关效应进行建模，包括粗糙度引起的模糊、厚度引起的重影以及反射率变化。基于合成数据，我们引入了 S2R-Removal，这是第一个基于扩散的视频反射去除模型，它通过反射感知的潜在自适应和单步像素几何细化来适配预训练的视频扩散先验，并在单个去噪步骤中恢复干净的透射图像。我们进一步构建了 S2R-Bench，这是第一个用于视频反射去除的基准，支持全参考评估和真实世界人类感知评估。在 S2R-Bench 和多个公共图像基准上的实验表明，其性能达到最先进水平，并且推理速度甚至比非扩散基线更快，同时验证了 S2R-Synthesis 的有效性。项目页面：https://codingwzp.github.io/VideoDereflection_S2R。
cs.LG / 40 / 2608.11643
Robustness of AI-Art Detectors under Generator Shift
生成器偏移下 AI 艺术检测器的鲁棒性
diffusion
扩散模型相关
Abstract
Text-to-image generative models have advanced rapidly, with modern Diffusion Transformer architectures producing images that are increasingly difficult to distinguish from human-created artwork. This development has raised significant concerns regarding copyright protection, misinformation, fraud, impersonation, and the authenticity of digital content. Most AI-art detectors are trained and evaluated on the same generator family, leaving robustness to newer architectures underexplored. In this chapter, we analyze generator shift based on a Stable Diffusion 3.5 Medium (SD3.5m) artwork dataset spanning ten art styles through reverse prompting of held-out human artwork samples. Five detectors are trained on U-Net-based latent diffusion artwork and evaluated in a zero-shot cross-generator setting on the SD3.5m dataset. Deep learning models perform strongly in-distribution but degrade under generator shift, misclassifying many SD3.5m images as human while human false positives remain low. The CLIP ViT-L/14 model performs best overall, while Grad-CAM analysis reveals weaker and more diffuse activation on false negatives. These findings highlight a generalization gap in current AI-art detectors and motivate the development of detectors as one component of a layered defense that remains reliable across rapidly evolving generative architectures.
Chinese Translation
文本到图像生成模型发展迅速，现代 Diffusion Transformer 架构生成的图像越来越难以与人类创作的艺术作品区分。这一发展引发了人们对版权保护、错误信息、欺诈、冒充以及数字内容真实性的重大担忧。大多数 AI 艺术检测器在同一生成器家族上进行训练和评估，导致其对新架构的鲁棒性尚未得到充分探索。在本章中，我们基于一个 Stable Diffusion 3.5 Medium（SD3.5m）艺术作品数据集来分析生成器偏移，该数据集通过对留出的人类艺术作品样本进行反向提示生成，涵盖十种艺术风格。五个检测器在基于 U-Net 的潜在扩散艺术作品上进行训练，并在 SD3.5m 数据集上以零样本跨生成器设置进行评估。深度学习模型在分布内表现强劲，但在生成器偏移下性能下降，将许多 SD3.5m 图像错误分类为人类作品，而人类作品的假阳性率仍然较低。CLIP ViT-L/14 模型总体表现最佳，而 Grad-CAM 分析显示，在假阴性样本上的激活更弱且更分散。这些发现凸显了当前 AI 艺术检测器在泛化方面的差距，并推动将检测器开发为分层防御的一个组成部分，使其在快速演进的生成架构中保持可靠。
cs.AI / 41 / 2608.11655
Motion-as-Prompt: Enhancing Motion Reasoning in Multimodal Large Language Models via Motion-Guided Cross-Frame Visual Prompting
Motion-as-Prompt：通过运动引导的跨帧视觉提示增强多模态大语言模型中的运动推理
large language model
大语言模型相关
Abstract
Motion-centric video reasoning is fundamental to interactive applications such as robotic manipulation and autonomous navigation. However, multimodal large language models (MLLMs) typically process videos through sparse uniform sampling to control visual-token and attention costs. This strategy may discard critical transitions between sampled frames, limiting reasoning about object movement, collisions, and causal interactions. To mitigate this issue, we propose Motion-as-Prompt (MaP), a track-guided cross-frame visual prompting framework. MaP recovers dense point trajectories, selects motion-informative frames, and marks the trajectories accumulated between consecutive sampled frames directly onto the visual inputs, making otherwise hidden displacement, direction changes, and interactions observable to frozen MLLMs. Experiments on CLEVRER and Something-Something-v2 show that MaP consistently improves average motion-reasoning accuracy, yielding gains of 4.2% and 8.9% for GPT-5.5, respectively. Notably, these improvements are obtained without degrading non-motion understanding, highlighting the robustness of MaP. These results demonstrate that MaP provides a simple and effective solution for enhancing motion-centric video reasoning without model training or architectural modification. Project page:https://github.com/SunVictor23/MaP.
Chinese Translation
以运动为中心的视频推理对于机器人操作和自主导航等交互式应用至关重要。然而，多模态大语言模型（MLLMs）通常通过稀疏均匀采样来处理视频，以控制视觉 token 和注意力成本。这种策略可能丢弃采样帧之间的关键变化，从而限制对物体运动、碰撞和因果交互的推理。为了缓解这一问题，我们提出了 Motion-as-Prompt（MaP），一个轨迹引导的跨帧视觉提示框架。MaP 恢复稠密点轨迹，选择运动信息丰富的帧，并将连续采样帧之间累积的轨迹直接标记到视觉输入上，使得原本隐藏的位移、方向变化和交互对冻结的 MLLMs 变得可观测。在 CLEVRER 和 Something-Something-v2 上的实验表明，MaP 持续提升平均运动推理准确率，分别使 GPT-5.5 获得 4.2% 和 8.9% 的提升。值得注意的是，这些提升是在不降低非运动理解的情况下取得的，突显了 MaP 的鲁棒性。这些结果表明，MaP 为增强以运动为中心的视频推理提供了一种简单而有效的解决方案，且无需模型训练或架构修改。项目页面：https://github.com/SunVictor23/MaP。
cs.AI / 42 / 2608.11738
Advancing MLLM-based UAV Image Understanding and Reasoning: A Benchmark and a Training-Free Multi-Agent System
推进基于 MLLM 的无人机图像理解与推理：一个基准与一个免训练多智能体系统
large language model
大语言模型相关
Abstract
Multimodal Large Language Model (MLLM)-based UAV aerial image understanding and reasoning is essential for aerial intelligence yet poses distinct challenges arising from extreme scale variation, arbitrary camera orientations, and high object density. Despite growing interest, existing evaluations remain fragmented across individual datasets and narrow tasks, leaving a critical gap in unified assessment of UAV understanding and reasoning capabilities. To fill this gap, we construct UAVQA-Bench, a benchmark of 1,500 human-annotated QA pairs drawn from 13 public UAV datasets, covering 6 capability dimensions and 16 tasks in both multiple-choice and visual grounding formats. Systematic evaluation of a broad range of open-source and closed-source MLLMs as well as agent-based systems on UAVQA-Bench identifies three key failure modes: domain-toolset mismatch, unchecked error propagation, and static reasoning. Motivated by these findings, we propose UAV-MAS, a training-free multi-agent system for MLLM-based UAV aerial image understanding and reasoning, comprising a Domain-Specific Perception Engine (DSPE) that routes queries to task-appropriate visual tools, a Context-Aware Iterative Refinement module (CAIR) that validates intermediate reasoning to curb error accumulation, and a Difficulty-Aware Adaptive Search mechanism (DAAS) that adjusts search depth to question difficulty. UAV-MAS with a 32B open-source MLLM achieves 77.0% overall accuracy on UAVQA-Bench, surpassing Gemini 3 Pro by 4.0\%, while the 8B variant improves 8.7\% over its base model.
Chinese Translation
基于多模态大语言模型（MLLM）的无人机航拍图像理解与推理对于空中智能至关重要，但也带来了由极端尺度变化、任意相机朝向和高目标密度引起的独特挑战。尽管相关研究兴趣日益增长，现有评估仍分散于各个单独的数据集和狭窄任务中，在无人机理解与推理能力的统一评估方面留下了关键空白。为了填补这一空白，我们构建了 UAVQA-Bench，一个由来自 13 个公开无人机数据集的 1,500 个人工标注问答对组成的基准，涵盖了 6 个能力维度和 16 个任务，并同时采用多项选择和视觉定位两种格式。在 UAVQA-Bench 上对广泛的开源和闭源 MLLM 以及基于智能体的系统进行系统评估后，我们识别出三种关键失败模式：领域工具集不匹配、未受抑制的错误传播以及静态推理。受这些发现启发，我们提出了 UAV-MAS，一个无需训练的、用于基于 MLLM 的无人机航拍图像理解与推理的多智能体系统，该系统包括：领域特定感知引擎（DSPE），负责将查询路由到适合任务的视觉工具；上下文感知迭代细化模块（CAIR），用于验证中间推理以抑制错误累积；以及难度感知自适应搜索机制（DAAS），可根据问题难度调整搜索深度。搭载 32B 开源 MLLM 的 UAV-MAS 在 UAVQA-Bench 上实现了 77.0% 的整体准确率，比 Gemini 3 Pro 高出 4.0%，而 8B 变体相较于其基础模型提升了 8.7%。
cs.AI / 43 / 2608.11741
JieZi: A Large-Scale Expert-Audited Dataset and Benchmark for Ancient Chinese Character Exegesis
JieZi：面向古代汉字训诂的大规模专家审核数据集与基准
large language model
大语言模型相关
Abstract
The scholarly exegesis of ancient Chinese characters demands integrating visual observation, linguistic analysis, and historical context. However, existing computational approaches focus narrowly on subtasks such as character recognition and retrieval, lacking the structured datasets and benchmarks required for comprehensive scholarly analysis. To address this limitation, we introduce Ancient Chinese Character Exegesis (ACCE), a vision-language question answering (VQA) task that models the scholarly exegesis process. ACCE is organized into four progressive levels: basic character identification, glyph-form analysis, meaning exegesis, and diachronic evolution analysis. To support this task, we construct two complementary resources. JieZi-Dataset is the first large-scale, expert-audited VQA training dataset for ACCE, comprising over 500K QA pairs. It is constructed via a pipeline that reduces factual errors by constraining generation with expert-designed templates and source-text references. Human verification is further applied at each key stage to ensure scholarly accuracy. JieZi-Bench is an evaluation benchmark aligned with the exegesis process, constructed and verified by human experts to ensure evaluation reliability. It consists of four levels with reference answers curated from authoritative lexicographic works held separate from the training data. Experiments on multimodal large language models show that current models perform well on basic identification but struggle with glyph analysis, semantic reasoning, and diachronic understanding. Fine-tuning on JieZi-Dataset substantially improves performance across all four levels. Code and dataset are available at https://github.com/Ran00w/JieZi.
Chinese Translation
古代汉字的学术训诂需要整合视觉观察、语言分析和历史语境。然而，现有的计算方法仅狭隘地关注字符识别和检索等子任务，缺乏全面学术分析所需的结构化数据集和基准。为解决这一局限，我们提出了古代汉字训诂（ACCE），这是一项视觉-语言问答（VQA）任务，对学术训诂过程进行建模。ACCE 分为四个递进层次：基础字符识别、字形分析、意义训诂和历时演变分析。为支持该任务，我们构建了两种互补资源。JieZi-Dataset 是首个面向 ACCE 的大规模、经专家审核的 VQA 训练数据集，包含超过 50 万个问答对。该数据集通过一条流水线构建，该流水线利用专家设计的模板和源文本参考文献约束生成过程，以减少事实错误；并在每个关键阶段进一步进行人工验证，以确保学术准确性。JieZi-Bench 是一个与训诂过程对齐的评估基准，由人类专家构建并验证，以确保评估可靠性。它由四个层次组成，参考答案选自权威辞书著作，并与训练数据保持分离。在多模态大语言模型上的实验表明，当前模型在基础识别上表现良好，但在字形分析、语义推理和历时理解方面存在困难。在 JieZi-Dataset 上进行微调可显著提升四个层次上的性能。代码和数据集可在 https://github.com/Ran00w/JieZi 获取。
cs.AI / 44 / 2608.12032
LoSA: Near-Lossless Sparse Attention for Training-Free Video Diffusion Acceleration
LoSA：用于免训练视频扩散加速的近无损稀疏注意力
diffusion
扩散模型相关
Abstract
Video diffusion transformers are costly to sample: every denoising step applies self-attention over a long 3D token sequence, a quadratic cost that dominates as resolution and duration grow. Sparse attention reduces this cost without retraining, but existing methods pursue aggressive sparsity, where further speedup costs disproportionately more attention fidelity. We target the opposite end of this trade-off: fix near-lossless fidelity by construction, and remove as much computation as this constraint permits. Two observations make this regime practical: roughly 40% of block interactions can be removed while retaining 99% of the attention mass, and the high-mass support remains stable across denoising steps. We propose LoSA, a training-free sparse-attention method that fixes a retained-mass threshold of 99% rather than a sparsity ratio: it measures exact block attention masses at one early dense step, keeps, for each head and query block, the smallest key/value block set meeting the threshold, and reuses the frozen block indices for all remaining steps. On Wan2.1-1.3B, LoSA alone gives a $1.36\times$ speedup with a 0.06-point VBench Overall drop. The benefit is largest under composition: combined with feature caching, LoSA reaches a $3.2\times$ speedup on HunyuanVideo at a 0.02-point drop, versus 0.32 points for the strongest sparse baseline at comparable speed. Across three video diffusion transformers and speedups up to $3.2\times$, LoSA consistently achieves the best training-free speed-quality trade-off.
Chinese Translation
视频扩散 Transformer 的采样成本很高：每个去噪步骤都会在长 3D token 序列上应用自注意力，随着分辨率和时长增加，这一二次方代价会占据主导。稀疏注意力无需重新训练即可降低这一成本，但现有方法追求激进稀疏性，此时进一步加速会不成比例地损失更多注意力保真度。我们瞄准这一权衡的另一端：通过构造方式固定近无损保真度，并在该约束允许的范围内尽可能多地移除计算量。两个观察使这一方案切实可行：在保留 99% 注意力质量的同时，大约 40% 的块交互可以被移除；并且高注意力质量的支撑集跨去噪步骤保持稳定。我们提出 LoSA，一种免训练的稀疏注意力方法，该方法固定 99% 的保留注意力质量阈值而非稀疏率：它在早期某个稠密步骤中测量精确的块注意力质量；针对每个头和查询块，保留满足该阈值的最小键/值块集合；并在所有剩余步骤中复用冻结的块索引。在 Wan2.1-1.3B 上，仅 LoSA 即可带来 $1.36\times$ 的加速，而 VBench Overall 仅下降 0.06 分。其优势在组合使用时最为明显：与特征缓存相结合，LoSA 在 HunyuanVideo 上达到 $3.2\times$ 的加速，且仅下降 0.02 分；而在相近速度下，最强的稀疏基线下降 0.32 分。在三个视频扩散 Transformer 上，在高达 $3.2\times$ 的加速条件下，LoSA 始终取得最佳的免训练速度-质量权衡。
cs.AI / 45 / 2608.12262
Diagram-MMU: A Multi-Modal Benchmark for Scientific Diagrams
Diagram-MMU：科学图表的多模态基准
large language model
大语言模型相关
Abstract
Multimodal Large Language Models (MLLMs) have been growing the capability for scientific writing and collaboration. For example, OpenAI Prism is a free workspace for scientific writing and collaboration. One important feature in Prism is turning scientific diagrams directly into LaTeX TikZ code. In this paper, we build a benchmark, Diagram-MMU, a multi-modal benchmark designed to assess MLLMs' ability for scientific diagram parsing and understanding. Diagram-MMU features 3.7k curated diagrams and 18.3k human-validated questions across six domains. It evaluates MLLMs on three tasks common in vibe writing workspaces: diagram-to-code parsing, diagram-to-code editing, and diagram question answering, alongside agentic settings per task. The evaluation of 12 MLLMs reveals that diagram-to-code tasks are more challenging than diagram question answering: models can reason well over diagrams but struggle to parse and edit them, underscoring the need for methods to enhance MLLMs' capability in diagram-to-code generation. Under agentic settings, most models improve parsing and editing performance but degrade on question answering, while Claude-4.6 Opus consistently improves across all three tasks. Project Page: https://vi-ocean.github.io/projects/diagram-mmu.
Chinese Translation
多模态大语言模型（MLLMs）在科学写作与协作方面的能力不断增强。例如，OpenAI Prism 是一个用于科学写作与协作的免费工作空间。Prism 的一个重要功能是将科学图表直接转换为 LaTeX TikZ 代码。在本文中，我们构建了一个基准 Diagram-MMU，这是一个多模态基准，旨在评估 MLLMs 在科学图表解析与理解方面的能力。Diagram-MMU 包含 3.7k 个精选图表和 18.3k 个人工验证的问题，涵盖六个领域。它在 vibe 写作工作空间中常见的三个任务上评估 MLLMs：图表到代码解析、图表到代码编辑和图表问答，以及每个任务的智能体设置。对 12 个 MLLMs 的评估表明，图表到代码任务比图表问答更具挑战性：模型能够很好地对图表进行推理，但在解析和编辑它们方面存在困难，这凸显了需要方法来增强 MLLMs 在图表到代码生成方面的能力。在智能体设置下，大多数模型的解析和编辑性能有所提升，但在问答方面有所下降，而 Claude-4.6 Opus 在所有三项任务上均持续改进。项目页面：https://vi-ocean.github.io/projects/diagram-mmu。
cs.AI / 46 / 2608.12290
Beyond Trial-and-Error: Agentic Optimization for Image-to-Video Adherence
超越试错：面向图像到视频一致性的智能体优化
large language model
大语言模型相关
Abstract
Modern black-box Image-to-Video (I2V) models offer powerful capabilities in automated content creation, yet their lack of fine-grained control and reliability presents significant challenges in professional workflows. Their inherent stochasticity causes minor variations in textual prompts or hyperparameters to yield drastically different outputs often necessitating inefficient, brute-force trial-and-error processes. To address these limitations, we introduce the ``Agentic Self-Improvement" framework, which reframes video synthesis into a closed-loop, goal-directed optimization. Our framework systematically navigates the generation parameter space using a novel two-stage approach. In the first stage, an iterative prompt optimization loop uses a multimodal Large Language Model (mLLM) to refine the input prompt. This refinement implements two automated evaluations: Davidsonian Scene Graph (DSG) queries ensure semantic adherence, and Common Mistake Questions (CMQ) for artifact detection. At the second stage, we use Bayesian optimization to efficiently co-optimize stochastic seeds and CFG scales. This search is guided by a suite of quality metrics, including the novel Video-Text Adherence (VTA) score derived from the DSG and CMQ evaluations. Our framework significantly outperforms unguided search methods: in human preference studies, videos generated via our agentic approach were strongly preferred over baseline outputs, achieving win rates up to 69\%. This work provides a practical and extensible methodology for enhancing the predictability and control of state-of-the-art video generation models, moving the field beyond speculative curiosities toward reliable, production-ready tools.
Chinese Translation
现代黑盒图像到视频（I2V）模型在自动化内容创作中提供了强大能力，但缺乏细粒度控制和可靠性，在专业工作流中带来了重大挑战。它们固有的随机性导致文本提示或超参数的微小变化会产生截然不同的输出，往往需要低效的蛮力试错过程。为解决这些局限，我们引入了“智能体自我改进”框架，该框架将视频合成重新定义为一个闭环、目标导向的优化过程。我们的框架通过一种新颖的两阶段方法系统地探索生成参数空间。在第一阶段，一个迭代式提示优化循环使用多模态大语言模型（mLLM）来优化输入提示。这一优化过程实施两种自动化评估：Davidsonian 场景图（DSG）查询用于确保语义一致性，常见错误问题（CMQ）用于伪影检测。在第二阶段，我们使用贝叶斯优化高效地协同优化随机种子和 CFG 尺度。该搜索由一组质量指标引导，其中包括从 DSG 和 CMQ 评估中得出的新颖视频-文本一致性（VTA）分数。我们的框架显著优于无引导搜索方法：在人类偏好研究中，通过我们的智能体方法生成的视频相比基线输出获得了强烈偏好，胜率最高达69%。这项工作提供了一种实用且可扩展的方法，用于增强最先进视频生成模型的可预测性和可控性，推动该领域超越投机性新奇事物，迈向可靠、生产就绪的工具。
cs.AI / 47 / 2608.12308
DreamFly: Causal Memory and Receding-Horizon Diffusion Planning for Aerial Vision-Language Navigation
DreamFly：面向空中视觉语言导航的因果记忆与滚动时域扩散规划
diffusion
扩散模型相关
Abstract
Aerial vision-language navigation (VLN) requires an embodied agent to integrate visual evidence over time, plan future actions, and determine when it has reached a navigation goal under partial observability. Although recent VLA models offer a promising perception-to-action paradigm, adapting them to aerial navigation remains challenging due to limited historical context, short planning horizons, and unreliable implicit termination. To address these challenges, we propose DreamFly, a diffusion-based aerial VLN framework built on Dream-VLA. DreamFly introduces a causally aligned historical memory that augments the current visual representation using only observations preceding the current decision step, enabling temporal reasoning without future information leakage. We further formulate navigation as receding-horizon diffusion planning, where the policy predicts a $K$-step action chunk but executes only the first action before replanning. This plan-$K$, execute-one strategy uses future actions as auxiliary planning targets while preserving closed-loop visual feedback. Finally, LiteStop estimates the stop probability directly from action logits at the initial all-mask state, decoupling explicit termination from action generation. Experiments on the OpenFly benchmark demonstrate consistent improvements in seen and unseen environments. DreamFly achieves 32.04%/29.46% SR and 28.22%/23.54% SPL on the test-seen/test-unseen splits, respectively, outperforming all compared methods on both metrics while attaining the lowest navigation error. These results demonstrate the effectiveness of jointly modeling historical context, future action structure, and explicit termination for aerial VLN.
Chinese Translation
空中视觉语言导航（VLN）要求具身智能体在部分可观测条件下，随时间整合视觉证据、规划未来动作，并判断何时已到达导航目标。尽管最近的 VLA 模型提供了一种有前景的感知到动作范式，但由于历史上下文有限、规划时域较短以及隐式终止不可靠，将其适配到空中导航仍然具有挑战性。为应对这些挑战，我们提出了 DreamFly，一个基于 Dream-VLA 构建的、基于扩散的空中 VLN 框架。DreamFly 引入了一种因果对齐的历史记忆，该记忆仅使用当前决策步之前的观测来增强当前视觉表示，从而在避免未来信息泄漏的情况下实现时序推理。我们进一步将导航形式化为滚动时域扩散规划，其中策略预测一个 $K$ 步动作块，但在重新规划前仅执行第一个动作。这种“规划 $K$ 步、执行一步”的策略将未来动作用作辅助规划目标，同时保留闭环视觉反馈。最后，LiteStop 直接在初始全掩码状态下的动作 logits 上估计停止概率，从而将显式终止与动作生成解耦。在 OpenFly 基准上的实验表明，在已见和未见环境中均取得了一致的改进。DreamFly 在测试已见/测试未见划分上分别实现了 32.04%/29.46% 的 SR 和 28.22%/23.54% 的 SPL，在两项指标上均优于所有对比方法，同时取得了最低的导航误差。这些结果表明，联合建模历史上下文、未来动作结构和显式终止对于空中 VLN 是有效的。
cs.CL / 48 / 2608.11830
Quantifying the Relationship Between Clinical Safety and Environmental Impact in Therapeutic LLMs
量化治疗性大语言模型中临床安全性与环境影响之间的关系
large language model
大语言模型相关
Abstract
The deployment of large language models (LLMs) in mental health contexts raises questions about the relationship between clinical safety and environmental cost. In this paper, we examine this relationship by combining K-Bench clinical safety scores with EcoLogits life-cycle assessment estimates across 47 supported model configurations. We evaluate model performance and environmental impact across four dimensions: energy use, carbon emissions, water consumption, and abiotic depletion. The results indicate a non-linear trade-off at the upper end of the safety distribution: a 2.61 percentage-point increase in clinical safety score corresponded to an approximately 60-fold increase in estimated energy use per million output tokens. Row-level analyses further suggest that additional test-time compute did not consistently improve clinical safety and, in some configurations, was associated with lower clinical safety scores. These findings suggest that relying solely on larger models or additional inference-time computation may be an inefficient strategy for improving safety in therapeutic AI systems. We discuss the implications for sustainable deployment and highlight dynamic model selection, including model cascading, as a potential approach for reducing environmental impact while preserving clinical performance in higher-risk cases.
Chinese Translation
在心理健康情境中部署大语言模型（LLMs）引发了关于临床安全性与环境代价之间关系的问题。在本文中，我们通过将K-Bench临床安全评分与EcoLogits生命周期评估估计相结合，在47个受支持的模型配置上考察这一关系。我们从四个维度评估模型性能和环境影响：能源使用、碳排放、水资源消耗和非生物资源耗竭。结果表明，在安全性分布的高端存在非线性权衡：临床安全评分提高2.61个百分点，对应的每百万输出token的估计能源使用量约增加60倍。行级分析进一步表明，额外的测试时计算并未一致地提高临床安全性，并且在某些配置中与较低的临床安全评分相关。这些发现表明，仅依赖更大的模型或额外的推理时计算可能是提高治疗性人工智能系统安全性的一种低效策略。我们讨论了可持续部署方面的意义，并强调动态模型选择（包括模型级联）是一种在较高风险病例中保持临床性能的同时减少环境影响的潜在方法。
cs.AI / 49 / 2608.11889
DexterSQL: Deep Schema Exploration and Rule-based Correction for Text-to-SQL Generation
DexterSQL：用于文本到SQL生成的深度模式探索与基于规则的纠正
large language model
大语言模型相关
Abstract
Prompting-based (\textit{i}.\textit{e}., non-fine-tuning) Text-to-SQL methods, where underlying large language model parameters are not changed for the task, face three problems: (\textit{i})~relying on coarse-grained schema information that may not reveal the fine-grained relationships needed to distinguish ambiguous columns, (\textit{ii})~not capturing recurring SQL-generation failures, and (\textit{iii})~suffering from omission, hallucination, or misplacement of conditions in complex questions. This paper develops \textsc{DexterSQL}, a prompting/non-fine-tuning-based Text-to-SQL system that improves SQL generation with three novel components: (\textit{i})~\emph{deep schema explorator} that identifies ambiguous columns, analyzes their individual and joint data distributions to uncover their relationships and the distinct role of each, (\textit{ii})~\emph{database-agnostic rule creator} that mines mismatches between generated and gold SQL only on the training database and converts them into database-agnostic corrective rules that capture recurring LLM failure patterns; and (\textit{iii})~\emph{multi-path SQL generation} that introduces a dependency-tree-based intermediate representation that uses the question's sentence structure to guide its decomposition into an SQL skeleton for final SQL generation. \textsc{DexterSQL} achieves a higher accuracy compared to the state-of-the-art using both open-source/weight and closed-source/weight models. Particularly, \textsc{DexterSQL}'s shows a high improvement of at least 2.7\% using an open-weight model (GPT-OSS-120B) on BIRD-Dev, with total accuracy 67.6\%. \textsc{DexterSQL} also shows better improvement of at least 0.9\% using closed-weight models, with total accuracy 71.6\% and 72.2\% on BIRD-Dev with GPT-4o and GPT-5.2.
Chinese Translation
基于提示（即，非微调）的文本到SQL方法（其底层大语言模型参数不因该任务而改变）面临三个问题：（i）依赖粗粒度模式信息，可能无法揭示区分模糊列所需的细粒度关系；（ii）无法捕获反复出现的SQL生成失败；（iii）在复杂问题中会出现条件遗漏、幻觉或位置错误。本文开发了DexterSQL，一个基于提示/非微调的文本到SQL系统，它通过三个新颖组件改进SQL生成：（i）深度模式探索器，识别模糊列，分析它们的单独和联合数据分布，以揭示它们之间的关系以及每一列的不同作用；（ii）数据库无关的规则创建器，仅在训练数据库上挖掘生成SQL与金标准SQL之间的不匹配，并将其转换为数据库无关的纠正规则，以捕获反复出现的大语言模型失败模式；（iii）多路径SQL生成，引入基于依存树的中间表示，利用问题的句子结构引导其分解为SQL骨架，以用于最终SQL生成。DexterSQL在使用开源/开放权重和闭源/封闭权重模型时，相比最先进方法均取得了更高的准确率。特别是，DexterSQL在BIRD-Dev上使用开放权重模型（GPT-OSS-120B）时显示出至少2.7%的显著提升，总准确率达到67.6%。DexterSQL在使用封闭权重模型时还显示出至少0.9%的更好提升，在BIRD-Dev上使用GPT-4o和GPT-5.2时总准确率分别为71.6%和72.2%。
cs.AI / 50 / 2608.11415
TRACES: A Benchmark for Epistemic Reliability in Scientific Reasoning by LLMs
TRACES：大型语言模型科学推理中的认知可靠性基准
large language model
大语言模型相关
Abstract
Large language models are being proposed as agents in scientific workflows, in domains where no downstream verifier exists. Such deployment assumes the model can distinguish reliable scientific literature from unreliable literature, a capability that has not yet been directly measured. Existing benchmarks evaluate factuality on questions with known answers; the failure mode we target here is different. We introduce a probe corpus of 42 retracted, fraudulent, and pseudoscientific papers, paired with a methodology for eliciting and scoring single-shot model engagement with each paper's framing. Each probe pairs a preamble extracted near-verbatim from the target paper with a scientifically plausible study-design request. The probes span five claim types: fabricated observation, pseudophysical mechanism, magical premise, legitimization bridge, and cargo-cult experiment. Two complementary scores measure whether a model rejects the flawed premise outright (IFR-a) and whether it recognizes the unreliability while still engaging (IFR-i). A depth score, the Engagement Depth Index (EDI), quantifies reproduction of paper- or field-specific withheld details. Across 30 models and 10 repeated runs, aggregate IFR-a is 0.93 $\pm$ 0.004 and aggregate IFR-i is 0.809 $\pm$ 0.009. Models engaged with untenable premises in 95% of all non-empty responses. Every evaluated model fails more than 71% of agentic probes, and 22 of 30 models fail more than 90% of the time. Rejections are concentrated on a small number of high-notoriety topics and specific probes, and disappear under matched-structure controls. These results are consistent with topic-keyed safety behavior rather than robust epistemic competence, and indicate an urgent need for guardrail infrastructure for scientific deployment of language models.
Chinese Translation
大型语言模型正被提议用作科学工作流中的智能体，并用于不存在下游验证器的领域。此类部署假设模型能够将可靠的科学文献与不可靠的科学文献区分开来，而这一能力尚未得到直接测量。现有基准在已知答案的问题上评估事实性；我们在此所针对的失败模式有所不同。我们引入了一个包含 42 篇被撤稿、欺诈性和伪科学论文的探针语料库，并配套了一种方法，用于引发和评分模型对每篇论文框架的单次参与。每个探针将目标论文中近乎逐字抽取的导言与科学上合理的研究设计请求配对。这些探针涵盖五种主张类型：伪造观察、伪物理机制、神奇前提、合法化桥梁和货物崇拜式实验。两个互补分数衡量模型是否直接拒绝有缺陷的前提（IFR-a），以及模型是否在认识到不可靠性的同时仍进行参与（IFR-i）。一个深度分数，即参与深度指数（EDI），量化了对论文或领域所特有的未提供细节的复现。在 30 个模型和 10 次重复运行中，总体 IFR-a 为 0.93 $\pm$ 0.004，总体 IFR-i 为 0.809 $\pm$ 0.009。在所有非空响应中，模型在 95% 的情况下参与了站不住脚的前提。每个被评估模型在超过 71% 的智能体探针上失败，30 个模型中有 22 个在超过 90% 的情况下失败。拒绝集中在少数高知名度主题和特定探针上，并且在结构匹配的对照条件下消失。这些结果与按主题触发的安全行为而非稳健的认知能力相一致，并表明在语言模型的科学部署中迫切需要护栏基础设施。
cs.LG / 51 / 2608.11327
Long-Horizon Forecasting of Complete Financial Statements with Forma
使用Forma进行完整财务报表的长期预测
large language model
大语言模型相关
Abstract
Specialist training beats generalist scale when forecasting financial statements. To our knowledge, no prior work jointly forecasts complete financial statements beyond one year, yet in a discounted-cash-flow valuation most firm value sits past that window. We release ProForma-20Q, a reproducible benchmark for forecasting 78 statement line items 1-20 quarters ahead, for anonymized firms, from past statements and an industry code, scored by change-space $R^2$. On it, Forma, a transformer that reads statements as sets of (account, quarter, value) tuples and maximizes a masked-tuple Gaussian likelihood, beats every competitor we field: classical machine learning, chained gradient boosting, a zero-shot time-series foundation model, and frontier large language models. Its lead widens with horizon, where valuation needs accuracy most, and its Gaussian predictive intervals never under-cover. Forma's forecasts nearly satisfy accounting identities; exact coherence is recoverable at no statistically significant accuracy cost. Its tuple interface supports scenario analysis without retraining, and we show that pinning future revenue paths sharpens the rest of the statement.
Chinese Translation
在预测财务报表时，专业化训练胜过通用规模。据我们所知，此前没有工作联合预测超过一年的完整财务报表，但在折现现金流估值中，大部分公司价值位于该时间窗口之后。我们发布了ProForma-20Q，这是一个可复现的基准，用于针对匿名公司、基于过往报表和行业代码，预测未来1至20个季度的78个报表行项目，并以变化空间 $R^2$ 进行评分。在该基准上，Forma——一种将报表读取为（科目、季度、数值）元组的集合并最大化掩码元组高斯似然的Transformer——击败了我们所测试的每一个竞争对手：经典机器学习、链式梯度提升、零样本时序基础模型以及前沿大语言模型。其领先优势随着预测期延长而扩大，而估值最需要准确性的正是较长的预测期，并且其高斯预测区间从未出现欠覆盖。Forma的预测几乎满足会计恒等式；精确一致性可以在没有统计学显著准确率损失的情况下恢复。其元组接口支持无需重新训练的情景分析，并且我们表明，固定未来收入路径会使报表其余部分的预测更加精准。
cs.LG / 52 / 2608.11361
Lifecycle-Optimal Tokenization: Vocabulary Size as a Deployment-Regime-Dependent Infrastructure Parameter
生命周期最优分词：词表大小作为依赖于部署体制的基础设施参数
large language model
大语言模型相关
Abstract
Tokenizer vocabulary size is a foundational design choice in large language model (LLM) infrastructure, yet it is typically fixed at training time based on convention rather than deployment analysis. We show that the cost-optimal vocabulary is not a constant but a function of the serving regime. We formalize total deployment cost as $C_{lifecycle}(V) = C_{train}(V) + λ\cdot C_{infer}(V, B)$, where $λ$ is inference volume and $B$ is the serving batch size. Through controlled experiments on two GPU families spanning the memory-bound to compute-bound regimes (A10G, ridge $\approx$ 117 FLOP/byte; A100, ridge $\approx$ 183 FLOP/byte), we demonstrate: (1) the inference-optimal vocabulary shifts 16x with serving batch, from 32k at $B=1$ to 524k at $B=64+$, driven by amortization of the $V \times d$ unembedding matrix read; (2) at 1.3-2.3B model scale, quality (bits per byte, BPB) is optimized at $V=65$k, confirming scale-dependent vocabulary preference; (3) the lifecycle-optimal vocabulary diverges from training-optimal by up to 16x for production deployments. Quality is approximately invariant across the optimal range ($<$2% BPB spread), making vocabulary a pure systems optimization with no quality penalty in the measured range. Our results provide actionable capacity planning guidance: on-device deployments ($B=1$) should use $V \approx 32$k; datacenter serving ($B \geq 64$, $λ\geq 10$) should use $V \approx 131$-262k.
Chinese Translation
分词器词表大小是大语言模型（LLM）基础设施中的一项基础性设计选择，但通常在训练时依据惯例而非部署分析被固定下来。我们表明，成本最优的词表并不是一个常数，而是服务体制的函数。我们将总部署成本形式化为 $C_{lifecycle}(V) = C_{train}(V) + λ\cdot C_{infer}(V, B)$，其中 $λ$ 是推理量，$B$ 是服务批大小。通过在两个跨越内存受限到计算受限体制的 GPU 系列（A10G，ridge $\approx$ 117 FLOP/byte；A100，ridge $\approx$ 183 FLOP/byte）上进行受控实验，我们证明：（1）推理最优词表随服务批大小变化 16 倍，从 $B=1$ 时的 32k 到 $B=64+$ 时的 524k，其驱动因素是对 $V \times d$ 反嵌入矩阵读取的摊销；（2）在 1.3-2.3B 模型规模下，质量（每字节比特数，BPB）在 $V=65$k 时达到最优，证实了规模相关的词表偏好；（3）对于生产部署，生命周期最优词表与训练最优词表最多相差 16 倍。在最优范围内质量近似不变（BPB 差异 <2%），这使得词表成为一个纯粹的系统优化问题，在测量范围内没有质量损失。我们的结果提供了可操作的容量规划指导：设备端部署（$B=1$）应使用 $V \approx 32$k；数据中心服务（$B \geq 64$，$λ\geq 10$）应使用 $V \approx 131$-262k。
cs.LG / 53 / 2608.11419
Diffusion-Based Data-Driven Assortment Optimization
基于扩散的数据驱动品类优化
diffusion
扩散模型相关
Abstract
Assortment optimization is a fundamental problem in revenue management, typically addressed using parametric choice models such as the multinomial logit (MNL) and its variants. While these models enable tractable formulations, their performance is sensitive to model misspecification and often struggles to capture complex customer behavior. In this paper, we propose a model-agnostic framework for assortment optimization based on guided discrete diffusion. We represent assortments as binary vectors and perform stochastic search via a learned reverse diffusion process, avoiding explicit combinatorial enumeration. To incorporate decision objectives, we introduce a reward-guided mechanism that biases local transitions using estimates of expected revenue. This allows the method to effectively balance exploration and exploitation during generation. Empirically, we show that the proposed approach consistently identifies high-quality assortments and remains robust under model misspecification, often recovering near-optimal solutions in high-dimensional settings. Moreover, the generative nature of diffusion enables the production of diverse high-performing assortments, offering flexibility beyond a single deterministic solution. These results highlight the potential of generative modeling as a scalable and robust paradigm for combinatorial optimization in data-driven decision-making.
Chinese Translation
品类优化是收益管理中的一个基本问题，通常使用参数化选择模型（如多项Logit（MNL）模型及其变体）来解决。虽然这些模型能够实现易于处理的公式化表述，但其性能对模型误设敏感，并且往往难以捕捉复杂的顾客行为。在本文中，我们提出了一种基于引导离散扩散的模型无关的品类优化框架。我们将品类组合表示为二元向量，并通过学习到的逆向扩散过程执行随机搜索，从而避免显式的组合枚举。为了纳入决策目标，我们引入了一种奖励引导机制，该机制利用期望收益的估计值来偏置局部转移。这使得该方法在生成过程中能够有效地平衡探索与利用。通过实验，我们表明所提出的方法能够持续识别出高质量的品类组合，并且在模型误设下保持鲁棒，通常在高维设定中恢复出近最优解。此外，扩散的生成特性使得能够产生多样化的高性能品类组合，从而提供了超越单一确定性解的灵活性。这些结果凸显了生成建模作为数据驱动决策中组合优化的一种可扩展且稳健的范式的潜力。
cs.LG / 54 / 2608.11572
RECAST: A Machine-Learning Framework for Correction and Super-Resolution of Coarse-Grid PDE Solvers
RECAST：用于粗网格 PDE 求解器校正与超分辨率的机器学习框架
diffusion
扩散模型相关
Abstract
Coarse-grid numerical solvers can substantially reduce the computational cost of time-dependent PDE simulation, but under-resolution often degrades both the trajectory and the spatial fidelity of the solution. We introduce RECAST (Recurrent Error Correction And Super-resolution of coarse-grid Trajectories), a machine-learning framework designed to restore this lost accuracy while retaining coarse-grid evolution. RECAST combines learned correction within the numerical time-stepping loop with reconstruction of the corresponding fine-grid state from the corrected coarse history. We evaluate the framework on six one-dimensional PDE systems spanning transport, diffusion, dispersion, reaction, and wave dynamics, using spatial grids coarsened by factors of 8-16 and 1000-step closed-loop rollouts from unseen initial conditions. Across the test cases, RECAST remains closely aligned with the fine-grid reference solutions and reduces time-averaged relative error by approximately 50-92% compared with the corresponding uncorrected coarse-grid solvers. Additional tests show generalization to unseen PDE parameter values, while comparison with a contemporary coarse-correction architecture shows that RECAST achieves lower error and better long-horizon agreement with the fine-grid reference over 5000-step rollouts. These results demonstrate that the learned correction and reconstruction capabilities of RECAST can enable substantially coarser PDE evolution without the corresponding loss of solution fidelity, providing a proof-of-concept route toward machine-learning acceleration of higher-dimensional numerical simulations across science and engineering.
Chinese Translation
粗网格数值求解器可以大幅降低含时 PDE 模拟的计算成本，但分辨率不足常常会降低解的轨迹和空间保真度。我们引入了 RECAST（Recurrent Error Correction And Super-resolution of coarse-grid Trajectories，即粗网格轨迹的循环误差校正与超分辨率），这是一个机器学习框架，旨在恢复丢失的精度，同时保留粗网格演化。RECAST 将数值时间步进循环中的学习校正与从校正后的粗网格历史重建相应细网格状态结合起来。我们在六个一维 PDE 系统上评估该框架，这些系统涵盖输运、扩散、色散、反应和波动动力学，使用空间网格粗化 8–16 倍，并从未见过的初始条件进行 1000 步闭环推演。在所有测试案例中，RECAST 始终与细网格参考解保持高度一致，并且与相应的未校正粗网格求解器相比，将时间平均相对误差降低了约 50%–92%。额外测试表明其可泛化到未见过的 PDE 参数值，而与一种当代粗校正架构的比较表明，RECAST 在 5000 步推演中实现了更低误差以及与细网格参考更好的长时程一致性。这些结果表明，RECAST 的学习校正与重建能力能够实现显著更粗的 PDE 演化，而不会相应损失解的保真度，为在科学与工程领域中对高维数值模拟进行机器学习加速提供了一条概念验证路径。
cs.LG / 55 / 2608.11623
FM-LLM: A frequency-enhanced mixture-of-experts framework for adapting LLMs to time series forecasting
FM-LLM：一种用于将大语言模型适配到时间序列预测的频率增强混合专家框架
large language model
大语言模型相关
Abstract
Recent advances in Large Language Models (LLMs) have spurred cross-modal solutions for time-series forecasting. However, existing methods rely heavily on textual prompts for modality alignment-introducing nontrivial computational overhead and failing to leverage the rich spectral dynamics inherent in time-series data. To enable prompt-free, frequency-aware adaptation of frozen LLMs, we propose FM-LLM (Frequency-Enhanced Mixture-of-Experts for adapting LLMs to Time Series Forecasting), an autoregressive framework grounded in constrained asymmetric coupling. A Fourier Analysis Network (FAN)-based spectral token aligner injects structured harmonic representations directly into the frozen LLM with numerical compatibility. An asymmetric Mixture-of-Experts (MoE) decoder enforces role separation: shared experts with lightweight FAN layers reconstruct the global periodic backbone, while routed experts-restricted to standard FFNs-specialize in modeling non-periodic residual dynamics. A time-frequency hybrid loss function jointly optimizes temporal accuracy and spectral consistency, mitigating error accumulation during long-horizon autoregressive rollouts. Evaluated across eleven public benchmarks, FM-LLM achieves state-of-the-art performance on 59 out of 78 evaluation metrics. Compared to the strongest autoregressive LLM-based baseline, it delivers average improvements of 5.3% in MSE and 5.6% in MAE, with maximum gains reaching 8.0% for MSE and 8.4% for MAE. FM-LLM also demonstrates robust transferability, maintaining superior performance in 10% few-shot and zero-shot forecasting scenarios.
Chinese Translation
近来大语言模型（LLMs）的进展推动了时间序列预测的跨模态解决方案。然而，现有方法严重依赖文本提示进行模态对齐——这引入了不小的计算开销，并且未能利用时间序列数据中固有的丰富频谱动态。为了实现冻结大语言模型的无提示、频率感知适配，我们提出 FM-LLM（用于将大语言模型适配于时间序列预测的频率增强混合专家方法），一个基于约束非对称耦合的自回归框架。基于傅里叶分析网络（FAN）的频谱令牌对齐器将结构化谐波表示以数值兼容的方式直接注入冻结的大语言模型。非对称混合专家（MoE）解码器强制角色分离：带有轻量级 FAN 层的共享专家重建全局周期主干，而被限制在标准 FFN 上的路由专家专门建模非周期残差动态。时频混合损失函数联合优化时间精度和频谱一致性，减轻长时域自回归推演过程中的误差累积。在十一个公开基准上评估后，FM-LLM 在 78 项评估指标中的 59 项上取得了最先进的性能。与最强的基于自回归大语言模型的基线相比，它在 MSE 上平均提升 5.3%，在 MAE 上平均提升 5.6%，最大增益在 MSE 上达到 8.0%，在 MAE 上达到 8.4%。FM-LLM 还展现出稳健的迁移能力，在 10% 少样本和零样本预测场景中保持优越性能。
cs.LG / 56 / 2608.11656
Continuous-Latent Predictive Modeling with Semantic Alignment for EEG-Language Foundation Models
面向EEG-语言基础模型的连续潜在预测建模与语义对齐
large language model
大语言模型相关
Abstract
Recent advances in EEG foundation models have demonstrated the potential of large-scale pretraining to enable generalizable neural decoding across subjects, recording environments, and datasets. However, dominant pretraining paradigms face key challenges: masked autoencoding tends to prioritize low-level signal reconstruction over task-relevant semantics, while autoregressive modeling creates a mismatch between continuous neural dynamics and discrete token spaces. To address these challenges, new strategies are needed to effectively align continuous EEG representations with natural-language semantics and enable their integration with large language models. Accordingly, we propose Brain Latent Predictive Model (BLPM), an EEG-language foundation model that reformulates heterogeneous EEG decoding tasks as a continuous semantic embedding prediction problem. BLPM introduces a Continuous EEG Latent Predictive (CELP) encoder that learns transferable representations through latent target prediction. Building on these representations, a Multi-Query Semantic Decomposition (MQSD) module extracts task-relevant information and aligns continuous EEG representations with textual semantics within a shared latent space according to their semantic relationships. Experiments across multiple benchmarks demonstrate consistent generalization performance across diverse tasks, establishing continuous latent semantic prediction as an effective paradigm for EEG-language foundation models.
Chinese Translation
近年来EEG基础模型的进展已经展示出大规模预训练在实现跨被试、记录环境和数据集的泛化神经解码方面的潜力。然而，主流的预训练范式面临关键挑战：掩码自编码倾向于优先进行低层信号重建，而非任务相关语义；而自回归建模则在连续神经动力学与离散token空间之间造成了不匹配。为了应对这些挑战，需要新的策略来有效地将连续EEG表示与自然语言语义对齐，并使其能够与大型语言模型集成。因此，我们提出了大脑潜在预测模型（BLPM），这是一种EEG-语言基础模型，它将异构EEG解码任务重新表述为连续语义嵌入预测问题。BLPM引入了一种连续EEG潜在预测（CELP）编码器，通过潜在目标预测来学习可迁移表示。在这些表示的基础上，多查询语义分解（MQSD）模块提取任务相关信息，并根据语义关系在共享潜在空间中将连续EEG表示与文本语义对齐。多个基准上的实验展示了在多种任务上一致的泛化性能，从而确立了连续潜在语义预测作为EEG-语言基础模型的一种有效范式。
cs.LG / 57 / 2608.11674
GCPO: Diagnosing and Constraining Subspace Geometry in Rollout RL for LLMs
GCPO：诊断与约束面向 LLM 的 Rollout RL 中的子空间几何
large language model
大语言模型相关
Abstract
On-policy rollout methods such as GRPO are central to post-training of large language models, yet they frequently suffer from training instabilities, cross-task capability degradation, and response-length inflation. Although prior work has characterized the subspace geometry of aggregate updates, the stepwise variation of this geometry and its relationship to model performance remain unclear. We introduce Principal-Subspace Overlap, a dimension-corrected measure of individual rollout updates relative to the dominant singular subspaces of pretrained weights. Despite low average overlap, transient spikes often precede performance degradation. To address this, we propose GCPO (Geometrically Constrained Policy Optimization), which applies hard bilateral orthogonal projections to constrain updates to the complementary subspaces, preventing such excursions by construction. Across mathematical reasoning, code generation, and tool-use tasks on Qwen3-8B and GLM4-9B, GCPO consistently outperforms GRPO and recent variants, including DAPO and GSPO, improving over the base models and the strongest baseline by up to 27.69 and 2.37 points, respectively. Furthermore, GCPO preserves general capabilities, eliminates response-length inflation, and stabilizes policy entropy. Our findings provide a new diagnostic lens and a principled design perspective for stable reinforcement learning post-training.
Chinese Translation
诸如 GRPO 之类的同策略 rollout 方法是大语言模型后训练的核心，但它们经常遭受训练不稳定、跨任务能力退化以及响应长度膨胀的问题。尽管先前工作已经刻画了聚合更新的子空间几何，但这种几何的逐步变化及其与模型性能的关系仍不清楚。我们引入了主子空间重叠（Principal-Subspace Overlap），这是一种针对单个 rollout 更新相对于预训练权重主导奇异子空间的维度校正度量。尽管平均重叠较低，瞬时的尖峰往往先于性能退化出现。为解决这一问题，我们提出了 GCPO（几何约束策略优化），它应用硬性双边正交投影将更新约束到互补子空间，从而在构造上防止此类偏离。在 Qwen3-8B 和 GLM4-9B 上的数学推理、代码生成和工具使用任务中，GCPO 始终优于 GRPO 及近期变体（包括 DAPO 和 GSPO），相较基础模型和最强基线分别最高提升 27.69 和 2.37 个点。此外，GCPO 保持通用能力，消除响应长度膨胀，并稳定策略熵。我们的研究结果为稳定的强化学习后训练提供了新的诊断透镜和原则性设计视角。
cs.LG / 58 / 2608.11785
TradingMoE: Routing the Right Experts in Evolving Markets
TradingMoE：在不断变化的市场中路由合适的专家
large language model
大语言模型相关
Abstract
Large language models (LLMs) have shown strong potential for financial analysis and trading, but direct trading remains challenging because the predictive capabilities required can vary across assets, decision fields, and market conditions. Existing LLM-based trading systems either coordinate human-defined external experts or adopt conventional internal Mixture-of-Experts (MoE) routers that do not directly evaluate how individual experts contribute to trading decisions. Moreover, these routers receive no direct signal indicating when an inactive expert has become more suitable as market conditions change. We find that native router scores poorly reflect how much individual experts improve trading decisions, frequently leaving better alternatives unselected. We further reveal that token-specific expert usefulness exhibits a compact low-dimensional structure. Based on these findings, we propose TradingMoE, a trading-oriented sparse MoE that augments a frozen dense LLM with lightweight residual experts. We introduce a Query-Key router that represents the expertise required by each token under the current market context as a low-dimensional query and matches it with learnable expert keys. We further propose a sparse expert selection update mechanism that samples a few inactive experts during training and estimates whether they should replace the weakest expert in the current Top-k route. This mechanism enables the router to update expert selection as market conditions change while preserving sparse computation. Experiments against 22 baselines on stock and cryptocurrency markets show that TradingMoE improves cumulative return over the best-performing baselines by 30.89% and 30.7%, respectively. Rolling paper-trading experiments further demonstrate that its advantage persists under forward-only deployment.
Chinese Translation
大语言模型（LLM）在金融分析与交易方面已展现出强大潜力，但直接交易仍然具有挑战性，因为所需的预测能力可能因资产、决策领域和市场条件而异。现有的基于LLM的交易系统要么协调人工定义的外部专家，要么采用传统的内部混合专家（MoE）路由器，而这些路由器并不直接评估各个专家对交易决策的贡献。此外，这些路由器接收不到直接信号来指示某个非活跃专家在市场条件变化时何时已变得更合适。我们发现，原生路由器评分不能很好地反映各个专家对交易决策的改进程度，经常使更优的替代专家未被选中。我们进一步揭示，特定于token的专家有用性表现出紧凑的低维结构。基于这些发现，我们提出TradingMoE，这是一种面向交易的稀疏MoE，通过轻量级残差专家增强冻结的稠密LLM。我们引入一种Query-Key路由器，该路由器将当前市场背景下每个token所需的专业知识表示为一个低维查询，并将其与可学习的专家键进行匹配。我们进一步提出一种稀疏专家选择更新机制，该机制在训练期间采样少量非活跃专家，并估计它们是否应替换当前Top-k路由中最弱的专家。该机制使路由器能够在市场条件变化时更新专家选择，同时保持稀疏计算。在股票和加密货币市场上与22个基线进行的实验表明，TradingMoE相较于表现最佳的基线分别将累计收益提高了30.89%和30.7%。滚动模拟交易实验进一步表明，其优势在仅前向部署下持续存在。
cs.LG / 59 / 2608.11967
LoongReflect: Boosting Long-Horizon Reflection in Search Agents via Global Perspective Distillation
LoongReflect：通过全局视角蒸馏提升搜索智能体的长程反思
large language model
大语言模型相关
Abstract
Large language model agents increasingly rely on long-horizon reasoning to solve complex tasks involving planning, tool use, and memory. A critical capability in such settings is reflection: assessing trajectory progress, identifying missing evidence and unreliable intermediate states, and deciding whether to continue, revise, or abandon the current branch. Learning effective reflection, however, is challenging because reflection is performed locally within the current branch, whereas its utility can only be determined by its contribution to the final trajectory outcome. This local-global mismatch makes outcome-based reinforcement learning provide only local, sparse and delayed supervision for reflective decisions. To solve these, we propose LoongReflect, a training framework that formulates reflection as a memory-control policy. The agent operates over a reversible trajectory tree using explicit reflect and backtrack actions. Reflection consolidates verified facts, missing evidence, and branch-specific risks into working memory, while backtracking removes an unreliable branch from the active context and preserves a concise corrective lesson. To learn this policy, LoongReflect combines two complementary signals through a look-ahead, extragradient-style coordination mechanism. A fast channel distills globally informed reflective behavior from a privileged teacher, with supervision restricted to reflection and backtracking tokens. A slow channel optimizes complete trajectories using outcome-based GRPO, aligning local control decisions with final task success. Experiments on multi-hop retrieval-augmented generation and mathematical reasoning benchmarks demonstrate consistent improvements over outcome-only reinforcement learning and self-distillation baselines.
Chinese Translation
大型语言模型智能体越来越依赖长程推理来解决涉及规划、工具使用和记忆的复杂任务。在此类场景中，一项关键能力是反思：评估轨迹进展、识别缺失证据和不可靠的中间状态，并决定是继续、修正还是放弃当前分支。然而，学习有效的反思具有挑战性，因为反思是在当前分支内局部执行的，而其效用只能通过它对最终轨迹结果的贡献来确定。这种局部与全局之间的不匹配使得基于结果的强化学习只能为反思性决策提供局部的、稀疏的且延迟的监督信号。为了解决这些问题，我们提出了 LoongReflect，一个将反思形式化为记忆控制策略的训练框架。该智能体在可逆轨迹树上运行，使用显式的反思（reflect）和回溯（backtrack）动作。反思将已验证的事实、缺失的证据和分支特定的风险整合到工作记忆中，而回溯则从活动上下文中移除不可靠的分支，并保留一条简洁的纠正性教训。为了学习这一策略，LoongReflect 通过一种前瞻的、超梯度风格的协调机制结合两个互补信号。快速通道从特权教师中蒸馏出具有全局信息的反思行为，并且监督仅限于反思和回溯 token。慢速通道使用基于结果的 GRPO 优化完整轨迹，使局部控制决策与最终任务成功保持一致。在多跳检索增强生成和数学推理基准上的实验表明，与仅基于结果的强化学习和自蒸馏基线相比，该方法取得了一致的改进。
cs.LG / 60 / 2608.12083
Faithful, Sufficient and Understandable: Rethinking Graph Counterfactual Explanations via Discrete Diffusion Inversion
忠实、充分且可理解：通过离散扩散反演重新思考图反事实解释
diffusion
扩散模型相关
Abstract
Graph Neural Networks (GNNs) achieve strong predictive performance on graph-structured data across domains such as chemistry, biology, and network analysis, yet they provide no intrinsic explanation of their predictions. This limits their adoption in high-stakes and safety-critical settings. Counterfactual explanations address this by revealing the minimal structural modifications that would change a model's prediction. On graphs, however, such a modification is hard to produce. The search space is discrete and combinatorial, and a valid answer must respect categorical node and edge types together with domain rules such as chemical valency in the case of molecular graphs. Existing explainers give up one of two things. Either edits are not held on the data manifold, or the search does not span the full edit space. We propose Graph Diffusion Counterfactual Explanation via Inversion (GDCE-I), which gives up neither. A discrete denoising diffusion model with a novel discrete inversion scheme enables distribution-aware edits leveraging the whole domain edit space. We further address the incomplete and inconsistent evaluation of graph counterfactuals by deriving a framework of explanation desiderata and applying it to every method under one shared protocol. Across four benchmarks, GDCE-I outperforms related work by a large margin on the defined framework. For the molecular domain, we further qualitatively show that GDCE-I attains interpretable in-distribution solutions.
Chinese Translation
图神经网络（GNN）在化学、生物学和网络分析等领域中的图结构数据上取得了强大的预测性能，但它们没有为其预测提供内在解释。这限制了它们在高风险和安全关键场景中的应用。反事实解释通过揭示会改变模型预测的最小结构修改来解决这一问题。然而，在图数据上，这样的修改很难生成。搜索空间是离散且组合的，并且一个有效答案必须同时遵守类别型节点和边类型以及领域规则，例如分子图中的化学价。现有的解释器放弃了两者之一。要么编辑不保持在数据流形上，要么搜索没有覆盖完整的编辑空间。我们提出了通过反演进行图扩散反事实解释（GDCE-I），该方法两者都不放弃。一个具有新颖离散反演方案的离散去噪扩散模型能够利用整个领域编辑空间进行分布感知的编辑。我们进一步通过推导一个解释需求框架，并在一个统一协议下将其应用于每种方法，来解决图反事实评估不完整和不一致的问题。在四个基准上，GDCE-I在所定义的框架上大幅优于相关工作。对于分子领域，我们进一步定性地表明GDCE-I获得了可解释的分布内解决方案。
cs.MA / 61 / 2608.11407
Top-down Traffic Scenario Generation via Joint Initial-Goal Diffusion and Trajectory Infilling
通过联合初始-目标扩散与轨迹填充的自上而下交通场景生成
diffusion
扩散模型相关
Abstract
Robust traffic simulators are crucial for developing and testing autonomous vehicles to reduce the costly, labor-intensive real-world data collection process and the need for physical presence on the road. However, existing simulators require agents' initial states to generate trajectories, which limits scalability and diversity due to restrictions on the given initial states. While data-driven agent initialization has been widely studied, the generated initial states are not interpretable in terms of why the agents are initialized at those specific locations. Given known initial states, trajectory generation is also a challenging problem, as the model must learn the variability of the destination and how agents should reach it over time. In this paper, we propose TrafficDiffuser, a top-down traffic scenario generation framework that generates high-level traffic scenarios, defined by initial and goal state pairs, by jointly modeling them. The high-level scenario generation makes initial states better interpretable and reduces trajectory generation into as simple as an infilling problem. We demonstrate how the generated high-level traffic scenarios can be used, including constraining based on different trajectory modes and integrating them with existing trajectory generation models. We conduct extensive experiments on the Argoverse 2 motion prediction dataset to evaluate how well the generated outputs capture real-world distributions. In addition to generating goal states, TrafficDiffuser outperforms the next-best approach for agent initialization, reducing speed distribution distance by 55.3% and the off-road rate by 2.8%.
Chinese Translation
稳健的交通模拟器对于开发和测试自动驾驶汽车至关重要，以减少成本高昂、劳动密集型的真实世界数据采集过程以及对道路实车存在的需求。然而，现有模拟器需要智能体的初始状态来生成轨迹，这由于对给定初始状态的限制而限制了可扩展性和多样性。尽管数据驱动的智能体初始化已被广泛研究，但所生成的初始状态在智能体为何被初始化在这些特定位置方面缺乏可解释性。在给定已知初始状态的情况下，轨迹生成也是一个具有挑战性的问题，因为模型必须学习目的地的可变性以及智能体应如何随时间到达该目的地。在本文中，我们提出 TrafficDiffuser，一种自上而下的交通场景生成框架，通过联合建模由初始状态和目标状态对定义的高层交通场景来生成这些场景。高层场景生成使初始状态更具可解释性，并将轨迹生成简化为一个简单的填充问题。我们展示了所生成的高层交通场景的使用方式，包括基于不同轨迹模式进行约束，以及将其与现有轨迹生成模型集成。我们在 Argoverse 2 运动预测数据集上进行了大量实验，以评估生成输出捕捉真实世界分布的效果。除了生成目标状态外，TrafficDiffuser 在智能体初始化方面优于次优方法，将速度分布距离降低了 55.3%，将偏离道路率降低了 2.8%。
cs.SE / 62 / 2608.11386
The Devil Is in the Interface: Evaluating How Tool Architecture Shapes Coding Agent Behavior
魔鬼在接口中：评估工具架构如何塑造编码智能体行为
large language model
大语言模型相关
Abstract
As large language models continue to improve, agentic systems are becoming increasingly important, and tools are a key design dimension because they determine how agents access information and take action in their environments. Prior work on agent tooling has primarily focused on expanding what agents can do, but has paid less systematic attention to how those capabilities are organized and exposed to the model. We refer to this latter design dimension as tool architecture. We study tool architecture in coding agents through controlled experiments on repository-level issue fixing, comparing six tool architectures that hold the underlying information and actions similar while varying how they are organized and exposed to the model, across three actors and a total of 11,700 trajectories. Our experiments show that, even when tools provide similar capabilities, tool architecture changes agent behavior: Compared to a basic architecture where the agent has only the bash tool, more structured low-level interfaces improve consistency across repeated attempts by up to 4.7 $\times$; natural-language search broadens repository exploration and increases access to relevant files by more than 11%; and Python CodeAct-style interfaces achieve similar task performance with 41.6% fewer steps and 56.3% lower token usage. By contrast, lightweight text-based cognitive-scaffolding tools, such as tools that let the agent record intermediate reasoning, have limited effect on actor behavior.
Chinese Translation
随着大型语言模型不断改进，智能体系统变得越来越重要，而工具是一个关键的设计维度，因为工具决定了智能体如何在其环境中获取信息和采取行动。先前关于智能体工具设计的工作主要侧重于扩展智能体能够做什么，但对这些能力如何被组织并暴露给模型的关注则较为缺乏系统性。我们将后一个设计维度称为工具架构。我们通过在仓库级问题修复上的受控实验来研究编码智能体中的工具架构，比较了六种工具架构；这些架构在底层信息和动作上保持相似，但在如何组织和暴露给模型方面有所不同；实验涉及三个行动者，共计11,700条轨迹。我们的实验表明，即使工具提供相似的能力，工具架构也会改变智能体行为：与智能体仅拥有 bash 工具的基础架构相比，更结构化的低层接口将重复尝试之间的一致性最多提高 4.7 $\times$；自然语言搜索扩大了仓库探索范围，并将对相关文件的访问增加了超过 11%；Python CodeAct 风格接口以更少的步骤和更低的 token 用量实现了相似的任务性能，步骤减少 41.6%，token 用量降低 56.3%。相比之下，轻量级的基于文本的认知脚手架工具（例如允许智能体记录中间推理的工具）对行动者行为的影响有限。
cs.SE / 63 / 2608.11394
GraphAlignCoder: Aligning Program and Proof Graphs for Code Generation
GraphAlignCoder：对齐程序图与证明图以用于代码生成
large language model
大语言模型相关
Abstract
Code large language models (LLMs) can generate syntactically plausible programs that nevertheless violate hidden semantic constraints. Existing execution-feedback training methods identify whether a completed program fails, but provide limited supervision about how a correct solution should be organized. We introduce GraphAlignCoder, a training framework that transfers explicit correctness structure into code generation. GraphAlignCoder constructs an implementation graph that captures control and dependence among program regions. In parallel, a constrained Lean pipeline produces proof traces, from which we extract a formal proof-flow graph. The model first learns executable code together with graph-derived descriptions of why individual program regions are correct, and then consolidates this knowledge into code generation. GraphAlignCoder consistently outperforms the base model, code-only SFT, and CodeRL across all benchmarks. Compared with CodeRL, it increases the solved count from 38 to 50 on LiveCodeBench v6 and from 16 to 23 on BigCodeBench Hard, corresponding to relative gains of 31.6% and 43.8%, while also improving BigCodeBench Full from 359 to 363 tasks. The ablation study further shows that verification-graph injection produces the initial reasoning gain, while verification to code consolidation is essential for robust cross-benchmark transfer.
Chinese Translation
代码大语言模型（LLMs）可以生成语法上看似合理、但仍违反隐藏语义约束的程序。现有的执行反馈训练方法能够识别已完成的程序是否失败，但关于正确解决方案应如何组织，提供的监督有限。我们提出 GraphAlignCoder，一个将显式正确性结构迁移到代码生成中的训练框架。GraphAlignCoder 构建一个实现图，捕获程序区域之间的控制与依赖关系。同时，一个受约束的 Lean 流水线生成证明轨迹，我们从中提取形式化的证明流图。模型首先学习可执行代码以及由图导出的、关于各个程序区域为何正确的描述，然后将这些知识整合到代码生成中。GraphAlignCoder 在所有基准上一致优于基础模型、仅代码 SFT 和 CodeRL。与 CodeRL 相比，它在 LiveCodeBench v6 上将已解决数量从 38 提高到 50，在 BigCodeBench Hard 上从 16 提高到 23，对应相对提升分别为 31.6% 和 43.8%，同时还将 BigCodeBench Full 从 359 个任务提高到 363 个任务。消融研究进一步表明，验证图注入带来了最初的推理提升，而验证到代码的整合对于稳健的跨基准迁移至关重要。
cs.SE / 64 / 2608.11454
Simplifying Requirements Engineering in the Context of the LGPD: An LLM-Based Investigation
在LGPD背景下简化需求工程：一项基于大语言模型的研究
large language model
大语言模型相关
Abstract
Compliance with privacy legislation poses a complex challenge to Requirements Engineering (RE): translating legal norms into software requirements. In this context, this study investigates whether Large Language Models (LLMs) can simplify RE within the framework of the Brazilian General Data Protection Law (LGPD). The proposed approach utilizes current legislation to automatically generate User Stories and Acceptance Test Scenarios. The evaluation results demonstrated high performance, confirming the potential of LLMs to ensure regulatory compliance from the software's inception.
Chinese Translation
遵守隐私立法对需求工程（RE）构成了复杂挑战：将法律规范转化为软件需求。在此背景下，本研究调查了大语言模型（LLMs）是否能够在巴西《通用数据保护法》（LGPD）框架内简化需求工程。所提出的方法利用现行立法自动生成用户故事和验收测试场景。评估结果表明其具有高性能，证实了大语言模型从软件初始阶段即确保法规合规性的潜力。
cs.SE / 65 / 2608.11513
Do Influence Tactics Matter? Investigating Prompt Framing Effects in LLM Code Generation
影响策略重要吗？探究提示框架在LLM代码生成中的效应
large language model
大语言模型相关
Abstract
Large Language Models (LLMs) are increasingly integrated into software engineering workflows, helping developers write, debug, test, and maintain code. While prompt wording and structure are known to influence model performance, the impact of psychologically inspired prompt framings remains unexplored. This study investigates whether different psychology-based communication strategies that humans use to persuade or motivate others can lead to more effective prompt framing, which may, in turn, affect LLM behaviour in coding tasks. Drawing on Yukl & Falbe's well-known taxonomy, we operationalized eight influence tactics (like rational persuasion, ingratiation, and exchange) into reproducible prompt templates. These prompt templates were evaluated across five leading open-weight LLMs using two widely adopted benchmarks: LiveCodeBench and SWE-bench Verified. We assessed the resulting code output on four key software quality dimensions: functional correctness, quality, maintainability, and security. Our results show that certain influence-induced prompt framings, particularly those emphasizing urgency, were associated with reduced correctness and security. This work presents the first large-scale empirical study of influence-induced prompt framing in software engineering tasks, offering insights into how linguistic cues may shape LLM outputs. We conclude with practical insights for designing transparent and interpretable human-AI interactions in code generation.
Chinese Translation
大语言模型（LLMs）正越来越多地融入软件工程工作流，帮助开发人员编写、调试、测试和维护代码。虽然提示的措辞和结构已知会影响模型性能，但受心理学启发的提示框架的影响尚未被探索。本研究探讨人类用于说服或激励他人的不同基于心理学的沟通策略是否能够带来更有效的提示框架，进而可能影响LLM在编码任务中的行为。借鉴Yukl和Falbe的著名分类法，我们将八种影响策略（如理性说服、逢迎和交换）操作化为可复现的提示模板。我们在五个领先的开放权重LLM上，使用两个广泛采用的基准（LiveCodeBench和SWE-bench Verified）对这些提示模板进行了评估。我们从四个关键软件质量维度评估了生成的代码输出：功能正确性、质量、可维护性和安全性。我们的结果表明，某些由影响策略诱导的提示框架，尤其是那些强调紧迫性的框架，与正确性和安全性的降低相关。本研究提出了软件工程任务中影响策略诱导的提示框架的首次大规模实证研究，为语言线索如何塑造LLM输出提供了见解。最后，我们为在代码生成中设计透明且可解释的人机交互提供了实践见解。
cs.SE / 66 / 2608.11965
Developing LLM-based Multi-Agent Systems in Software Engineering: A Mixed-Method Experience Report
在软件工程中开发基于大语言模型的多智能体系统：一项混合方法经验报告
large language model
大语言模型相关
Abstract
The proliferation of Generative Artificial Intelligence (Gen AI) powered by large language models (LLMs) has transformed the software development process, introducing new paradigms for code generation, debugging, testing, and maintenance. While early applications focused on leveraging single, independent LLMs to assist developers with isolated tasks, recent advances have shifted toward multi-agent systems (MAS) that orchestrate multiple LLM-based agents working collaboratively toward common objectives. Despite their promising potential, using MAS encompasses a set of challenges for developers who have to carefully select the right technology, devise proper coordination rules, and design specific roles for the involved agents. In this paper, we provide a comprehensive overview of the existing tools and frameworks for implementing MAS in software engineering. First, we conducted a quantitative analysis of the most relevant open source MAS frameworks by evaluating their documentation, features, and capabilities from the developers' perspective. Second, we performed a qualitative evaluation of a subset of the selected frameworks by implementing a common use case: the summarization of README.MD files. The findings show that the selected frameworks provide a good coverage of fundamental components of MAS, though advanced features such as telemetry of agents are still missing. In addition, the empirical evaluation shows that there is no significant difference in terms of ROUGE scores considering the summarization task. Finally, we provide a set of lessons learned and challenges that can help researchers and practitioners to select a suitable MAS framework according to their needs.
Chinese Translation
由大语言模型（LLMs）驱动的生成式人工智能（Gen AI）的普及已经改变了软件开发过程，为代码生成、调试、测试和维护引入了新的范式。虽然早期应用侧重于利用单一、独立的大语言模型来帮助开发者完成孤立任务，但近期进展已转向多智能体系统（MAS），该系统协调多个基于大语言模型的智能体协作以实现共同目标。尽管多智能体系统具有广阔潜力，使用多智能体系统给开发者带来了一系列挑战，他们必须谨慎选择合适技术、制定恰当协调规则，并为所涉及的智能体设计具体角色。在本文中，我们对用于在软件工程中实现多智能体系统的现有工具和框架进行了全面概述。首先，我们从开发者视角出发，通过评估最相关的开源多智能体系统框架的文档、功能和能力，对其进行了定量分析。其次，我们通过实现一个常见用例——README.MD文件的摘要生成——对所选框架的一个子集进行了定性评估。研究结果表明，所选框架对多智能体系统的基本组件提供了较好的覆盖，但仍缺少智能体遥测等高级功能。此外，实证评估表明，就摘要任务而言，在ROUGE分数方面没有显著差异。最后，我们提供了一组经验教训和挑战，可帮助研究者和实践者根据自身需求选择合适的多智能体系统框架。
cs.SE / 67 / 2608.12004
RealisticTritonBench: A Benchmark for Triton-Kernel Generation in Real-World AI Frameworks
RealisticTritonBench：面向真实世界AI框架中Triton内核生成的基准测试
large language model
大语言模型相关
Abstract
In modern AI frameworks, GPU kernels are key to overall system performance. Combining usability, portability, and near-handwritten CUDA performance, Triton is widely adopted for implementing GPU kernels. Recent advances show the potential of large language models (LLMs) to automatically generate Triton kernels, reducing the manual effort required from expert kernel developers. Several benchmarks evaluate LLM-generated Triton kernels. However, they suffer from three key limitations: (1) they restrict tasks to PyTorch-to-Triton translation, failing to reflect the diversity and complexity of real-world Triton tasks; (2) they evaluate only individual-kernel performance rather than end-to-end performance, the core criterion for real-world deployment in AI frameworks; and (3) they rely on manually written evaluation scripts for individual kernels, which may contain flaws that models can exploit to bypass correctness checks and obtain inflated scores. To address these limitations, we introduce RealisticTritonBench, the first benchmark to derive Triton kernel generation tasks from real-world pull requests in popular AI frameworks, enabling realistic, production-like evaluation. RealisticTritonBench systematically extracts PRs that modify Triton kernels from popular open-source AI frameworks and transforms them into generation tasks with concrete engineering contexts. Each task takes a natural language requirement as input and requires a corresponding Triton kernel implementation, with a complete and reproducible evaluation environment. Unlike prior benchmarks focused on isolated kernel performance, RealisticTritonBench integrates generated kernels into their original frameworks and evaluates them using end-to-end tests, enabling a more faithful assessment. We evaluate leading LLMs on RealisticTritonBench and find that they still struggle with real-world Triton kernel generation tasks.
Chinese Translation
在现代AI框架中，GPU内核是整体系统性能的关键。Triton 兼具易用性、可移植性和接近手写 CUDA 的性能，被广泛用于实现 GPU 内核。近期进展表明，大语言模型（LLMs）有潜力自动生成 Triton 内核，从而减少对专家级内核开发人员手工工作的需求。一些基准测试对 LLM 生成的 Triton 内核进行了评估。然而，它们存在三个关键局限：（1）它们将任务限制在 PyTorch 到 Triton 的翻译上，未能反映真实世界 Triton 任务的多样性和复杂性；（2）它们只评估单个内核的性能，而非端到端性能，而端到端性能才是 AI 框架中实际部署的核心标准；（3）它们依赖针对单个内核的手工编写评估脚本，这些脚本可能存在缺陷，模型可利用这些缺陷绕过正确性检查并获得虚高分数。为解决这些局限，我们提出了 RealisticTritonBench，这是首个从流行 AI 框架中的真实世界拉取请求（pull requests）中提取 Triton 内核生成任务的基准测试，能够实现真实、接近生产环境的评估。RealisticTritonBench 系统地从流行的开源 AI 框架中提取修改 Triton 内核的 PR，并将其转化为具有具体工程上下文的生成任务。每个任务以自然语言需求作为输入，要求实现相应的 Triton 内核，并提供完整且可复现的评估环境。与以往专注于孤立内核性能的基准测试不同，RealisticTritonBench 将生成的内核集成到其原始框架中，并使用端到端测试进行评估，从而实现更忠实的评估。我们在 RealisticTritonBench 上评估了领先的 LLM，发现它们在真实世界的 Triton 内核生成任务上仍然面临困难。
cs.SE / 68 / 2608.12025
From Safety Documentation to Safety Knowledge Support: An Evidence-Grounded LLM Framework for Medical Devices
从安全文档到安全知识支持：一种面向医疗器械的证据锚定大语言模型（LLM）框架
large language model
大语言模型相关
Abstract
Medical devices are becoming more software-intensive, connected, and AI-enabled. Their development requires risk-management evidence aligned with ISO 14971 and, for software, IEC 62304. This evidence must be kept consistent across requirements, design decisions, software changes, verification results, complaints, and post-market data. These tasks are costly and depend on scarce safety and domain experts. Large language models (LLMs) may reduce parts of this effort because medical-device safety work is highly document-based. However, current LLM-based safety-engineering studies often address isolated methods, rely on generic prompting or public examples, and provide limited support for source links, traceability, uncertainty handling, lifecycle updates, and recorded expert review. This limits their use in regulated medical-device development. This paper argues that the central research problem is not safety-text generation, but source-linked safety-knowledge support. We propose an evidence-grounded framework that connects device artifacts, controlled knowledge storage and retrieval, method-specific generation of candidate safety items, critique and uncertainty checks, and recorded expert review. The framework prepares, links, checks, and updates candidate safety artifacts for expert decision-making. It does not decide whether a device is safe and does not provide regulatory approval. We also outline an evaluation strategy using non-public or newly built medical-device case studies and expert reference analyses to assess coverage, correctness, relevance, traceability, duplicate rate, unsupported claims, and review effort.
Chinese Translation
医疗器械正变得更加软件密集、互联且具备人工智能能力。其开发需要与 ISO 14971 相一致的风险管理证据，对于软件还需与 IEC 62304 相一致。这些证据必须在需求、设计决策、软件变更、验证结果、投诉和上市后数据之间保持一致。这些任务成本高昂，并依赖稀缺的安全专家和领域专家。大语言模型（LLMs）可能减少部分此类工作，因为医疗器械安全工作高度以文档为基础。然而，当前基于大语言模型的安全工程研究通常只处理孤立的方法，依赖通用提示或公开示例，并在来源链接、可追溯性、不确定性处理、生命周期更新和记录的专家评审方面提供有限支持。这限制了它们在受监管的医疗器械开发中的应用。本文认为，核心研究问题不是安全文本生成，而是来源关联的安全知识支持。我们提出了一个证据锚定的框架，该框架将设备工件、受控知识存储与检索、候选安全条目的方法特定生成、批判与不确定性检查以及记录的专家评审连接起来。该框架为专家决策准备、链接、检查并更新候选安全工件。它不决定设备是否安全，也不提供监管批准。我们还概述了一种评估策略，该策略使用非公开或新建的医疗器械案例研究和专家参考分析来评估覆盖度、正确性、相关性、可追溯性、重复率、无支持的声明以及评审工作量。
cs.SE / 69 / 2608.12228
Towards Automated Domain Model Extraction from Source Code using Heuristics and Open-Source LLMs
迈向使用启发式方法和开源大语言模型从源代码自动提取领域模型
large language model
大语言模型相关
Abstract
Large language models (LLMs) have recently shown strong capabilities for code understanding, making them promising for reverse engineering domain models from source code. However, state-ofthe- art proprietary LLMs cannot be used in many industrial contexts due to privacy and confidentiality constraints, while compact open-source LLMs that can run locally are limited by their context window and cannot process large code bases directly. In this paper, we propose an automated approach to extract domain models from source code using lightweight, locally deployable LLMs. Our method combines structural and semantic heuristics with iterative LLM-based reasoning to overcome context limitations. By progressively analyzing ranked subsets of code elements, the approach identifies domain concepts and refines domain boundaries without requiring full-system context. Our approach achieves high F1-scores on a dataset of ten projects, each comprising a curated domain model and its corresponding implementation, while remaining fully executable on locally deployable LLMs. This makes it particularly suitable for reverse engineering tasks in privacy-sensitive industrial environments.
Chinese Translation
大语言模型（LLMs）最近在代码理解方面展现出强大能力，这使它们有望用于从源代码逆向工程领域模型。然而，最先进的专有 LLM 由于隐私和保密限制在许多工业场景中无法使用，而能够在本地运行的紧凑开源 LLM 受其上下文窗口的限制，无法直接处理大型代码库。在本文中，我们提出一种使用轻量级、可本地部署的 LLM 从源代码中提取领域模型的自动化方法。我们的方法将结构启发式和语义启发式与基于 LLM 的迭代推理相结合，以克服上下文限制。通过逐步分析排序后的代码元素子集，该方法识别领域概念并细化领域边界，而无需完整的系统上下文。我们的方法在一个包含十个项目的数据集上取得了较高的 F1 分数，每个项目都包含一个精心构建的领域模型及其对应实现，同时仍然完全可在本地部署的 LLM 上执行。这使得它特别适用于隐私敏感的工业环境中的逆向工程任务。
cs.SE / 70 / 2608.12311
The Role Specialization Model (RSM): Coordinating LLM-Based Tools in Agentic Software Development - An Exploratory Case Study
角色专业化模型（RSM）：在智能体化软件开发中协调基于大语言模型的工具——一项探索性案例研究
large language model
大语言模型相关
Abstract
The integration of large language models (LLMs) into software development workflows has given rise to a paradigm known as Agentic Software Engineering (SE 3.0), in which autonomous agents manage full development life cycles under human supervision. This paper presents an exploratory case study in which three LLM-based tools, Antigravity (an agentic IDE with a Gemini 2.5 backend), Gemini CLI, and Qwen Code (local execution via Ollama), are coordinated according to a role-distribution framework proposed in this work as the Role Specialization Model (RSM). Three research questions guide the study: (RQ1) how can LLM-based tools with distinct capabilities be coordinated through the RSM in a real development workflow; (RQ2) what deviations from the planned role distribution emerge during RSM execution and what factors explain them; and (RQ3) how does the resulting product compare against the ISO/IEC 25010 quality model. The objective was to propose the RSM with the incremental development of a Python desktop application for interactive climate-data visualization. The workflow, observed deviations, prompt-hardening techniques, and a qualitative quality assessment are documented. Results suggest that explicit role coordination can support development cycle organization and architectural quality, but requires deliberate coordination strategies, context management, and human verification of agent-generated outputs.
Chinese Translation
将大语言模型（LLMs）集成到软件开发工作流中，催生了一种被称为智能体化软件工程（SE 3.0）的范式，其中自主智能体在人类监督下管理完整的开发生命周期。本文介绍了一项探索性案例研究，其中三个基于大语言模型的工具——Antigravity（一个采用 Gemini 2.5 后端的智能体集成开发环境）、Gemini CLI 和 Qwen Code（通过 Ollama 本地执行）——根据本文提出的角色分配框架（即角色专业化模型，RSM）进行协调。三个研究问题指导本研究：（RQ1）在真实开发工作流中，如何通过 RSM 协调具有不同能力的基于大语言模型的工具；（RQ2）在 RSM 执行过程中会出现哪些偏离计划角色分配的情况，以及哪些因素可以解释这些偏离；（RQ3）最终产品与 ISO/IEC 25010 质量模型相比如何。研究目标是通过增量式开发一个用于交互式气候数据可视化的 Python 桌面应用程序来提出 RSM。文中记录了工作流、观察到的偏离、提示加固技术以及定性质量评估。结果表明，明确的角色协调能够支持开发周期的组织和架构质量，但需要有意识的协调策略、上下文管理以及对智能体生成输出的人工验证。
cs.AI / 71 / 2608.12197
NetlistBench: Evaluating LLM Reliability in SPICE Netlist Recognition and Manipulation
NetlistBench：评估大语言模型在 SPICE 网表识别与操作中的可靠性
large language model
大语言模型相关
Abstract
Large Language Models (LLMs) are increasingly used in circuit design workflows, yet their reliability on simulator-facing SPICE netlist recognition and manipulation remains poorly understood and is rarely separated from high-level design reasoning. Although netlists are textual, they encode structured circuit objects through topology and parameters. We present \textbf{NetlistBench}, a structure-verified benchmark for SPICE netlist recognition and manipulation. NetlistBench contains 2,342 cases across 24 task families, covering parameter and connectivity recognition and edits, hierarchical operations, equivalence judgment, and long-horizon compound editing. Model outputs are evaluated by a deterministic structure-aware oracle. Across six non-thinking LLMs, performance varies substantially with operation-level structural complexity. Simple local edits reach $96\%$--$100\%$ accuracy, while device addition drops to $41\%$--$83\%$ and equivalence judgment to $49\%$--$90\%$. Enabling reasoning substantially improves weaker models but does not eliminate structure-preservation failures, with performance still degrading sharply as the edit horizon increases. NetlistBench identifies netlist reliability as a distinct bottleneck for trustworthy LLM-based circuit design automation.
Chinese Translation
大语言模型（LLMs）越来越多地用于电路设计流程，但它们在面向仿真器的 SPICE 网表识别与操作方面的可靠性仍然了解不足，并且很少与高层设计推理区分开来。尽管网表是文本形式的，但它们通过拓扑和参数编码了结构化的电路对象。我们提出 \textbf{NetlistBench}，一个面向 SPICE 网表识别与操作的结构验证基准。NetlistBench 包含 2,342 个用例，横跨 24 个任务族，涵盖参数与连接关系的识别和编辑、层次化操作、等价性判断以及长时域复合编辑。模型输出由一个确定性的结构感知 oracle 进行评估。在六个非思维型大语言模型上，性能随操作层面的结构复杂性而显著变化。简单的局部编辑达到 $96\%$--$100\%$ 的准确率，而器件添加降至 $41\%$--$83\%$，等价性判断降至 $49\%$--$90\%$。启用推理可以大幅提升较弱模型的表现，但并不能消除结构保持方面的失败，随着编辑视界的增大，性能仍会急剧下降。NetlistBench 将网表可靠性确定为基于 LLM 的可信电路设计自动化中的一个独特瓶颈。
cs.AI / 72 / 2608.11510
Conflict and Congruency Effects in Large Language Models: In-Weight and In-Context Competition in a Verbal Conflict Task
大语言模型中的冲突与一致性效应：一项言语冲突任务中的权重内与上下文内竞争
large language model
大语言模型相关
Abstract
Congruency effects, observed in conflict tasks such as Stroop and flanker tasks, have been investigated for nearly a century in psychology and neuroscience, but their mechanistic basis is not fully understood. We introduce a verbal-only LLM conflict task in which a prompt stem elicits a default same-color completion and an explicit rule either agrees with (congruent condition) or conflicts with (incongruent condition) the completion. Gemma-2-2B and six Pythia models ranging from 410M to 12B parameters showed strong default same-color tendencies, and six of seven models showed strong congruency effects. Using causal attribution analysis, attention analysis, and attention ablations, we identified distinct processing pathways in these LLMs: a pathway involving short-range attention to a superficial color cue that is preferentially activated in the congruent condition, and a pathway involving long-range attention to the rule prefix that is preferentially activated in the incongruent condition. Fine-tuning that strengthened the default same-color tendency had divergent effects on task conditions, reducing incongruent performance while increasing congruent performance. In contrast, increasing rule set size selectively impaired incongruent performance. These converging findings support an account in which congruency effects in this task arise from competition between an in-weight default mapping and an in-context rule-based mapping. More broadly, our findings illustrate how LLMs can serve as model systems for mechanistic analysis of competition between default and rule-governed response tendencies within a single learned network.
Chinese Translation
在诸如Stroop任务和flanker任务等冲突任务中观察到的一致性效应，已在心理学和神经科学中被研究了近一个世纪，但其机制基础尚未被完全理解。我们引入了一个纯言语的大语言模型冲突任务，其中提示词主干会引出一个默认的同色补全，而一条明确规则与该补全一致（一致条件）或冲突（不一致条件）。Gemma-2-2B 以及参数量从 410M 到 12B 的六个 Pythia 模型表现出强烈的默认同色倾向，并且七个模型中有六个表现出强烈的一致性效应。通过因果归因分析、注意力分析和注意力消融，我们在这些大语言模型中识别出不同的处理通路：一条通路涉及对表面颜色线索的短程注意力，并在一致条件下被优先激活；另一条通路涉及对规则前缀的长程注意力，并在不一致条件下被优先激活。强化默认同色倾向的微调对任务条件产生了不同效应：降低不一致条件下的表现，同时提高一致条件下的表现。相比之下，增加规则集大小选择性地损害了不一致条件下的表现。这些汇聚的发现支持这样一种解释：该任务中的一致性效应源于权重内默认映射与上下文内基于规则的映射之间的竞争。更广泛地说，我们的发现说明了大语言模型如何能够在单一已学习网络内作为模型系统，用于对默认反应倾向与规则支配的反应倾向之间的竞争进行机制分析。
cs.LG / 73 / 2608.11475
Probing and steering biology across Boltz-1s trunk-diffusion boundary
跨越 Boltz-1 主干-扩散边界的生物学探测与引导
diffusion
扩散模型相关
Abstract
AlphaFold3-class structure predictors pair a representational trunk, which processes sequence and context, with a diffusion module, which generates atomic coordinates. How biological information changes as it crosses this architectural boundary remains poorly understood. We analyze per-residue activations from the Pairformer trunk and diffusion module of Boltz-1 using linear probes, sparse autoencoders (SAEs), and causal interventions. From the trunk, both geometry (secondary structure, disorder) and sequence chemistry (amino-acid identity, signal peptides, disulfide-bond annotations) are linearly decodable. In the diffusion module, the two diverge. Secondary structure transfers essentially unchanged, whereas sequence chemistry is strongly attenuated. We then test whether decodable directions can steer the model, intervening on the final trunk single representation that conditions the diffusion module. Helix and coil directions change predicted structure dose-dependently against matched-norm random controls, but a beta-strand direction that is highly predictive (F1 =0.82) produces no measurable increase in strand content: linear decodability does not imply causal influence at the site we tested. The same probes also score markedly lower against sparse SwissProt annotations than against dense DSSP labels, because unannotated residues that the model gets right are charged as false positives; such scores are therefore lower bounds. Finally, supervised probes outscore single SAE features wherever a label already exists. We release the trained trunk and diffusion SAEs, Boltz-1 per-residue activations, and the analysis code.
Chinese Translation
AlphaFold3 类结构预测器将一个处理序列和上下文信息的表示主干与一个生成原子坐标的扩散模块配对使用。生物信息在跨越这一架构边界时如何变化，目前仍知之甚少。我们使用线性探针、稀疏自编码器（SAE）和因果干预，分析了 Boltz-1 的 Pairformer 主干和扩散模块中的逐残基激活。来自主干的信息中，几何特征（二级结构、无序区域）和序列化学特征（氨基酸身份、信号肽、二硫键注释）都是线性可解码的。在扩散模块中，这两类信息发生了分化。二级结构几乎原样传递，而序列化学特征则被强烈衰减。随后，我们测试可解码方向是否能够引导模型，具体做法是对用于调节扩散模块的最终主干单一表示进行干预。与匹配范数的随机对照相比，螺旋和卷曲方向会以剂量依赖的方式改变预测结构；但一个具有高预测能力的 β-折叠方向（F1=0.82）并未产生可测量的折叠含量增加：线性可解码性并不意味着在我们测试的位置上具有因果影响。同样的探针在稀疏的 SwissProt 注释上的得分也显著低于在密集的 DSSP 标签上的得分，因为模型预测正确但未被注释的残基会被计为假阳性；因此这类分数只是下界。最后，只要已有标签存在，监督式探针的得分就优于单个 SAE 特征。我们发布了训练好的主干和扩散 SAE、Boltz-1 逐残基激活以及分析代码。
cs.CL / 74 / 2608.12283
Large Language Model-Driven Small-Capitalization Trading: Integrating Financial News Sentiment, Macroeconomic Indicators, and Technical Signals
大语言模型驱动的小市值交易：整合财经新闻情绪、宏观经济指标与技术信号
large language model
大语言模型相关
Abstract
Large language models can extract richer signals from financial news than fixed sentiment lexicons, and recent work has explored feeding such signals into portfolio construction. We study an uncertainty-aware construction that feeds model-predicted risk -- decomposed into aleatoric and epistemic components -- directly into the covariance matrix of portfolio allocators, rather than treating portfolio risk as fixed or adjusting only expected returns. We evaluate the pipeline on Russell 2000 equities under three stock-selection regimes: a pure-alpha trigger that isolates abnormal stock moves not explained by macro indicators, a pure-beta trigger that captures macro-indicator moves before the stock itself fires, and a beta trigger in which both channels agree. Across the full holding-period grid, the separated pure-alpha and pure-beta legs usually dominate the beta intersection on Sharpe and return. Two horizons are especially informative. At one day, pure beta can work under low and moderate transaction costs because it captures immediate lead-lag spillovers from liquid macro and sector indicators into exposed small-cap stocks, but this advantage disappears at 100 bps when turnover and microstructure noise dominate. At 40 days, pure beta works for a different reason: slower macro repricing overtakes the firm-specific pure-alpha channel. The strongest conservative row is pure beta with GPT-4o mini sentiment, a Student-t target, a 40-day holding period, and risk parity allocation, reaching Sharpe 2.33 at 100 bps. The results suggest that stock-selection regime and allocator choice matter at least as much as the sentiment model, and that separating firm-specific and macro-exposure triggers is more informative than requiring both to fire simultaneously.
Chinese Translation
大语言模型能够从财经新闻中提取比固定情绪词典更丰富的信号，近期研究已探索将此类信号用于投资组合构建。我们研究了一种不确定性感知的构建方法，将模型预测的风险——分解为偶然不确定性（aleatoric）和认知不确定性（epistemic）成分——直接输入投资组合配置器的协方差矩阵，而不是将组合风险视为固定不变，或仅调整预期收益。我们在罗素2000成分股上评估该流程，并设置三种选股机制：纯阿尔法触发条件，用于分离出宏观经济指标无法解释的异常个股波动；纯贝塔触发条件，用于在个股自身发出信号之前捕捉宏观经济指标的变动；以及双渠道一致的贝塔触发条件。在整个持有期网格上，分离的纯阿尔法和纯贝塔分支通常在夏普比率和收益上优于贝塔交叉组合。两个持有期尤其具有信息价值。在持有期为1天时，纯贝塔策略在较低和中等交易成本下能够有效，因为它捕捉到了流动性较高的宏观和行业指标向暴露的小市值股票传导的即时领先-滞后溢出效应，但在100个基点成本下，当换手率和微观结构噪声占据主导时，这一优势会消失。在持有期为40天时，纯贝塔策略有效的原因不同：较慢的宏观重新定价超过了公司层面的纯阿尔法渠道。表现最强的保守配置行是采用GPT-4o mini情绪、Student-t目标、40天持有期和风险平价配置的纯贝塔策略，其在100个基点成本下夏普比率达到2.33。结果表明，选股机制和配置器选择的重要性至少与情绪模型相当，而且将公司特定触发条件与宏观敞口触发条件分开，比要求两者同时触发更具信息价值。
cs.AI / 75 / 2608.11911
Hamilton-Zero: A Neural Tensor-Network Foundation Model for Ground States of Arbitrary Quadratic Qubit Hamiltonians
Hamilton-Zero：用于任意二次量子比特哈密顿量基态的神经张量网络基础模型
large language model
大语言模型相关
Abstract
A central promise of useful quantum advantage is the ability to compute ground states of Hamiltonian systems beyond the reach of classical simulation methods. Here we demonstrate that this problem can be effectively amortized across an arbitrary and universal set of Hamiltonians by a foundation model with $\sim0.5$B variational parameters, trained with contemporary techniques from large language models and deep reinforcement learning. To do this, we formulate $\text{spin-}1/2$ quantum ground-state learning as manifold variational optimisation over centrally odd scalar functions on $\mathrm{SU}(2)^N$. This replaces explicit Hilbert-space vector amplitudes with manifold functions on which the Hamiltonian acts through Lie derivatives, evaluated by custom automatic differentiation primitives. We prove that the resulting variational principle on this manifold preserves the $\text{spin-}1/2$ sector's ground-state upper bound using the Peter-Weyl theorem, then pre-train our foundation model on a dataset of hundreds of thousands of different Hamiltonian systems, varying the connection topology, system size, interaction types and strengths, bringing together a century of many-body literature. Using a novel $\mathrm{SU}(2)$ replica-exchange Langevin sampler and sharded natural-gradient optimisation, we train our model with our own extension of the Kronecker-Factored Approximate Curvature (KFAC) optimiser on system sizes up to 64 qubits. On a held-out generalisation dataset, we fine-tune our model on system sizes of up to 1024 qubits, and evaluate on systems up to 8100 qubits.
Chinese Translation
实用量子优势的一个核心前景是能够计算超出经典模拟方法能力范围的哈密顿量系统的基态。在这里，我们证明该问题可以通过一个具有 $\sim0.5$B 变分参数的基础模型，在任意且通用的哈密顿量集合上被有效摊销；该模型使用来自大语言模型和深度强化学习的当代技术进行训练。为此，我们将 $\text{spin-}1/2$ 量子基态学习表述为在 $\mathrm{SU}(2)^N$ 上中心奇标量函数的流形变分优化。这用流形函数取代了显式的希尔伯特空间向量振幅，哈密顿量通过李导数作用于这些流形函数，并由定制的自动微分原语进行求值。我们利用 Peter-Weyl 定理证明，该流形上所得的变分原理保持了 $\text{spin-}1/2$ 扇区的基态上界；然后，我们在一个包含数十万个不同哈密顿量系统的数据集上对我们的基础模型进行预训练，这些系统在连接拓扑、系统尺寸、相互作用类型和强度上各不相同，汇集了一个世纪以来的多体文献。利用一种新颖的 $\mathrm{SU}(2)$ 副本交换朗之万采样器和分片自然梯度优化，我们在最高达 64 量子比特的系统尺寸上，使用我们自己对 Kronecker-Factored Approximate Curvature（KFAC）优化器的扩展来训练我们的模型。在一个留出的泛化数据集上，我们在最高达 1024 量子比特的系统尺寸上对模型进行微调，并在最高达 8100 量子比特的系统上进行评估。
人工智能 (cs.AI)
75
cs.AI / 1 / 2608.11323
Deployment Decision Reliability: A Generalizability-Theory Framework for Sizing Long-Horizon Agent Evaluations
部署决策可靠性：一种用于确定长时程智能体评估规模的概化理论框架
Abstract
Enterprise practitioners read agent leaderboards as if they ranked agent capability. We show, across three open agent-trace benchmarks (TheAgentCompany, $τ^2$-bench, and AppWorld), that the agent main effect accounts for less than 3% of total variance in every dataset and check type, while the agent-by-task interaction accounts for 7-23%. Leaderboards rank specialization, not capability. We arrive at this through a four-facet Generalizability Theory variance decomposition, fit with three estimators (Henderson Method-I, REML via lme4, and a Bayesian binomial GLMM) that agree to three decimal places. Four further findings sharpen what the leaderboard is hiding. First, aggregate reliability collapses on the hardest task quartile: $Eρ^2$ on $τ^2$ action_checks falls from 0.752 to 0.000. Second, training-cell reliability negatively correlates with held-out reliability ($r = -0.90$ on $τ^2$), meaning the designs that look most reliable replicate worst. Third, population-level diagnostics transfer across enterprise benchmarks (capability-gap ratio stable at 0.35-0.40) but per-family agent rankings invert. Fourth, on the MAST failure taxonomy, trace-level mode profiles are idiosyncratic (MAE = 0.261) while cell-level profiles generalise (MAE = 0.056, $r = 0.83$). We package these into Deployment Decision Reliability (DDR), a one-page reporting discipline that turns the variance-component table into five decisions an enterprise buyer can defend. All code, data loaders, and fit artifacts are released under an open-source license.
Chinese Translation
企业从业者阅读智能体排行榜，仿佛它们对智能体能力进行了排名。我们表明，在三个开放智能体轨迹基准（TheAgentCompany、$τ^2$-bench 和 AppWorld）中，在每个数据集和检查类型上，智能体主效应占总方差的比例不到3%，而智能体与任务的交互效应占7-23%。排行榜对专长进行排名，而非能力。我们通过一个四层面概化理论方差分解得出这一结论，并使用三种估计方法（Henderson Method-I、通过 lme4 的 REML 以及贝叶斯二项 GLMM）进行拟合，它们在小数点后三位一致。另外四项发现进一步揭示了排行榜所隐藏的内容。第一，总体信度在最难任务四分位上崩溃：$τ^2$ action_checks 上的 $Eρ^2$ 从 0.752 降至 0.000。第二，训练单元信度与留出信度呈负相关（在 $τ^2$ 上 $r = -0.90$），这意味着看起来最可靠的设计在复制中表现最差。第三，总体层面的诊断指标可跨企业基准迁移（能力差距比率稳定在 0.35-0.40），但按系列划分的智能体排名会发生反转。第四，在 MAST 失败分类体系上，轨迹层面的模式特征是特异的（MAE = 0.261），而单元层面的特征具有可推广性（MAE = 0.056，$r = 0.83$）。我们将这些整合为部署决策可靠性（DDR），这是一种一页纸的报告规范，将方差分量表转化为企业买方能够为其辩护的五项决策。所有代码、数据加载器和拟合产物均以开源许可证发布。
cs.AI / 2 / 2608.11341
Apodex Discovery: Reality Benchmarks and Environments for Evaluating and Building Discoverative Artificial Intelligence
Apodex Discovery：用于评估和构建发现型人工智能的现实基准与环境
Abstract
Apollo did not reach the Moon merely because its engineers could solve difficult equations. It succeeded by turning a distant ambition into a mission architecture of explicit objectives, simulation, verification, and repeated correction. AI now faces a similar transition: frontier models can solve difficult tasks once the problem, tools, and success criteria are specified, yet consequential real-world challenges rarely arrive in an executable or verifiable form. We introduce Apodex Discovery, a framework for building and evaluating discoverative AI through the heavy-duty solver, a system comprising a foundation model, harness, tools, and control policies that pursues extended, stateful, verifiable investigations. It has three core components. First, a problem-scouting process surveyed 561 industries across 16 sectors, assembled 423 high-value real-world problems, and selected 20 for the initial release. Second, a common environment-task-episode abstraction provides data, tools, constraints, feedback, trajectory recording, and verification of intermediate artifacts and final submissions. Third, HDS6 evaluates Tools, Repair, Alternatives, Coherence, Evidence, and Scope independently of final-task success. In AAV capsid design, Apodex surpassed the published state of the art by 7% across viability, tropism, structure prediction, and generative design. In drug repurposing and reformulation, a task-specific biomedical environment improved the mean normalized prediction score of GPT-5.5 and GPT-5.6-sol by 2.5 and 7.6 points over the same closed-book backbone. Controlled ablations show that the fixed TRACES episode interface enables attribution of performance differences to specific solver components. Apodex Discovery moves AI evaluation beyond predefined benchmarks toward verifiable investigations aimed at genuine discovery.
Chinese Translation
阿波罗之所以能够抵达月球，并不仅仅因为其工程师能够求解困难的方程。它之所以成功，是因为将一个遥远的抱负转化为由明确目标、仿真、验证和反复修正组成的任务架构。人工智能如今正面临类似的转型：一旦问题、工具和成功标准被明确给定，前沿模型就能解决困难任务，但具有重大影响的现实挑战很少以可执行或可验证的形式出现。我们提出 Apodex Discovery，一个通过重载求解器来构建和评估发现型人工智能的框架；该重载求解器由基础模型、执行框架、工具和控制策略组成，致力于开展持续、有状态且可验证的调研。该框架包含三个核心组成部分。第一，问题搜寻过程调研了跨越 16 个领域的 561 个行业，汇集了 423 个高价值现实问题，并选出 20 个作为初始发布。第二，统一的环境—任务—回合抽象提供数据、工具、约束、反馈、轨迹记录，以及对中间产物和最终提交的验证。第三，HDS6 独立于最终任务成功与否，对工具、修复、替代方案、一致性、证据和范围进行评价。在 AAV 衣壳设计中，Apodex 在活性、组织嗜性、结构预测和生成式设计方面比已发表的最先进水平高出 7%。在药物再利用和再配方方面，一个任务特定的生物医学环境使 GPT-5.5 和 GPT-5.6-sol 的平均归一化预测得分相对于相同的闭卷主干分别提高了 2.5 分和 7.6 分。受控消融实验表明，固定的 TRACES 回合接口能够将性能差异归因于特定的求解器组件。Apodex Discovery 将 AI 评估从预定义基准推向旨在实现真正发现的可验证调研。
cs.AI / 3 / 2608.11354
Inverse Theory of Mind Modeling for Content Recommendation: From Web Browsing to Dynamic Intelligent Interfaces
面向内容推荐的反向心智理论建模：从网页浏览到动态智能界面
Abstract
Modern recommender systems treat observed actions as reliable proxies for user preferences, yet interactions often reflect exploration or comparison rather than stable preference expression. As interfaces evolve from static layouts toward generative UIs and immersive extended reality (XR), the need for deeper, modality-agnostic user understanding grows: these adaptive environments must decide not only what to present but where, when, how prominently, and most importantly why a user acts. We propose an Inverse Theory of Mind (IToM) pipeline that reasons backward from observed interactions to infer the beliefs, preferences, and decision-making traits that explain behavior. The pipeline reconstructs each user's decision context, including what was chosen and what alternatives were available, applies LLM-driven counterfactual reasoning to produce evidence-grounded natural-language belief statements, and synthesizes these beliefs through multi-hypothesis abductive inference into a structured user persona. We evaluate on the OPeRA dataset against ground-truth personality assessments, attitudinal surveys, and interview-based personas across four tasks: next action prediction, shopping attitude alignment, Big Five personality inference, and held-out category prediction. Results show that inferred personas match or exceed ground-truth personas and that multi-hypothesis reasoning is essential for accurate personality prediction. We further demonstrate cross-modal transferability with a persona-driven spatial banking application on VisionOS.
Chinese Translation
现代推荐系统将观察到的行为视为用户偏好的可靠代理，然而交互往往反映的是探索或比较，而非稳定的偏好表达。随着界面从静态布局向生成式用户界面和沉浸式扩展现实（XR）演进，对更深入、与模态无关的用户理解的需求日益增长：这些自适应环境不仅必须决定呈现什么，还要决定在哪里、何时、以何种显著程度呈现，以及最重要的是，用户为什么会采取某种行为。我们提出了一种反向心智理论（IToM）流水线，它从观察到的交互出发进行逆向推理，以推断能够解释行为的信念、偏好和决策特征。该流水线重建每个用户的决策情境，包括选择了什么以及有哪些备选项，应用由大语言模型（LLM）驱动的反事实推理来生成基于证据的自然语言信念陈述，并通过多假设溯因推理将这些信念综合为结构化的用户画像。我们在OPeRA数据集上进行评估，以真实人格评估、态度调查和基于访谈的用户画像为基准，涵盖四项任务：下一步行为预测、购物态度对齐、大五人格推断和留出类别预测。结果表明，推断出的用户画像达到或超过了真实用户画像，并且多假设推理对于准确的人格预测至关重要。我们进一步通过在VisionOS上的一个由用户画像驱动的空间银行应用，展示了跨模态可迁移性。
cs.AI / 4 / 2608.11381
From Numbers to Judgment: Specialist LLM Agents and Reinforcement Learning for European Listed Real Estate
从数字到判断：面向欧洲上市房地产的专家型 LLM 智能体与强化学习
Abstract
We study whether the localized numerical operations and integrative judgments of financial analysis benefit from the same form of LLM specialization. Larix maps a 16-lens European listed-real-estate analysis framework to eight lens-aligned specialists; we compare a frontier LLM under monolithic versus specialist-decomposed prompting while holding the model, source evidence, task instructions, output schema, and scoring fixed. Across 19 firms spanning seven regulatory wrappers, decomposition improves the numerical-task aggregate by 15.8 percentage points but does not reliably improve, and can reduce, performance on judgment tasks, a pattern stable across four frozen-template dispatches; a single-agent control given the complete framework does not reproduce the numerical gain. Post-training Qwen3.5-9B with GRPO using task-aligned structured rewards then raises the development-split score by 12.0 points and the judgment aggregate by 14.2 points, with gains on all four sub-ceiling tasks; the gains transfer to unseen firms (+15.2 points overall; +40.4 on covenant stress) and to unseen regulatory wrappers (+4.3), with positive transfer on all three anti-memorization splits. Prompt-level decomposition thus improves modular numerical execution, whereas targeted parameter adaptation improves integrative financial judgment.
Chinese Translation
我们研究财务分析中的局部化数值运算和整合性判断是否受益于同一种形式的 LLM 专业化。Larix 将一个 16 视角的欧洲上市房地产分析框架映射为八个与视角对齐的专家；我们在整体式提示与专家分解式提示下比较前沿 LLM，同时保持模型、来源证据、任务指令、输出模式和评分不变。在涵盖七种监管包装的 19 家公司上，分解将数值任务综合得分提高了 15.8 个百分点，但并未可靠地改善判断任务表现，并且可能降低判断任务表现；这一模式在四次冻结模板分派中保持稳定；给定完整框架的单智能体对照未能重现该数值提升。随后，使用任务对齐的结构化奖励并通过 GRPO 对 Qwen3.5-9B 进行后训练，将开发划分得分提高 12.0 分，将判断综合得分提高 14.2 分，并在全部四个未达到上限的任务上取得提升；这些提升迁移至未见过的公司（总体 +15.2 分；契约压力方面 +40.4）和未见过的监管包装（+4.3），并且在所有三个防记忆划分上均实现正向迁移。因此，提示层面的分解改善了模块化数值执行，而针对性的参数适配则改善了整合性财务判断。
cs.AI / 5 / 2608.11403
When Self-Consistency Backfires: Majority Vote Hurts the Majority of Hard Science Problems for Small LLMs
当自洽性适得其反：多数投票损害了小规模 LLM 在大多数硬科学问题上的表现
Abstract
Self-consistency (SC) via majority vote is a widely used way to spend inference-time compute: sample N chains of thought, return the plurality answer. On the full GPQA Diamond benchmark (198 graduate-level science questions), majority voting reduces per-problem accuracy on a majority of problems for two instruction-tuned models from different families: 56.6% of problems for Qwen2.5-7B and 65.7% for Llama-3-8B, with Qwen the primary demonstration and Llama corroborating the direction from a near-chance baseline. The effect was pre-registered on a 151-problem confirmatory split after being observed on 47 exploratory problems, and all four confirmatory hypotheses passed. A grid oracle that routes each problem to the best N across {1, 2, 4, 8, 16, 32, 64} marks a theoretical upper bound 14 accuracy points above N = 1 for Qwen and 17 for Llama, an oracle bound requiring ground truth rather than a deployable method. No verifier-free gate reaches it: neither a plurality-agreement gate nor a token-entropy gate moves accuracy more than 0.002 from fixed-budget voting at N = 64. The mechanism is direct: confidence does not track correctness on these problems. In the highest-agreement bin the plurality answer is correct about half the time for Qwen, and for Llama that bin is less accurate than its lowest-agreement bin. We pre-register and confirm these findings on small instruction-tuned models; we do not test reasoning-native models, which we flag as the central open question.
Chinese Translation
通过多数投票实现的自洽性（SC）是一种广泛使用的推理时计算投入方式：采样 N 条思维链，返回得票最多的答案。在完整的 GPQA Diamond 基准（198 道研究生水平的科学问题）上，对于两个来自不同系列的指令微调模型，多数投票在大多数问题上降低了逐题准确率：Qwen2.5-7B 有 56.6% 的问题、Llama-3-8B 有 65.7% 的问题准确率下降，其中 Qwen 是主要演示模型，Llama 从接近随机水平的基线出发佐证了这一方向。该效应在 47 道探索性问题上被观察到后，在一个 151 道问题的验证划分上进行了预注册，且全部四个验证性假设均通过。一个网格预言机（grid oracle）将每个问题路由到 {1, 2, 4, 8, 16, 32, 64} 中最佳的 N，其给出的理论上界在 Qwen 上比 N = 1 高出 14 个准确率百分点，在 Llama 上高出 17 个百分点；这一预言机上界需要真实标签，而不是一种可部署的方法。没有任何无需验证器的门控方法能达到该上界：在 N = 64 时，无论是多数一致性门控还是 token 熵门控，相对于固定预算投票，准确率变化都不超过 0.002。机制是直接的：在这些问题上，置信度并不与正确性保持一致。在一致性最高的分组中，Qwen 的多数答案大约只有一半时候是正确的；而对 Llama 而言，该分组的准确率低于其一致性最低的分组。我们在小型指令微调模型上预注册并确认了这些发现；我们没有测试原生推理模型，并将其标记为核心开放问题。
cs.AI / 6 / 2608.11420
Social Chain of Thought: A Multi-Agent Architecture Grounded in Medical Differential Diagnosis Methodology
社会思维链：一种基于医学鉴别诊断方法论的多智能体架构
Abstract
Medical diagnostic reasoning is a high-impact use case for LLMs that carries significant implications for the health and wellbeing of users. When OpenAI (2026) reports that more than 5% of ChatGPT messages globally are healthcare-related, the transparency of these systems becomes a serious design concern. This is especially true for complex cases, where differential diagnosis often requires integrating multiple forms of specialist reasoning. Existing work has proposed multi-agent approaches to medical diagnosis, but it remains unclear when such systems are needed, why they help, and where they outperform monolithic inference. We introduce Social Chain of Thought (SCoT),a multi-round pipeline for medical differential diagnosis that structures multi-agent interaction as a deliberative framework for collabora. tive LLM reasoning. Evaluating SCoT against single-agent baselines, one-agent pipeline ablations, and best-of-n scaling, we show that its recall advantage is not reproduced by monolithic inference alone. SCoT is most successful in the hardest diagnostic cases, where multiple rounds of specialist conversation help recover ground-truth diagnoses and converge on a higher-recall differential.
Chinese Translation
医学诊断推理对LLM来说是一个高影响用例，对用户的健康和福祉具有重大影响。当OpenAI（2026）报告称全球ChatGPT消息中超过5%与医疗保健相关时，这些系统的透明度就成为一个严肃的设计问题。对于复杂病例尤其如此，因为鉴别诊断通常需要整合多种形式的专科推理。已有工作提出了用于医学诊断的多智能体方法，但仍不清楚此类系统在何时需要、为何有帮助，以及在哪些情况下优于单体式推理。我们提出社会思维链（SCoT），这是一种用于医学鉴别诊断的多轮流水线，将多智能体交互构建为协作式LLM推理的审议框架。通过将SCoT与单智能体基线、单智能体流水线消融和best-of-n扩展进行评估，我们表明其召回率优势不能仅靠单体式推理再现。SCoT在最困难的诊断病例中最为成功；在这些病例中，多轮专科对话有助于恢复金标准诊断，并收敛到召回率更高的鉴别诊断。
cs.AI / 7 / 2608.11434
Benchmarking LLM Judges for Mobile Agent Evaluation
面向移动智能体评估的LLM评判器基准测评
Abstract
Mobile agent benchmarks increasingly rely on LLM-based judges to evaluate task completion, yet the reliability of these judges on mobile agent trajectories remains largely unexamined. We introduce MobileJudgeBench, a benchmark for systematically evaluating LLM-as-judge methods on mobile agent trajectories. Our benchmark comprises 931 human-annotated trajectories spanning 6 mobile agent benchmarks, 4 agent models, and 68 apps. We evaluate 6 judge methods (five adapted from SPA-Bench, A3 with two modes, AndroidArena, and AgentRewardBench, plus a simple baseline we design) across multiple LLM backends. Our experiments reveal three key findings. First, a simple baseline judge with sampled screenshots is competitive with, and often exceeds, purpose-built methods, indicating that more elaborate judge pipelines do not consistently improve judge quality; among competitive methods, the LLM backbone is the primary driver. Second, benchmark quality metrics reliably predict real-world judge utility: they correlate with both agent ranking fidelity for evaluation and downstream performance when judges serve as reward signals for on-policy reinforcement learning. Third, failure analysis across two LLM backends uncovers qualitatively opposite failure profiles, one conservative and the other permissive, linked to the backbone's precision-recall characteristics.
Chinese Translation
移动智能体基准测试越来越依赖基于LLM的评判器来评估任务完成情况，然而这些评判器在移动智能体轨迹上的可靠性在很大程度上仍未得到检验。我们提出了MobileJudgeBench，一个用于在移动智能体轨迹上系统评估LLM作为评判器方法的基准测试。我们的基准包含931条人工标注轨迹，涵盖6个移动智能体基准、4个智能体模型和68个应用。我们在多个LLM后端上评估了6种评判器方法（其中五种分别改编自SPA-Bench、具有两种模式的A3、AndroidArena和AgentRewardBench，另加一个我们设计的简单基线方法）。我们的实验揭示了三个关键发现。第一，一个使用采样截图的简单基线评判器与专用方法相比具有竞争力，并且经常超过专用方法，这表明更复杂的评判器流水线并不能持续提升评判质量；在有竞争力的方法中，LLM主干模型是主要驱动因素。第二，基准质量指标能够可靠地预测真实场景下的评判器效用：它们既与评估中的智能体排名保真度相关，也与评判器作为同策略强化学习奖励信号时的下游性能相关。第三，对两个LLM后端进行的失败分析揭示了性质相反的失败特征：一个偏保守，另一个偏宽松，这与主干模型的精确率-召回率特性有关。
cs.AI / 8 / 2608.11483
A Modular Agentic Framework for Synthetically Constrained Multi-Objective Hit-to-Lead Optimization
一种面向合成约束条件下多目标Hit-to-Lead优化的模块化智能体框架
Abstract
Hit-to-lead optimization requires iterative design of hit analogs across competing potency, selectivity, physicochemical, pharmacokinetic, safety, and synthetic constraints. We present SABLE (Synthetically-accessible Agentic Bayesian Ligand Exploration), an open-source framework that employs natural-language orchestration to guide chemical structure optimization. SABLE uses an LLM to interpret user-defined goals and route tasks, while specialized tools perform reaction-templated analog enumeration, physicochemical and ADMET property prediction, structure-based affinity scoring, and Bayesian optimization. The resulting workflow is a computational twin of the analytical and prioritization stages of the design-make-test-analyze cycle, providing provenance of each numerical output. Across single, and multi-objective optimization studies, SABLE enriches candidate sets for user-defined computational objectives while evaluating only a subset of the enumerated search space. Its modular architecture allows tools and characterization backends to be replaced by editing a simple config file, without modifying operational logic. SABLE provides an extensible decision-support framework for prioritizing synthetically constrained analogs in early-stage drug discovery.
Chinese Translation
Hit-to-Lead优化需要在相互制约的效力、选择性、理化性质、药代动力学、安全性和合成约束下，对命中类似物进行迭代设计。我们提出了SABLE（合成可及的智能体贝叶斯配体探索），这是一个采用自然语言编排来指导化学结构优化的开源框架。SABLE使用大语言模型（LLM）来解释用户定义的目标并分配任务，而专用工具则执行基于反应模板的类似物枚举、物理化学和ADMET性质预测、基于结构的亲和力打分以及贝叶斯优化。由此产生的工作流程是设计-合成-测试-分析循环中分析与优先级排序阶段的计算孪生，可提供每个数值输出的来源追溯。在单目标和多目标优化研究中，SABLE仅评估枚举搜索空间的一个子集，即可针对用户定义的计算目标富集候选集。其模块化架构允许通过编辑一个简单的配置文件来替换工具和表征后端，而无需修改操作逻辑。SABLE为早期药物发现中合成受限类似物的优先级排序提供了一个可扩展的决策支持框架。
cs.AI / 9 / 2608.11584
EnterpriseRAG: Benchmarking LLM Instruction Adherence and Robustness under Non-Ideal Enterprise Retrieval
EnterpriseRAG：在非理想企业检索条件下对LLM指令遵循性与鲁棒性的基准评测
Abstract
Enterprise RAG deployments face a critical reliability gap: while LLMs satisfy 80% of individual constraints, only 26.8% of responses meet all requirements simultaneously, revealing a 57-point orchestration gap. Existing benchmarks assume clean retrieval with simple queries, failing to capture production conditions where noisy documents and multi-dimensional constraints coexist. We introduce EnterpriseRAG, a benchmark of 983 expert-validated samples across six domains that systematically simulates three failure modes absent from prior work: retrieval noise, knowledge gaps, and factual conflicts, coupled with complex instructions. Evaluation of 13 state-of-the-art LLMs reveals a severe instruction adherence collapse, where high per-constraint satisfaction masks low holistic compliance. Critical findings expose deep barriers under knowledge gaps and factual conflicts, even with reasoning-enhanced inference, indicating production RAG requires explicit context-aware protocols and calibrated judgment. EnterpriseRAG provides a reproducible foundation for measuring and closing these gaps, directly informing deployment decisions for enterprise-scale RAG systems. We will release the benchmark and evaluation framework upon publication.
Chinese Translation
企业级RAG部署面临一个关键可靠性差距：虽然LLM能满足80%的单项约束，但只有26.8%的响应能同时满足所有要求，这暴露出57个百分点的编排差距。现有基准假设采用干净检索和简单查询，未能捕捉到生产环境中噪声文档与多维约束共存的情况。我们提出了EnterpriseRAG，这是一个涵盖六个领域、包含983个经专家验证样本的基准，它系统性地模拟了以往工作中缺失的三种失效模式：检索噪声、知识缺口和事实冲突，并与复杂指令相结合。对13个最先进LLM的评估揭示了严重的指令遵循性崩溃：较高的单项约束满足率掩盖了较低的整体遵循度。关键发现揭示了在知识缺口和事实冲突下存在的深层障碍，即使采用推理增强推断也是如此，这表明生产环境RAG需要显式的上下文感知协议和经过校准的判断。EnterpriseRAG为衡量和弥合这些差距提供了可复现的基础，并直接为企业级RAG系统的部署决策提供依据。我们将在论文发表后发布该基准和评估框架。
cs.AI / 10 / 2608.11588
CoAdapt-GUI: Joint Workflow Context and Policy Adaptation for Unseen GUI Applications
CoAdapt-GUI：面向未见 GUI 应用的联合工作流上下文与策略自适应
Abstract
Mobile GUI agents remain brittle when deployed to applications absent from source training. We study novel-app generalization under a limited target interaction budget and without target demonstrations. We introduce CoAdapt-GUI, a test-time adaptation (TTA) framework that jointly adapts structured workflow context and policy from the agent's own target-app rollouts and rewards. The workflow context retains transferable procedures, failure modes, and verification rules while excluding app-bound source details. This separation allows reusable workflow knowledge to guide adaptation without transferring source-interface state. For policy adaptation, task-context-matched group-relative optimization updates a LoRA adapter on a frozen vision-language model. Across two unseen-app evaluations, CoAdapt-GUI reaches 45.0% on AndroidWorld-Generalization, compared with 37.5% for the reported Policy-Only TTA baseline, and raises AndroidWorld Plus performance from 38.6% to 52.9%. These results show that transfer-constrained workflow context provides substantial gains and that joint policy adaptation further improves held-out performance.
Chinese Translation
当部署到源训练中不存在的应用程序时，移动 GUI 智能体仍然很脆弱。我们研究在有限的目标交互预算且没有目标演示的情况下，新应用程序的泛化能力。我们提出了 CoAdapt-GUI，一个测试时自适应（TTA）框架，它根据智能体自身在目标应用中的 rollout 和奖励，联合自适应结构化工作流上下文和策略。工作流上下文保留可迁移的过程、失败模式和验证规则，同时排除与源应用绑定的细节。这种分离使得可复用的工作流知识能够指导自适应，而不会迁移源界面状态。在策略自适应方面，任务上下文匹配的组相对优化在冻结的视觉语言模型上更新 LoRA 适配器。在两项未见应用评估中，CoAdapt-GUI 在 AndroidWorld-Generalization 上达到 45.0%，而所报告的仅策略 TTA 基线为 37.5%，并将 AndroidWorld Plus 的性能从 38.6% 提高到 52.9%。这些结果表明，受迁移约束的工作流上下文带来了显著增益，并且联合策略自适应进一步提升了留出应用上的性能。
cs.AI / 11 / 2608.11605
Foresight Without Seeing: Latent Futures for World Action Models
未见而预见：世界动作模型的潜在未来
Abstract
World Action Models (WAMs) couple future visual prediction with robot action generation, enabling policies to model how the physical world evolves during interaction. Existing WAMs differ in how predictive dynamics are exposed to the action pathway. Explicit-future WAMs provide direct access to predicted scene evolution, but incur substantial inference costs from iterative video denoising. In contrast, direct-policy WAMs efficiently predict actions from the current observation but lack an explicit inference-time interface for exposing predictive dynamics to the Action DiT. To bridge this gap, we propose ForeWAM, a dynamics-conditioned direct-policy WAM that provides predictive context for action generation without decoding future videos. At its core, Future-KV performs a single Video DiT prefill over the current visual latent and stochastic future slots, and reuses the resulting layer-wise key-value states throughout action denoising. We further introduce dynamics registers supervised by a frozen latent action teacher, encouraging the implicit future states to capture interaction-induced transitions such as object motion, contact changes, and task progress. Ground-truth future observations and the teacher are used only during training; deployment requires neither and performs no future video generation. Without embodied robot data pretraining, the standard and accelerated variants of ForeWAM achieve average success rates of 96.7% and 96.9% on LIBERO, respectively. The standard variant further achieves 61.6% success on LIBERO-Plus. These results demonstrate that direct-policy WAMs can retain efficient action prediction while exposing predictive dynamics to the action pathway without explicitly generating future observations.
Chinese Translation
世界动作模型（WAMs）将未来视觉预测与机器人动作生成结合起来，使策略能够建模物理世界在交互过程中如何演化。现有 WAMs 在预测动态如何暴露给动作通路方面存在差异。显式未来 WAMs 提供对预测场景演化的直接访问，但会因迭代视频去噪而产生大量推理成本。相比之下，直接策略 WAMs 能够从当前观测中高效预测动作，但缺乏在推理时将预测动态暴露给 Action DiT 的显式接口。为弥合这一差距，我们提出了 ForeWAM，一种以动态为条件的直接策略 WAM，它在不解码未来视频的情况下为动作生成提供预测上下文。其核心是 Future-KV，它对当前视觉潜变量和随机未来槽执行一次 Video DiT 预填充，并在整个动作去噪过程中重用所得的逐层键值状态。我们进一步引入由冻结的潜在动作教师监督的动态寄存器，以促使隐式未来状态捕获由交互引起的转变，例如物体运动、接触变化和任务进展。真实未来观测和该教师仅在训练期间使用；部署时二者均不需要，也不进行任何未来视频生成。在不使用具身机器人数据预训练的情况下，ForeWAM 的标准变体和加速变体在 LIBERO 上分别取得了 96.7% 和 96.9% 的平均成功率。标准变体还在 LIBERO-Plus 上取得了 61.6% 的成功率。这些结果表明，直接策略 WAMs 能够在保持高效动作预测的同时，将预测动态暴露给动作通路，而无需显式生成未来观测。
cs.AI / 12 / 2608.11625
Making AI-Generated Feedback Matter: From Provision to Student Enactment
让AI生成的反馈产生实效：从提供反馈到学生践行
Abstract
Feedback processes strongly influence student learning, yet their educational value depends on addressing two distinct challenges: providing high-quality, timely, and individualised feedback at scale, and supporting students to interpret, evaluate, and act on that feedback productively. Generative AI offers a credible means of addressing the provision challenge, but students' uptake of AI-generated feedback remains limited. We conducted a large-scale quasi-experimental sequential cohort study comparing three AI-mediated feedback workflows across 13,037 students and 51,296 student-authored resources. In Directed Feedback (n = 3,723), students received AI-generated feedback comments without structured support. In Self-Directed Feedback (n = 3,951), students could initiate optional AI-supported dialogue. In Enacted Feedback (n = 5,363), students were prompted to select feedback suggestions, evaluate their relevance, and engage in targeted AI-supported dialogue anchored to those selections. Enacted Feedback was associated with significantly higher uptake of AI-generated feedback, with an estimated probability of 26.2%, compared with 14.1% for Directed Feedback and 0.1% for Self-Directed Feedback. It was also associated with significantly higher self-assessment confidence and submitted-work quality than both comparison conditions. These findings suggest that the educational value of AI-generated feedback depends not only on the quality of feedback comments, but also on workflows that actively structure students' enactment of feedback literacy processes. The results have implications for the design of AI feedback systems that position learners as active participants in judgement, dialogue, and improvement rather than passive recipients of comments. Overall findings show that AI access alone is insufficient; purposeful workflow design is central to productive feedback use.
Chinese Translation
反馈过程深刻影响学生学习，但其教育价值取决于应对两个不同的挑战：大规模地提供高质量、及时且个性化的反馈，以及支持学生富有成效地解读、评估并依据反馈采取行动。生成式AI为解决提供反馈这一挑战提供了一种可信的手段，但学生对AI生成反馈的采纳仍然有限。我们开展了一项大规模准实验序贯队列研究，在13,037名学生和51,296份学生创作的资源中比较了三种由AI介导的反馈工作流程。在定向反馈（n = 3,723）中，学生收到AI生成的反馈评语，但没有结构化支持。在自主反馈（n = 3,951）中，学生可以发起可选的AI支持对话。在践行反馈（n = 5,363）中，学生被提示选择反馈建议、评估其相关性，并围绕这些选择开展有针对性的AI支持对话。践行反馈与显著更高的AI生成反馈采纳率相关，估计概率为26.2%，而定向反馈为14.1%，自主反馈为0.1%。与两个对照条件相比，它还伴随着显著更高的自我评估信心和提交作业质量。这些发现表明，AI生成反馈的教育价值不仅取决于反馈评语的质量，还取决于能够积极组织学生践行反馈素养过程的工作流程。这些结果对AI反馈系统的设计具有启示意义，即应将学习者定位为判断、对话和改进的主动参与者，而不是评语的被动接受者。总体研究结果表明，仅获得AI访问权限是不够的；有目的的工作流程设计对于富有成效地使用反馈至关重要。
cs.AI / 13 / 2608.11676
XBridge: Entity-Grounded Latent Bridge for Heterogeneous LLM Communication
XBridge：用于异构大语言模型通信的实体锚定隐空间桥
Abstract
Heterogeneous multi-agent LLM systems, where agents are powered by different model families, can outperform homogeneous configurations by reducing redundant reasoning patterns. Yet existing communication protocols either operate through text, discarding the sender's internal representations, or require architectural homogeneity for latent-level transfer. We identify the entity grounding problem in cross-architecture communication: cross-attention bridges that transfer continuous representations across different LLM families suffer from rare-token compression collapse, where entity identity is lost in the continuous bottleneck (bridge-only F1 ~30%). We propose XBRIDGE, a decode-free communication protocol that addresses this through two mechanisms. Lexical Anchor Mapping (LAM) maps the sender's original context tokens to the receiver's vocabulary, providing discrete entity anchors. A Latent Enrichment Bridge (LEB) lets the receiver query the sender's hidden states for contextual enrichment. The entity anchors ground the bridge's contextual signals to specific entities through the receiver's own self-attention. Across three model families (Llama, Qwen, and Mistral), seven benchmarks, and both communication directions, XBRIDGE outperforms text-based communication on all seven tasks for each model pair while achieving 11x lower latency, and in a same-architecture setting it also exceeds a KV-sharing baseline on six of seven tasks. LEB requires only 264M trainable parameters (3.8% of the receiver), is trained on a small balanced sample set, and adds negligible inference overhead.
Chinese Translation
异构多智能体大语言模型系统中，各智能体由不同的模型家族驱动，能够通过减少冗余推理模式而优于同构配置。然而，现有的通信协议要么通过文本进行，丢弃发送方的内部表示，要么需要架构同质性才能进行隐空间层面的传输。我们识别了跨架构通信中的实体锚定问题：在不同大语言模型家族之间传输连续表示的交叉注意力桥会遭受稀有词元压缩坍缩，即实体身份在连续瓶颈中丢失（仅桥接的 F1 ~30%）。我们提出了 XBRIDGE，一种无需解码的通信协议，通过两种机制解决这一问题。词法锚映射（LAM）将发送方原始上下文词元映射到接收方的词表，从而提供离散的实体锚点。隐空间增强桥（LEB）允许接收方查询发送方的隐藏状态以进行上下文增强。实体锚点通过接收方自身的自注意力机制，将桥的上下文信号锚定到特定实体上。在三个模型家族（Llama、Qwen 和 Mistral）、七个基准以及两个通信方向上，XBRIDGE 在每一对模型的所有七个任务上均优于基于文本的通信，同时实现 11 倍的更低延迟；在同架构设置下，它还在七个任务中的六个上超过了 KV 共享基线。LEB 仅需要 264M 可训练参数（占接收方的 3.8%），在一个小型平衡样本集上训练，并且增加的推理开销可以忽略不计。
cs.AI / 14 / 2608.11683
FrontierFinance: A Challenging Benchmark for Measuring Frontier Intelligence of Finance Agents
FrontierFinance：衡量金融智能体前沿智能的具有挑战性的基准
Abstract
AI agents are increasingly deployed for professional investment research, yet no benchmark captures the complexity of the full investor workflow. Existing benchmarks mainly target financial data extraction, a narrow slice that current models have largely saturated, while reference-based metrics and generic LLM-as-a-judge scoring fall short on the open-ended, long-form answers that real analyst queries demand. We introduce FrontierFinance, a fully open benchmark of 220 expert-crafted queries and 11,543 source-attributed rubrics spanning six crucial use cases across the full investor workflow. FrontierFinance is both broader and harder than existing public finance benchmarks. Evaluating frontier models and agent systems under a common harness restricted to publicly available data, we find that the tool harness, not the model alone, strongly shapes quality and efficiency; that Samaya's in-house system leads at 56.0%, ahead of the strongest frontier model (Claude Fable 5, 49.2%) at roughly 2.2x lower cost; and that the best open-weight model (Kimi K3, 46.4%) nearly matches the best proprietary model at 4.5x lower cost. Screening & Discovery and Sector, Industry & Macro remain the hardest use cases across all systems, where even the best systems reach only 33% and 39%. We make the dataset and grading code publicly available.
Chinese Translation
人工智能智能体日益被部署用于专业投资研究，然而目前没有基准能够捕捉完整投资者工作流的复杂性。现有基准主要针对金融数据提取，这是当前模型已基本饱和的一个狭窄部分，而基于参考的指标和通用的LLM-as-a-judge评分在真实分析师查询所需的开放式、长文本答案上存在不足。我们推出了FrontierFinance，一个完全开放的基准，包含220个专家编写的查询和11,543个来源标注的评分标准，涵盖完整投资者工作流中的六个关键用例。FrontierFinance比现有的公共金融基准更广泛且更难。在仅限于公开可用数据的统一测试框架下评估前沿模型和智能体系统，我们发现工具框架而非模型本身在很大程度上决定了质量和效率；Samaya的内部系统以56.0%领先，领先于最强前沿模型（Claude Fable 5，49.2%），且成本约低2.2倍；最佳开放权重模型（Kimi K3，46.4%）几乎以低4.5倍的成本达到最佳专有模型的水平。在所有系统中，筛选与发现以及板块、行业与宏观仍然是最难的用例，即使是最好的系统也仅达到33%和39%。我们将数据集和评分代码公开发布。
cs.AI / 15 / 2608.11692
HUGIN: Enhancing Vision-Language Planning for Autonomous Logistics Sorting
HUGIN：增强面向自主物流分拣的视觉语言规划
Abstract
Autonomous logistics sorting systems (ALSS) are an important industrial application of embodied AI, which requires joint planning over spatially disjoint camera views. We formulate this setting as Joint Multi-Scene Understanding (JMSU). With open-world visual understanding and task-planning capabilities, vision-language models (VLMs) are promising candidates for JMSU. However, directly applying existing VLMs to JMSU is non-trivial due to scarce cross-scene supervision and attention dispersion caused by long visual context in JMSU. To address these challenges, we propose HUGIN, a training framework with two complementary components. Endogenous Data Augmentation recombines verified atomic facts under operating constraints, while Global Context Ranking aligns the instruction representation more strongly with the complete visual context than with a partial visual context. To support ongoing research, we construct a high-quality industrial sorting dataset and benchmark named SortingBench from four layouts of autonomous logistics sorting systems. Across five open VLMs, HUGIN consistently outperforms matched baselines; for example, the accuracy on SortingBench of Qwen3-VL-8B increases from 63.6% to 78.8%. Additional experiments verify the effectiveness of each component and JMSU's spillover benefits in embodied tasks. Deployment tests involving more than 15,000 packages support the practical viability of VLM-based planning for autonomous logistics sorting.
Chinese Translation
自主物流分拣系统（ALSS）是具身AI的重要工业应用，需要在空间上不相交的相机视图之间进行联合规划。我们将这一设定形式化为联合多场景理解（JMSU）。凭借开放世界视觉理解和任务规划能力，视觉语言模型（VLM）是JMSU的有前景的候选方案。然而，将现有VLM直接应用于JMSU并非易事，因为JMSU中跨场景监督稀缺，且长视觉上下文导致注意力分散。为解决这些挑战，我们提出了HUGIN，一个具有两个互补组件的训练框架。内源性数据增强在操作约束下重组经过验证的原子事实，而全局上下文排序使指令表示与完整视觉上下文之间的对齐强于其与部分视觉上下文之间的对齐。为了支持持续研究，我们基于自主物流分拣系统的四种布局构建了一个高质量的工业分拣数据集和基准，名为SortingBench。在五个开放VLM上，HUGIN始终优于匹配的基线；例如，Qwen3-VL-8B在SortingBench上的准确率从63.6%提升到78.8%。额外实验验证了每个组件的有效性以及JMSU在具身任务中的溢出效益。涉及超过15,000个包裹的部署测试支持了基于VLM的自主物流分拣规划的实际可行性。
cs.AI / 16 / 2608.11724
Proportional Analogies on Probability Distributions via Bayesian Updating
基于贝叶斯更新的概率分布比例类比
Abstract
Analogies are quaternary relations of the form "A is to B as C is to D". Among the various formalizations of analogical reasoning, proportional analogies provide an important axiomatic framework by characterizing valid analogies through a set of postulates. While proportional analogies have been extensively studied over Boolean, symbolic, and real-valued domains, their extension to probability distributions remains largely unexplored. In this paper, we introduce a notion of proportional analogy for probability distributions based on Bayesian updating. Our approach builds upon the idea that two distributions are related whenever one can be transformed into the other through Bayesian updating induced by a suitable set of observations. We investigate this framework for several standard members of the exponential family and discuss how it naturally extends to arbitrary probability distributions through Gaussian mixture approximations.
Chinese Translation
类比是“A之于B如同C之于D”这种形式的四元关系。在类比推理的各种形式化中，比例类比通过一组公设来刻画有效类比，提供了一个重要的公理化框架。尽管比例类比在布尔、符号和实值域上已得到广泛研究，但它们在概率分布上的扩展在很大程度上仍未得到探索。本文中，我们基于贝叶斯更新引入了一个针对概率分布的比例类比概念。我们的方法建立在如下思想上：只要一个分布能够通过由一组合适的观测所诱导的贝叶斯更新变换为另一个分布，这两个分布就是相关的。我们针对指数族的若干标准成员研究了该框架，并讨论了它如何通过高斯混合近似自然地扩展到任意概率分布。
cs.AI / 17 / 2608.11727
Harness-IF: Evaluating Instruction Following Across Instruction Surfaces in Coding Agents
Harness-IF：评估编码智能体中跨指令表面的指令遵循
Abstract
When a coding agent obeys a rule, it may simply have been going to do that anyway. Existing instruction-following benchmarks cannot tell the difference: they concentrate rules in the user turn, while coding-agent benchmarks emphasize final task success. We introduce Harness-IF, which scores operational rules one at a time from execution evidence: 60 realistic multi-turn coding items drawn from a 642-rule library, 256 rules receiving verdicts, placed on the five configurable surfaces a deployed agent reads. To separate compliance from coincidence we introduce Against-Prior Accuracy (AP-Acc), which scores only rules labeled as opposing unprompted defaults, observed by re-running tasks with the rule withheld across nine probe builds and curated otherwise. Across 12 frontier models, accuracy spans 72.1-85.9% and AP-Acc 66.1-78.6%; every model is worse on against-prior rules, by 3.6 to 7.4 points (mean 5.81), and the direction survives a common-support analysis with item-clustered intervals. Aggregate scores therefore overstate compliance by a model-specific margin: prior control leaves the top build unchanged and exchanges three adjacent rank pairs. A counterbalanced conflict pilot on nine separate builds adds a second result: pooled precedence does not follow prompt depth, with system prompts, project files, and user instructions ahead of tool and skill descriptions.
Chinese Translation
当一个编码智能体遵守某条规则时，它可能本来就打算那样做。现有指令遵循基准无法区分这一差别：它们将规则集中在用户回合中，而编码智能体基准强调最终任务成功。我们引入了 Harness-IF，它根据执行证据逐条对操作性规则进行评分：60 个现实的多轮编码条目取自一个包含 642 条规则的规则库，其中 256 条规则获得判定，并被放置在已部署智能体所读取的五个可配置表面上。为了将遵从与巧合区分开，我们引入了反先验准确率（AP-Acc），该指标仅对标记为与未提示默认行为相反的规则进行评分，并通过在九个探测构建中移除相应规则后重新运行任务、其余部分经过整理来观测这些规则。在 12 个前沿模型中，准确率范围为 72.1–85.9%，AP-Acc 为 66.1–78.6%；每个模型在反先验规则上的表现都更差，差幅为 3.6 到 7.4 个百分点（平均 5.81），且这一方向在带有项目聚类区间的共同支持分析中仍然成立。因此，总体得分以模型特定的幅度高估了遵从性：先验控制使最佳构建保持不变，并交换了三个相邻排名对。在九个独立构建上进行的平衡冲突试点增加了第二个结果：汇总优先级并不遵循提示深度，系统提示、项目文件和用户指令优先于工具和技能描述。
cs.AI / 18 / 2608.11768
HyperANFIS: Enhancing Rule Representation and Interpretability in Adaptive Neuro-Fuzzy Systems via Hyperbolic Geometry
HyperANFIS：通过双曲几何增强自适应神经模糊系统中的规则表示与可解释性
Abstract
The adaptive neuro-fuzzy inference system (ANFIS) is an interpretable reasoning framework capable of generating explicit IF-THEN fuzzy rules, making it suitable for tasks requiring transparent reasoning. However, existing ANFIS models generally construct rule antecedents and perform inference in Euclidean space, limiting their representational capacity and predictive performance. To address this issue, we propose Hyperbolic ANFIS (HyperANFIS), a hyperbolic extension of ANFIS. HyperANFIS preserves the fuzzy semantics and core architecture of conventional ANFIS while performing rule-prototype learning, rule activation, and consequent aggregation in hyperbolic space. It also retains the ability to generate interpretable IF-THEN rules. By exploiting the representational properties of hyperbolic geometry, HyperANFIS strengthens the fuzzy inference process, thereby improving predictive accuracy, inter-rule collaboration, and the credibility of its interpretable rules. Experimental results show that HyperANFIS consistently outperforms the standard ANFIS baseline and various ANFIS variants across all datasets, while also generating higher-quality fuzzy rules.
Chinese Translation
自适应神经模糊推理系统（ANFIS）是一种可解释的推理框架，能够生成显式的 IF-THEN 模糊规则，因此适用于需要透明推理的任务。然而，现有 ANFIS 模型通常在欧几里得空间中构建规则前件并执行推理，限制了其表示能力和预测性能。为解决这一问题，我们提出了双曲 ANFIS（HyperANFIS），它是 ANFIS 的一种双曲扩展。HyperANFIS 保留了传统 ANFIS 的模糊语义和核心架构，同时在双曲空间中执行规则原型学习、规则激活和结论聚合。它还保留了生成可解释 IF-THEN 规则的能力。通过利用双曲几何的表示特性，HyperANFIS 增强了模糊推理过程，从而提高了预测准确性、规则间协作性以及其可解释规则的可信度。实验结果表明，HyperANFIS 在所有数据集上均一致优于标准 ANFIS 基线和各种 ANFIS 变体，同时生成了更高质量的模糊规则。
cs.AI / 19 / 2608.11775
The Sleeping Agent: What Gist-Based Context Compression Loses and Why
沉睡的智能体：基于要点的上下文压缩丢失了什么及其原因
Abstract
Gist-based context compression---summarising older conversation history into compact representations---is a common approach in long-horizon language model agents, yet its effect on different types of memory retrieval is poorly understood. We use Salience-Weighted Consolidation (SWC), a biologically-inspired compression framework motivated by sleep-based memory consolidation, as a diagnostic probe to study when gist compression helps and when it hurts. SWC scores conversation history by salience, partitions it into priority tiers, and applies structured gist abstraction to mid-priority content. Evaluating four conditions on all ten LoCoMo conversations---1,935 matched text-only questions in total, 1,501 used in the primary aggregate after excluding Category 5 (adversarial) questions---at temperature 0, we find a consistent task-type interaction: gist compression substantially outperforms truncation on multi-hop reasoning and single-hop factual questions, but temporal questions remain substantially harder under compression, with compressed conditions scoring well below the full-context reference on the conversations where both are evaluated. We trace this failure to a specific mechanism: the gist abstraction prompt preserves relational and event structure while discarding dates and times. A preservation analysis across all ten conversations confirms the mechanism: an approximately 20-fold increase in temporal expression preservation (3.05% to 62.39%) with a one-sentence prompt modification, while named entity and event preservation rates barely change (x1.02 and x1.11), demonstrating that the fix is a precision instrument. The prompt modification recovers +0.314 [0.254, 0.375] judge accuracy on category-2 (temporal) questions in the matched set. Code and results: https://github.com/kyrkewood/sleeping-agent.
Chinese Translation
基于要点的上下文压缩——将较早的对话历史总结为紧凑表示——是长时程语言模型智能体中的一种常见方法，然而它对不同类型记忆检索的影响尚不清楚。我们使用显著性加权巩固（Salience-Weighted Consolidation, SWC）——一种受睡眠记忆巩固启发的生物启发式压缩框架——作为诊断探针，以研究基于要点的压缩何时有帮助、何时有害。SWC 根据显著性对对话历史进行评分，将其划分为优先级层级，并对中等优先级内容应用结构化要点抽象。在温度 0 下，对全部十个 LoCoMo 对话评估四种条件——共 1,935 个匹配的纯文本问题，在排除第 5 类（对抗性）问题后有 1,501 个用于主要汇总——我们发现一种一致的任务类型交互：基于要点的压缩在多跳推理和单跳事实性问题上显著优于截断，但时间性问题在压缩下仍然显著更难，压缩条件在同时评估两者的对话上得分远低于完整上下文参考。我们将这一失败追溯到一个具体机制：要点抽象提示在保留关系和事件结构的同时丢弃了日期和时间。对全部十个对话的保留分析证实了这一机制：通过一句提示修改，时间表达的保留率提高了约 20 倍（从 3.05% 到 62.39%），而命名实体和事件的保留率几乎不变（x1.02 和 x1.11），表明该修正是精准工具。该提示修改在匹配集中的第 2 类（时间性）问题上恢复了 +0.314 [0.254, 0.375] 的评判准确率。代码与结果：https://github.com/kyrkewood/sleeping-agent。
cs.AI / 20 / 2608.11888
Agent Skills Can Be Harmful: An Empirical Study of Skill-Induced Failures in LLM Agents
智能体技能可能有害：一项关于 LLM 智能体中技能引发失败的实证研究
Abstract
Agent skills are the de facto mechanism for extending LLM agents with reusable guidance. A skill can shape the agent's task execution, including planning, tool use, problem-solving, and validation. Prior work reported mixed results of agent skills: some skills improve task success rates, while others have no effect, increase token use and execution time, and even reduce success rates. This paper presents a comprehensive analysis of skill-induced agent failures by attributing task failures and cost regressions to specific loaded skills. We introduce a differential analysis framework that attributes a failure or regression to a skill by comparing a target skill-guided run against a no-skill or semantically matched skill reference run that solves the same task, or solves it more cheaply. We instantiate this framework on SkillsBench and SWE-Skills-Bench, yielding 307 skill-induced failures, including 125 functional failures and 182 efficiency regressions. We also build SkillTriage, a taxonomy-guided attribution tool that normalizes paired cases, extracts differential evidence, and produces triage reports. Our major findings include: (1) Skill induced functional failures are rarely caused by obviously irrelevant skills; instead, seemingly relevant skills often make the agent incorrectly implement or omit task-required implementation elements. (2) Skill-induced efficiency regressions are not explained by prompt length alone. (3) The largest sources within Excessive Procedure are excessive verification and heavy implementation pipelines, contributing 67 and 30 cases, respectively. This shows that skills often turn validation checklists and construction recipes into mandatory work. Based on our findings, we propose research topics and tooling improvements for safer and more cost-aware skill reuse.
Chinese Translation
智能体技能是通过可复用指导扩展 LLM 智能体的事实标准机制。技能可以塑造智能体的任务执行过程，包括规划、工具使用、问题求解和验证。先前工作报告了智能体技能好坏参半的结果：一些技能提高了任务成功率，而另一些技能没有效果、增加了 token 使用量和执行时间，甚至降低了成功率。本文通过将任务失败和成本回退归因于具体加载的技能，对技能引发的智能体失败进行了全面分析。我们引入了一个差分分析框架，该框架通过将目标技能引导的运行与无技能或语义匹配技能的参考运行进行比较，将失败或回退归因于某个技能；该参考运行要么解决相同任务，要么以更低成本解决该任务。我们在 SkillsBench 和 SWE-Skills-Bench 上实例化了该框架，共得到 307 个技能引发的失败，包括 125 个功能性失败和 182 个效率回退。我们还构建了 SkillTriage，一个由分类法指导的归因工具，它可对配对案例进行规范化、提取差异证据并生成分诊报告。我们的主要发现包括：（1）技能引发的功能性失败很少由明显不相关的技能引起；相反，看似相关的技能常常使智能体错误地实现或遗漏任务所需的实现要素。（2）技能引发的效率回退不能仅用提示长度来解释。（3）在“过度流程（Excessive Procedure）”类别中，最大的来源是过度验证和繁重的实现流水线，分别贡献了 67 个和 30 个案例。这表明，技能常常把验证清单和构建配方变成强制性工作。基于这些发现，我们为更安全、更具成本意识的技能复用提出了研究课题和工具改进。
cs.AI / 21 / 2608.11905
Policy-as-logic for robust reasoning over rules
策略即逻辑：规则上的鲁棒推理
Abstract
In many practical applications of generative AI systems, from tax rules to airline baggage allowance, responses to natural language queries must respect written policies or rules. We present a hybrid symbolic approach that expresses policies in formal logic and at inference time exploits the representation power of language models for fact extraction to ground predicates, and an answer set solver for reasoning such that responses are interpretable, auditable, and as we show, accurate and robust under input perturbations. Specifically, we show this separation of extraction and reasoning steps outperforms policy-as-prompt and policy-as-code methods in most cases with ~10x reduction in token usage. The results point to the value of structured reasoning and symbolic solvers in conjunction with generative models to make robust decisions involving objective criteria.
Chinese Translation
在生成式人工智能系统的许多实际应用中，从税务规则到航空公司行李限额，对自然语言查询的响应必须遵守书面政策或规则。我们提出一种混合符号方法，该方法将政策表示为形式逻辑，并在推理时利用语言模型的表示能力进行事实提取以实例化谓词，同时利用答案集求解器进行推理，从而使响应具有可解释性、可审计性，并且如我们所示，在输入扰动下具有准确性和鲁棒性。具体而言，我们表明，这种提取与推理步骤的分离在大多数情况下优于策略即提示和策略即代码方法，并且token使用量减少约10倍。结果表明，结构化推理和符号求解器与生成模型相结合，对于做出涉及客观标准的稳健决策具有重要价值。
cs.AI / 22 / 2608.11941
OEIS Open: How many conjectures can language models turn into theorems?
OEIS Open：语言模型能把多少猜想转化为定理？
Abstract
We construct OEIS Open, a benchmark based on 492 open mathematical conjectures from the OEIS, formalized in Lean by Tsoukalas et al. Whereas these conjectures had previously been attempted only with a bespoke agent, our open-source evaluation code runs any generic language model (LM) against them, and is secure against LM cheating attempts. We find that LMs equipped with a minimal set of tools resolve 147 of these conjectures with a budget of \$50 per attempt, scoring 30% on OEIS Open. OEIS Open Lite is a random subset of 100 conjectures for cheaper evaluation. When evaluated with a budget of \$200 per attempt, the best current LM scores 44% on OEIS Open Lite. Giving LMs access to the mathematics literature via 476,000 papers from arXiv did not increase performance on OEIS Open Lite, and nor did using more sophisticated agent loops. The conjectures covered in this work are of uncertain mathematical significance, and most have likely received little previous attention. Nevertheless, our results show that LMs can resolve open research conjectures autonomously and at modest cost.
Chinese Translation
我们构建了 OEIS Open，这是一个基于来自 OEIS 的 492 个开放数学猜想的基准，由 Tsoukalas 等人用 Lean 形式化。尽管这些猜想此前仅通过一个定制智能体尝试过，我们的开源评估代码可对它们运行任何通用语言模型（LM），并能防范语言模型的作弊尝试。我们发现，配备最少工具集的语言模型在每次尝试 50 美元的预算下解决了其中 147 个猜想，在 OEIS Open 上得分为 30%。OEIS Open Lite 是由 100 个猜想组成的随机子集，用于成本更低的评估。在每次尝试 200 美元的预算下评估时，当前最好的语言模型在 OEIS Open Lite 上得分为 44%。让语言模型通过来自 arXiv 的 476,000 篇论文访问数学文献，并没有提高其在 OEIS Open Lite 上的表现；使用更复杂的智能体循环也没有提高表现。本工作所涵盖的猜想在数学上的重要性尚不确定，且其中大多数此前很可能很少受到关注。尽管如此，我们的结果表明，语言模型能够以适度的成本自主解决开放研究猜想。
cs.AI / 23 / 2608.11949
ExRole: From Team Trajectories to Executable Roles in Multi-Agent Language Models
ExRole：从团队轨迹到多智能体语言模型中的可执行角色
Abstract
Roles provide an interpretable interface for organizing language-model agents, yet most multi-agent systems treat them as hand-written prompt labels disconnected from learned behavior and parameter updates. We argue that a useful role should instead be an executable control variable: it should summarize behavior predictive of future utility, guide subsequent interaction, and identify the trainable capacity responsible for that behavior. We introduce ExRole, a trajectory-to-role framework that learns future-aware role prototypes from prefix-local team traces, resolves them into readable instructions and token-aligned role markers, and optionally routes shared LoRA rank slots with turn-aligned credit. Across MuSiQue and 2WikiMultiHopQA, ExRole improves over single-agent search by 15.0/14.4 and 13.5/16.1 EM/F1 points, respectively. Against the strongest non-ExRole controls, the corresponding gains remain 11.5/11.6 and 7.7/9.7 points. Across both benchmarks, the controlled results consistently favor trajectory-induced role conditioning over role-free, manual, random, and shuffled alternatives. Role-Agent-Turn interventions further show that the induced roles capture transferable behavioral specialization beyond fixed agent identities or turn positions.
Chinese Translation
角色为组织语言模型智能体提供了一种可解释的接口，然而大多数多智能体系统将其视为手写提示标签，与学习到的行为和参数更新相脱节。我们认为，一个有用的角色应当是一种可执行的控制变量：它应总结能够预测未来效用的行为，指导后续交互，并识别负责该行为的可训练容量。我们提出了 ExRole，一个从轨迹到角色的框架，该框架从前缀局部团队轨迹中学习具有未来感知的角色原型，将其解析为可读指令和与 token 对齐的角色标记，并可选择性地以轮次对齐的信用分配方式路由共享 LoRA 秩槽位。在 MuSiQue 和 2WikiMultiHopQA 上，ExRole 相较于单智能体搜索分别将 EM/F1 提高了 15.0/14.4 和 13.5/16.1 个分数点。与最强的非 ExRole 对照方法相比，相应增益仍分别为 11.5/11.6 和 7.7/9.7 个分数点。在这两个基准上，受控结果一致表明，轨迹诱导的角色条件化优于无角色、手工设定、随机和打乱顺序的替代方案。Role-Agent-Turn 干预进一步表明，所诱导的角色捕获了可迁移的行为专门化，超越了固定的智能体身份或轮次位置。
cs.AI / 24 / 2608.11977
Retry, Switch, or Abstain? Learning Strategy-Aware Tool-Use Policies via Controlled Error Injection
重试、切换还是放弃？通过受控错误注入学习策略感知的工具使用策略
Abstract
Tool-using LLM agents are commonly trained and evaluated in environments where tool calls succeed reliably, yet deployed tools can fail transiently, persistently, or silently. Robust recovery therefore requires more than repeated retries: an agent may need to retry the same path, switch to an alternative, or recognize that no viable path remains. We present BENCH2ROBUST, a framework that converts failure-free tool-use benchmarks into controlled stochastic environments with scenario-controlled solvability, where episodes explicitly require retrying, switching, or stopping after available paths are exhausted. We use BENCH2ROBUST to study two complementary interventions: structured runtime recovery context through Bayesian Tool Memory (BTM), and curriculum-controlled reinforcement learning. Across 7 models from 4 families and two multi-turn benchmark families, tool failures produce a near-universal robustness gap. On held-out Retail tasks, BTM improves robustness by up to 16.8 percentage points without retraining, while RL learns complementary recovery behavior that remains beneficial without inference-time BTM. Combining the two reaches 40.8-45.5% under injection while preserving failure-free performance. These results suggest that robust tool use benefits from combining environment-specific recovery knowledge with learned recovery behavior.
Chinese Translation
使用工具的 LLM 智能体通常在工具调用能够可靠成功的环境中进行训练和评估，然而部署后的工具可能发生暂时性、持续性或静默性故障。因此，鲁棒恢复需要的不仅仅是反复重试：智能体可能需要重试同一路径、切换到替代方案，或者认识到已无可行路径。我们提出了 BENCH2ROBUST，一个将无故障的工具使用基准转换为受控随机环境的框架，该环境具有场景可控的可解性，其中每个回合明确要求在可用路径耗尽后进行重试、切换或停止。我们利用 BENCH2ROBUST 研究两种互补的干预措施：通过贝叶斯工具记忆（BTM）提供的结构化运行时恢复上下文，以及课程控制的强化学习。在来自 4 个系列的 7 个模型以及两个多轮基准系列上，工具故障导致了近乎普遍的鲁棒性差距。在留出的零售任务上，BTM 在无需重新训练的情况下将鲁棒性最多提高了 16.8 个百分点，而 RL 学习到的互补恢复行为在没有推理时 BTM 的情况下仍然有益。将两者结合在注入条件下达到 40.8-45.5%，同时保持无故障性能。这些结果表明，鲁棒的工具使用受益于将环境特定的恢复知识与学习到的恢复行为相结合。
cs.AI / 25 / 2608.11994
Claim-Level Reliability Assessment for Efficient Test-Time Reasoning
面向高效测试时推理的声明级可靠性评估
Abstract
We propose claim-level falsification as a principle for test-time scaling and instantiate it through Claim-Level Reliability Assessment (CLR), a training-free framework that reallocates test-time compute from additional solution sampling to targeted verification. Since whole-trace evaluation often obscures decisive errors due to signal dilution from routine tokens, CLR condenses each reasoning trace into a compact set of decision-critical claims, thereby isolating its logical anchors. Furthermore, recognizing the inherent difficulty of generating entirely correct solutions under fixed model capabilities, CLR shifts the focus to semantic falsification. This approach exploits a fundamental asymmetry between solution construction and claim refutation. Constructing a valid solution requires a flawless reasoning path, whereas refuting an incorrect claim requires identifying only a single decisive flaw. This targeted search for negative evidence systematically compresses the survival space of high-confidence incorrect traces, effectively suppressing erroneous consensus via nonlinear reliability scoring. Across four LLMs and four reasoning benchmarks under matched budgets, CLR generally improves upon pass@1 and self-consistency. On GPT-OSS-20B/CMIMC25, for instance, CLR exceeds pass@1 by 27.15 percentage-points and raises self-consistency accuracy from 77.50\% to 82.19\% with 37.0\% fewer tokens.
Chinese Translation
我们提出声明级证伪作为测试时扩展的一项原则，并通过声明级可靠性评估（CLR）将其具体化；CLR 是一个免训练框架，将测试时计算从额外的解决方案采样重新分配到针对性验证。由于整条轨迹评估常常因常规词元造成的信号稀释而掩盖决定性错误，CLR 将每条推理轨迹压缩为一组紧凑的决策关键声明，从而隔离其逻辑锚点。此外，认识到在固定模型能力下生成完全正确的解决方案具有固有困难，CLR 将重点转向语义证伪。该方法利用了解决方案构造与声明反驳之间的基本不对称性。构造一个有效解决方案需要一条无瑕疵的推理路径，而反驳一个错误声明只需识别出单个决定性缺陷。这种对负面证据的定向搜索系统性地压缩高置信度错误轨迹的生存空间，并通过非线性可靠性评分有效抑制错误共识。在匹配预算条件下，跨四个大语言模型和四个推理基准，CLR 总体上优于 pass@1 和自洽性。例如，在 GPT-OSS-20B/CMIMC25 上，CLR 比 pass@1 高出 27.15 个百分点，并将自洽性准确率从 77.50\% 提高到 82.19\%，同时 token 使用量减少 37.0\%。
cs.AI / 26 / 2608.12002
CTBench: Evaluating Troubleshooting Capabilities of AI Agents in Realistic Telecom Network Operations
CTBench：评估 AI 智能体在真实电信网络运维中的故障排除能力
Abstract
Agents are increasingly considered for automating network operations and maintenance, where engineers must diagnose network faults, optimize configurations to enhance services, and reduce operational costs while acting under strict constraints. However, existing evaluations fail to accurately model real network characteristics or assess agents under partially observable telecom environments with diverse vendors, devices, protocols, and interfaces. In this paper, we introduce CTBench, a public benchmark for assessing whether an agent behaves like a competent telecom troubleshooting engineer. CTBench focuses on root cause analysis and path restoration. Each task is constructed by experts and annotated with rich task metadata, including golden evidence steps. CTBench uses expert-grounded metrics that evaluate both final answers and the diagnostic evidence. Experiments with representative harness-model combinations show that state-of-the-art agents perform very well at identifying endpoints in path-restoration tasks but, more generally, underperform in root cause analysis. In particular, agents struggle with interface state, link-layer, service-management, and other operational faults. Most importantly, even when agents produce plausible or correct final answers, they often fail to provide the evidence-grounded diagnoses required in operational practice. Our results further show that path restoration is generally more resource expensive, yet larger resource usage does not necessarily translate into better diagnosis.
Chinese Translation
智能体正越来越多地被考虑用于自动化网络运营与维护；在此过程中，工程师必须诊断网络故障、优化配置以提升服务，并在严格约束下降低运营成本。然而，现有评估未能准确建模真实网络特征，也未能在具有多厂商、多设备、多协议和多接口的部分可观测电信环境中评估智能体。在本文中，我们引入了 CTBench，这是一个用于评估智能体是否表现得像一名称职的电信故障排除工程师的公开基准。CTBench 重点关注根因分析和路径恢复。每个任务均由专家构建，并标注了丰富的任务元数据，包括金标准证据步骤。CTBench 采用基于专家经验的指标，同时评估最终答案和诊断证据。对代表性执行框架-模型组合的实验表明，最先进的智能体在路径恢复任务中识别端点方面表现非常出色，但在根因分析方面总体上表现不佳。特别是，智能体在接口状态、链路层、服务管理及其他运维故障方面存在困难。最重要的是，即使智能体给出了看似合理或正确的最终答案，它们往往仍未能提供运维实践中所需的基于证据的诊断。我们的结果进一步表明，路径恢复通常资源开销更大，但更大的资源使用量并不一定能转化为更好的诊断。
cs.AI / 27 / 2608.12036
Mechanist: AI as a Scientific Instrument for Discovering the Mechanisms of Intelligence
Mechanist：AI 作为发现智能机制的科学仪器
Abstract
AI models have achieved remarkable success across diverse domains, yet the mechanisms underlying their capabilities and the risks they may pose remain poorly understood. As AI development becomes faster and increasingly automated, mechanistic exploration remains largely manual, widening the gap between what models can do and our ability to understand and control them. To bridge this gap, we introduce Mechanist, an agentic system that uses AI as a scientific instrument for the autonomous discovery of mechanisms underlying AI intelligence. To support autonomous mechanistic discovery, we construct an interpretability-focused knowledge graph of approximately 13,000 papers and integrate it with a multidisciplinary database of 43 million papers spanning 26 fields. We further curate a library of 32 foundational methods for mechanism analysis, causal intervention, and validation. Compared with Claude Code and existing AI-scientist systems, Mechanist generates more valuable mechanism hypotheses and executes experiments more reliably. Mechanist also demonstrates a progression from discovering model behaviors to explaining and controlling AI models. Specifically, Mechanist first uncovers a counterintuitive safety risk in scientific laboratories, showing that unsafe traits can transfer across modalities through apparently safe training data. Mechanist then develops a mechanism theory of belief, revealing how models represent world knowledge, form beliefs, infer the beliefs of others, and how these mechanisms emerge during pretraining. Finally, Mechanist translates these mechanistic insights into practical interventions that improve model performance across diverse scenarios and steer scientific foundation models toward generating DNA sequences with specified properties.
Chinese Translation
AI 模型已在众多领域取得显著成功，然而其能力背后的机制以及其可能带来的风险仍未被充分理解。随着 AI 开发变得更快且日益自动化，机制探索在很大程度上仍然是手动的，这拉大了模型所能做到的事情与我们理解和控制它们的能力之间的差距。为了弥合这一差距，我们提出了 Mechanist，这是一个智能体系统，它将 AI 用作科学仪器，以自主发现 AI 智能背后的机制。为支持自主机制发现，我们构建了一个以可解释性为重点的、包含约 13,000 篇论文的知识图谱，并将其与涵盖 26 个领域、包含 4300 万篇论文的多学科数据库集成。我们还整理了一个包含 32 种基础方法的库，用于机制分析、因果干预和验证。与 Claude Code 以及现有的 AI 科学家系统相比，Mechanist 能生成更有价值的机制假设，并更可靠地执行实验。Mechanist 还展示了从发现模型行为到解释和控制 AI 模型的进展。具体而言，Mechanist 首先揭示了科学实验室中一个反直觉的安全风险，表明不安全特征可以通过看似安全的训练数据跨模态迁移。随后，Mechanist 提出了关于信念的机制理论，揭示了模型如何表示世界知识、形成信念、推断他人信念，以及这些机制如何在预训练期间出现。最后，Mechanist 将这些机制性见解转化为实际干预措施，在多种场景下提升模型性能，并引导科学基础模型生成具有指定特性的 DNA 序列。
cs.AI / 28 / 2608.12097
Graph-Structured Rubrics: Compiling Rubrics into Typed Evaluation Graphs for LLM Judges
图结构化评分标准：将评分标准编译为面向 LLM 评判器的类型化评估图
Abstract
Rubric-based evaluators commonly treat rubrics as prompt context or flat criteria: they specify what to judge but leave criterion composition implicit, even when natural-language rules state it. We introduce Graph-Structured Rubrics (GSR), which compiles a rubric into a response-independent typed evaluation graph before observing responses. Criterion nodes elicit judgments; transformation, reduction, and gating operators compose them through named ports; and a task-specific output mapping, termed Readout, converts the unique sink into a score or preference. Compilation rejects malformed or type-incompatible graphs. Pointwise evaluation judges rubric dimensions separately before graph aggregation; pairwise evaluation reuses the graph with one judgment for each candidate under every criterion. Under GPT-OSS-120B, GSR improves exact score agreement by 0.62--6.75 percentage points over Prometheus-style scoring on four pointwise datasets and achieves the numerically highest end-to-end pairwise accuracy on two preference benchmarks under native tie and abstention policies.
Chinese Translation
基于评分标准的评估器通常将评分标准视为提示上下文或扁平化准则：它们规定了要评判什么，但即使自然语言规则已明确表述了准则组合方式，也仍将准则组合留作隐式处理。我们提出图结构化评分标准（GSR），它在观察响应之前将评分标准编译为与响应无关的类型化评估图。准则节点引出判断；变换、归约和门控算子通过命名端口组合这些节点；一个任务特定的输出映射（称为 Readout）将唯一的汇节点转换为分数或偏好。编译过程会拒绝格式错误或类型不兼容的图。逐点评估在图聚合之前分别评判各评分标准维度；成对评估复用该图，在每个准则下对每个候选对象进行一次判断。在 GPT-OSS-120B 下，GSR 在四个逐点数据集上相较于 Prometheus 风格评分将精确分数一致性提高了 0.62--6.75 个百分点，并在原生并列与弃权策略下，在两个偏好基准上取得了数值最高的端到端成对准确率。
cs.AI / 29 / 2608.12133
GUIDE: Governed Unified Intelligence for Document-to-Artifact Generation in Enterprise Settings
GUIDE：企业环境中用于文档到工件生成的受治理统一智能
Abstract
Enterprise guideline documents are heterogeneous and multimodal, combining narrative text, complex tables, and embedded images. Existing LLM and VLM systems face hallucinated content, table structure degradation, and lack governed workflows extending beyond extraction to validation and artifact generation. This leaves enterprises to perform this manually, consuming 2-3 days per document. To address this, we introduce GUIDE, a governed multi-agent framework built on a shared versioned rule store with schema-validated inter-agent contracts and end-to-end provenance tracking. Six specialized agents handle parsing, VLM-driven extraction, consistency checking, evaluation, human-in-the-loop (HITL) escalation, and persona-tailored artifact synthesis. Evaluated on 120 real-world enterprise guideline documents, GUIDE achieves 96% document success, extracts 3,896 rules with 71.4% auto-approved, produces 812 deployment-ready artifacts, and reduces turnaround to 40-125 minutes per document.
Chinese Translation
企业指南文档是异构且多模态的，结合了叙述性文本、复杂表格和嵌入图像。现有的LLM和VLM系统面临内容幻觉、表格结构退化，以及缺乏从提取延伸到验证和工件生成的受治理工作流等问题。这使得企业只能手动执行这些工作，每份文档耗时2至3天。为解决这一问题，我们提出了GUIDE，一个受治理的多智能体框架，构建在共享的版本化规则存储之上，具有经过模式验证的智能体间契约和端到端的溯源跟踪。六个专用智能体分别处理解析、VLM驱动的提取、一致性检查、评估、人在回路（HITL）升级以及按角色定制的工件合成。在120份真实企业指南文档上的评估中，GUIDE实现了96%的文档成功率，提取了3,896条规则，其中71.4%自动批准，生成了812个可部署工件，并将每份文档的处理周期缩短至40至125分钟。
cs.AI / 30 / 2608.12192
How to Spend Your Oracle Budget: Practical Guidance for Protein Structure Prediction Models
如何花费你的预言机预算：蛋白质结构预测模型的实用指南
Abstract
Foundation models for protein structure prediction remain unreliable on certain targets. External oracles can flag and correct these failures, but biological oracles are expensive, making oracle budget a critical constraint. Existing guidance methods, such as FK-steering, DPO, and Best K-of-N sampling, differ in how they spend this budget, yet no systematic comparison exists to guide method selection. To bridge this gap, we benchmark these methods alongside the recently proposed Optimisation Over Outputs (O3), which applies off-the-shelf optimisers within a generative model's latent subspace. We extend the usage of O3 to protein structure prediction models. Overall, our work provides the first practical reference for oracle budget-aware guidance. Our evaluation on two protein targets, calmodulin (1CLL) and E. coli aspartate transcarbamoylase (9EEH), reveals that no single method consistently dominates across all budgets and oracles. Specifically, O3 proves most effective at low oracle budgets, while FK-steering and DPO demonstrate improved performance as the budget increases. We distil these findings into actionable recommendations for practitioners operating under real-world oracle-budget constraints.
Chinese Translation
蛋白质结构预测的基础模型在某些靶标上仍然不可靠。外部预言机可以标记并纠正这些失败，但生物预言机成本高昂，使得预言机预算成为一个关键约束。现有的引导方法，例如 FK-steering、DPO 和 Best K-of-N 采样，在使用这一预算的方式上有所不同，但尚无系统性比较来指导方法选择。为弥合这一差距，我们将这些方法与最近提出的输出优化（O3）一起进行基准测试；O3 在生成模型的潜在子空间内应用现成的优化器。我们将 O3 的用途扩展到蛋白质结构预测模型。总体而言，我们的工作为感知预言机预算的引导提供了首个实用参考。我们对两个蛋白质靶标——钙调蛋白（1CLL）和大肠杆菌天冬氨酸转氨甲酰酶（9EEH）——的评估表明，没有任何单一方法能在所有预算和预言机条件下始终占优。具体而言，O3 在低预言机预算下被证明最有效，而 FK-steering 和 DPO 随着预算增加表现出更好的性能。我们将这些发现提炼为可操作的建议，供在现实世界预言机预算约束下工作的实践者参考。
cs.AI / 31 / 2608.12249
An Agentic Workflow for Legacy HPC Modernization: Converting the Two-Electron-Integral Core of GAMESS
一种用于遗留 HPC 现代化的智能体工作流：转换 GAMESS 的双电子积分核心
Abstract
Modernizing legacy Fortran is a problem of volume: the transformations are individually routine, but the codebases can be enormous, and across much of computational science the work simply goes undone. We propose an agentic workflow that takes this work on at production scale, and we set out to measure how far such delegation can reach. In this work, three prompt-specialized agent roles operate under a version-controlled specification that the agents themselves authored and revised, while humans hold a small number of gates. The arrangement is kept safe by an exact verification oracle inherited from the domain, and the boundary of safe delegation lies exactly where that oracle stops seeing. We apply the proposed workflow in a case study, converting the two-electron-integral routines of GAMESS (General Atomic and Molecular Electronic Structure System), a mature quantum-chemistry package with a 48-year development history, from fixed-form Fortran 77 to free-form Fortran 2008. The scope of this work was twelve source files, 56,448 lines, and 225 subroutines for computing electron repulsion integrals. The agents ran as three Claude Code roles in isolated worktrees, and the work spanned four Claude model generations. Because the GAMESS group ships a standard test suite whose printed energies its user community treats as canonical, we could adopt bit-for-bit reproduction of those energies as the merge criterion, where a deviation in the twelfth decimal place counts as a failure rather than drift. All twelve source files pass a 51-test validation battery comprising the 49 standard GAMESS tests and two additional calculations, and across 612 test runs the number of chemistry-relevant differences is zero, and every file also passes the Jenkins tests that are used for continuous integration.
Chinese Translation
对遗留 Fortran 进行现代化是一个规模问题：各项转换单独来看都是常规操作，但代码库可能极其庞大，而且在计算科学的许多领域中，这项工作干脆被搁置未做。我们提出了一种以生产规模承担这项工作的智能体工作流，并着手衡量这种委托能够延伸到多远。在这项工作中，三个经过提示词专门化的智能体角色在一个由这些智能体自行编写和修订的版本控制规范下运作，而人类只掌握少量关口。这种安排通过一个从该领域继承而来的精确验证预言机来保证安全，而安全委托的边界恰好位于该预言机不再能够看到的地方。我们在一个案例研究中应用了所提出的工作流，将具有 48 年开发历史的成熟量子化学软件包 GAMESS（通用原子与分子电子结构系统）的双电子积分例程从固定格式 Fortran 77 转换为自由格式 Fortran 2008。这项工作的范围是 12 个源文件、56,448 行代码以及 225 个用于计算电子排斥积分的子例程。这些智能体在隔离的工作树中以三个 Claude Code 角色的形式运行，该工作跨越了四代 Claude 模型。由于 GAMESS 团队发布了一个标准测试套件，且其用户社区将套件输出的能量视为权威标准，我们可以采用对这些能量进行逐位复现作为合并标准，其中第十二位小数的偏差被视为失败而非漂移。全部 12 个源文件通过了一个由 49 个标准 GAMESS 测试和两个附加计算组成的 51 项验证测试组，在 612 次测试运行中，与化学相关的差异数量为零，并且每个文件还通过了用于持续集成的 Jenkins 测试。
cs.AI / 32 / 2608.12282
VAKRA: Evaluating Multi-Hop Reasoning Across APIs and Retrieval Under Tool-Use Policies
VAKRA：在工具使用策略下评估跨 API 与检索的多跳推理
Abstract
Agents deployed in enterprise settings must reason across structured APIs and document collections, yet existing benchmarks evaluate these capabilities in isolation. We introduce VAKRA (e\textbf{V}aluating \textbf{A}PI and \textbf{K}nowledge \textbf{R}etrieval \textbf{A}gents), a benchmark of over $8{,}000$ executable APIs across $62$ domains with tasks spanning three settings of increasing difficulty: diverse API interaction styles, multi-hop reasoning over structured APIs, and multi-source reasoning with natural-language tool-use policy constraints. Correctness is verified by re-executing predicted tool calls against live APIs, accommodating multiple valid paths. Using a fixed ReAct harness to isolate model capabilities from agent architecture, we evaluate frontier and open-weight models and find that even the best model achieves only 70.4\% on single-hop endpoint-style tasks and drops to 50--51\% on compositional APIs; performance degrades by over 50\% as reasoning depth increases, and policy-constrained questions expose severe failures (as low as 2.4\% on unanswerable queries). Trace analysis shows failures concentrate at language-mediated reasoning - entity disambiguation, cross-source grounding, rather than tool invocation mechanics. Code is available https://github.com/IBM/VAKRA. Dataset is available https://huggingface.co/datasets/ibm-research/VAKRA
Chinese Translation
部署在企业环境中的智能体必须跨结构化 API 和文档集合进行推理，然而现有基准将这些能力孤立地评估。我们提出 VAKRA（评估 API 与知识检索智能体），这是一个覆盖 $62$ 个领域、包含超过 $8{,}000$ 个可执行 API 的基准，其任务横跨三种难度递增的设置：多样的 API 交互风格、基于结构化 API 的多跳推理，以及带有自然语言工具使用策略约束的多源推理。正确性通过在实时 API 上重新执行预测的工具调用来验证，从而容纳多条有效路径。我们使用固定的 ReAct 框架将模型能力与智能体架构隔离开来，评估前沿模型和开放权重模型，并发现即使最好的模型在单跳端点式任务上也仅达到 70.4%，在组合式 API 上降至 50--51%；随着推理深度增加，性能下降超过 50%，而策略约束问题暴露出严重失败（在不可回答查询上低至 2.4%）。轨迹分析显示，失败集中在以语言为中介的推理——实体消歧、跨源 grounding，而不是工具调用机制。代码可在 https://github.com/IBM/VAKRA 获取。数据集可在 https://huggingface.co/datasets/ibm-research/VAKRA 获取。
cs.AI / 33 / 2608.11317
Clinical Feasibility of Low-Magnification Fluorescence Imaging for Breast Cancer Margin Detection Using Texture Analysis and Deep Learning
基于纹理分析和深度学习的低放大倍率荧光成像用于乳腺癌切缘检测的临床可行性
Abstract
High-resolution images of unprocessed surgical breast tissue can be obtained using microscopy with ultraviolet surface excitation (MUSE). This technique is considered a promising method for checking surgical margins during breast cancer surgery. In this study, MUSE images at 4x and 10x magnifications were compared using patch-level classification methods. Texture analysis (TA) based on local binary patterns (LBP) and deep learning (DL) with a base Vision Transformer (ViT) model were used. Both methods achieved similar performance at both magnifications. Using DL method, both 4x and 10x magnifications achieved 96.30% sensitivity, 100% specificity and 98.18% accuracy. Using TA method, 4x achieved better specificity (100% vs 93.33%) and 10x yielded higher sensitivity (100% vs 93.33%), but both had the same accuracy (96.67%). No clear improvement in performance was observed with 10x magnification. These results show that 4x imaging achieves the same diagnostic accuracy as 10x imaging. At the same time, 4x offers a larger field of view and faster image capture. Therefore, lower magnification can be effectively used in MUSE systems for accurate and efficient intraoperative margin assessment.
Chinese Translation
使用紫外表面激发（MUSE）显微镜可以获得未经处理的手术乳腺组织的高分辨率图像。该技术被认为是乳腺癌手术中检查手术切缘的一种有前景的方法。在本研究中，使用图像块级分类方法对4倍和10倍放大倍率下的MUSE图像进行了比较。使用了基于局部二值模式（LBP）的纹理分析（TA）和采用基础Vision Transformer（ViT）模型的深度学习（DL）。两种方法在两种放大倍率下均取得了相似的性能。使用DL方法时，4倍和10倍放大倍率均实现了96.30%的灵敏度、100%的特异性和98.18%的准确率。使用TA方法时，4倍放大倍率实现了更好的特异性（100% 对比 93.33%），10倍放大倍率产生了更高的灵敏度（100% 对比 93.33%），但两者的准确率相同（96.67%）。10倍放大倍率下未观察到性能的明显改善。这些结果表明，4倍成像能够达到与10倍成像相同的诊断准确性。同时，4倍成像提供更大的视场和更快的图像采集速度。因此，在MUSE系统中可以有效地使用较低放大倍率，以实现准确且高效的术中切缘评估。
cs.AI / 34 / 2608.11335
Dual-Domain Cross-Modal Decoding for Clinical Text-Guided Medical Image Segmentation
面向临床文本引导医学图像分割的双域跨模态解码
Abstract
Clinical text can narrow down what to segment, but recent text-guided designs emphasize spatial alignment while overlooking frequency content that governs texture and boundaries. We propose Dual-Domain Cross-Modal Decoding (DD-CMD) for clinical text-guided pulmonary infection segmentation, integrating two complementary forms of language guidance during decoding. In the spatial domain, Text-Guided Spatial Cross-Attention (TGSA) aligns multi-scale visual tokens with text semantics and updates features through gated residual fusion. In the frequency domain, Spectral-Text Adaptive Modulation (STAM) applies a 2D DCT to compute learnable band-energy statistics and predicts text-conditioned FiLM parameters to recalibrate decoder channels for frequency-aware decoding. DD-CMD embeds TGSA and STAM into a coarse-to-fine decoder (7x7 to 56x56) and restores full-resolution masks using a lightweight two-stage refinement module. Experiments on QaTa-COV19 and MosMedData+ show that DD-CMD achieves 91.46% Dice / 84.26% mIoU and 81.95% Dice / 69.42% mIoU, respectively, with average gains of +1.96 Dice and +2.67 mIoU over the strongest prior baselines. Code: https://github.com/maklachur/DD-CMD.
Chinese Translation
临床文本可以缩小分割目标范围，但近期文本引导的设计强调空间对齐，而忽略了支配纹理和边界的频率内容。我们提出双域跨模态解码（DD-CMD），用于临床文本引导的肺部感染分割，在解码过程中整合两种互补的语言引导形式。在空间域中，文本引导空间交叉注意力（TGSA）将多尺度视觉 token 与文本语义对齐，并通过门控残差融合更新特征。在频率域中，频谱-文本自适应调制（STAM）应用二维 DCT 计算可学习的频带能量统计，并预测以文本为条件的 FiLM 参数，以重新校准解码器通道，实现频率感知解码。DD-CMD 将 TGSA 和 STAM 嵌入到由粗到细的解码器（7x7 到 56x56）中，并使用轻量级两阶段细化模块恢复全分辨率掩膜。在 QaTa-COV19 和 MosMedData+ 上的实验表明，DD-CMD 分别达到 91.46% Dice / 84.26% mIoU 和 81.95% Dice / 69.42% mIoU，相比最强先前基线平均提升 +1.96 Dice 和 +2.67 mIoU。代码：https://github.com/maklachur/DD-CMD。
cs.AI / 35 / 2608.11367
Gaze Target Estimation Anywhere with Concepts
基于概念的任意场景注视目标估计
Abstract
Estimating human gaze targets from images in-the-wild is an important and formidable task. Existing approaches primarily employ brittle, multi-stage pipelines that require explicit inputs, like head bounding boxes and human pose, in order to identify the subject of gaze analysis. As a result, detection errors can cascade and lead to failure. Moreover, these prior works lack the flexibility of specifying the gaze analysis task via natural language prompting, an approach which has been shown to have significant benefits in convenience and scalability for other image analysis tasks. To overcome these limitations, we introduce the Promptable Gaze Target Estimation (PGE) task, a new end-to-end, concept-driven paradigm for gaze analysis. PGE conditions gaze prediction on flexible user text or visual prompts (e.g., "the boy in the red shirt" or "person in point [0.52, 0.48]") to identify a specific subject for gaze analysis. This approach integrates subject localization with gaze estimation, and eliminates the rigid dependency on intermediate analysis stages. We develop a scalable data engine to generate Gaze-Co (Gaze Estimation with Concepts), a dataset and benchmark of 120K high-quality, prompt-annotated image pairs. We also propose GazeAnywhere, the first model designed for PGE. GazeAnywhere uses a transformer-based detector to fuse features from frozen encoders and simultaneously solves subject localization, in/out-of-frame presence, and gaze target heatmap estimation. GazeAnywhere achieves state-of-the-art performance on multiple PGE benchmarks, setting a strong baseline for this new problem even on a difficult out-of-domain, real-world clinical dataset. GazeAnywhere is open-sourced in github.com/IrohXu/GazeAnywhere.
Chinese Translation
在自然图像中估计人类注视目标是一项重要且艰巨的任务。现有方法主要采用脆弱的多阶段流程，需要显式输入，如头部边界框和人体姿态，以识别注视分析的对象。因此，检测错误可能级联并导致失败。此外，这些先前工作缺乏通过自然语言提示来指定注视分析任务的灵活性，而这种方式已被证明在其他图像分析任务中具有显著的便利性和可扩展性优势。为克服这些局限性，我们提出了可提示的注视目标估计（PGE）任务，这是一种新的端到端、概念驱动的注视分析范式。PGE 根据灵活的用户文本或视觉提示（例如“穿红衬衫的男孩”或“位于点 [0.52, 0.48] 处的人”）来条件化注视预测，以识别用于注视分析的特定主体。该方法将主体定位与注视估计相结合，并消除了对中间分析阶段的刚性依赖。我们开发了一个可扩展的数据引擎，以生成 Gaze-Co（基于概念的注视估计），这是一个包含 12 万对高质量、带提示标注图像的数据集和基准。我们还提出了 GazeAnywhere，这是首个为 PGE 设计的模型。GazeAnywhere 使用基于 Transformer 的检测器来融合来自冻结编码器的特征，并同时解决主体定位、画面内/外存在性判断以及注视目标热图估计问题。GazeAnywhere 在多个 PGE 基准上取得了最先进的性能，即使在困难的域外真实临床数据集上，也为这一新问题建立了强大的基线。GazeAnywhere 已在 github.com/IrohXu/GazeAnywhere 开源。
cs.AI / 36 / 2608.11452
TangPoetryBench: A Multi-Dimensional Benchmark and Rubric-Conditioned Evaluator for Poetry-to-Image Generation
TangPoetryBench：面向诗歌到图像生成的多维基准与以评分标准为条件的评估器
Abstract
Text-to-image (T2I) models are increasingly asked to illustrate literary and cultural content, yet we cannot measure how well an image renders the meaning of a poem. The task is many-sided: a good illustration must be visually sound, faithful to the poem's imagery and scene, culturally and stylistically apt, free of spurious text, and true to its emotion, and its deepest requirements, imagery and especially implicit emotion, are never stated in the words. Existing metrics (CLIPScore, BLIPScore, VQAScore) reward literal text-image correspondence and so cannot tell whether an illustration succeeds, let alone why, or even separate the best model from the worst. We introduce TangPoetryBench, a multi-dimensional benchmark of 1,280 images (320 classical Chinese Tang poems x 4 state-of-the-art T2I models) with quality-controlled human annotations across ten dimensions. Analyzing this data, we reveal the shared and model-specific strengths and weaknesses of current T2I models, including their ability to evoke a poem's implicit emotion. We further introduce PoemAutoEvaluator (PAE), an open, rubric-conditioned evaluator that reaches parity with a strong proprietary judge (Claude), generalizes to an unseen generator and a second poetic tradition (Song Ci), and lets the benchmark scale to new images without fresh human annotation. We release the benchmark, annotations, and evaluator.
Chinese Translation
文本到图像（T2I）模型越来越多地被要求为文学和文化内容绘制插图，但我们仍无法衡量一幅图像在多大程度上呈现了一首诗的含义。这项任务具有多面性：一幅好的插图必须在视觉上合理，忠实于诗歌的意象和场景，在文化和风格上恰当，不包含杂散文字，并忠实于其情感；而这项任务最深层次的要求，即意象、尤其是隐含情感，从未在文字中明说。现有指标（CLIPScore、BLIPScore、VQAScore）奖励的是字面上的文本-图像对应，因此无法判断一幅插图是否成功，更不用说判断其成功原因，甚至无法将最好的模型与最差的模型区分开来。我们提出了 TangPoetryBench，这是一个包含 1,280 张图像（320 首中国唐代古典诗歌 × 4 个最先进的 T2I 模型）的多维基准，并带有在十个维度上经过质量控制的人工标注。通过分析这些数据，我们揭示了当前 T2I 模型共有的和模型特有的优势与不足，包括它们唤起诗歌隐含情感的能力。我们进一步提出了 PoemAutoEvaluator（PAE），这是一个开放的、以评分标准为条件的评估器，其表现可与强大的专有评判模型（Claude）相媲美，能够泛化到未见过的生成器和第二种诗歌传统（宋词），并使该基准能够扩展到新图像而无需新的人工标注。我们发布了该基准、标注和评估器。
cs.AI / 37 / 2608.11582
A Hybrid Framework of Vision Transformer and Gated Recurrent Unit for Detection of Mosquito Diseases
一种用于检测蚊虫疾病的视觉Transformer与门控循环单元混合框架
Abstract
Identifying dengue virus-infected mosquitoes from control mosquitoes is a major challenge in analyzing mosquito locomotion behavior due to the small size and complexity of the video background. Conventional AI methods are often unable to extract accurate features from video frames and produce erroneous features. In this study, a three-step framework is introduced: first, mosquitoes are identified and the background is removed using the YOLO 11M model, then visual features are extracted using the Vision Transformer (ViT), and finally the videos are classified with a convolutional GRU (ConvGRU) classifier. A comparative analysis of different models, including Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and their convolutional versions showed that the ConvGRU model achieved the best performance; it achieved 88.88% accuracy, 84.45% precision, 82.82% recall, and 82.81% F1 score. These results demonstrate that combining convolutional models with sequence-based networks, especially in the ConvGRU model, allows the simultaneous extraction of precise spatial features and long-term temporal dependencies from mosquito movements. Finally, the proposed framework provides a reliable solution for analyzing mosquito behavior in complex environments.
Chinese Translation
从对照蚊子中识别登革热病毒感染蚊子是分析蚊子运动行为中的一项重大挑战，因为视频背景中的目标尺寸小且背景复杂。传统人工智能方法通常无法从视频帧中提取准确特征，并会产生错误特征。本研究引入了一个三步框架：首先，使用YOLO 11M模型识别蚊子并去除背景，然后使用Vision Transformer（ViT）提取视觉特征，最后使用卷积GRU（ConvGRU）分类器对视频进行分类。对不同模型（包括循环神经网络（RNN）、长短期记忆网络（LSTM）、门控循环单元（GRU）及其卷积版本）的比较分析表明，ConvGRU模型取得了最佳性能；其准确率为88.88%，精确率为84.45%，召回率为82.82%，F1分数为82.81%。这些结果表明，将卷积模型与基于序列的网络相结合，尤其是在ConvGRU模型中，能够同时从蚊子运动中提取精确的空间特征和长期时间依赖关系。最后，所提出的框架为在复杂环境中分析蚊子行为提供了一种可靠的解决方案。
cs.AI / 38 / 2608.11681
Learning from Multimodal Pseudo-Labels for Robust Open-Vocabulary Instance and Panoptic Segmentation
从多模态伪标签中学习以实现鲁棒的开放词汇实例分割与全景分割
Abstract
This work addresses the challenge of open-vocabulary instance segmentation (OVIS) and open-set panoptic segmentation (OSPS), which aim to recognize both predefined and unseen object categories without exhaustive human annotations. Existing methods often suffer from noisy pseudo-masks, limited visual-textual grounding, and difficulty handling synonyms or out-of-vocabulary (OOV) words. To overcome these challenges, we propose a multimodal framework that leverages pre-trained vision-language models for automatic pseudo-label generation, CLIP-guided synonym filtering, and GPT-based caption reconstruction. In our target-vocabulary-assisted pseudo-labeling setting, the framework first constructs pseudo segmentation masks, descriptive captions, and semantically aligned synonym sets using Grounded SAM, LLaVA, and CLIP, providing multimodal supervision without manual annotation. We then enhance visual-textual alignment through three complementary training objectives: an extended grounding loss that incorporates visually grounded synonyms, a semantic consistency loss, and a generative caption reconstruction loss. Extensive experiments on the COCO dataset demonstrate that the proposed method consistently outperforms previous state-of-the-art approaches under this protocol, achieving substantial improvements on both OVIS and OSPS benchmarks.
Chinese Translation
本工作致力于解决开放词汇实例分割（OVIS）和开放集全景分割（OSPS）的挑战，这两项任务旨在在无需详尽人工标注的情况下识别预定义和未见过的物体类别。现有方法通常受到噪声伪掩码、有限的视觉-文本对齐以及难以处理同义词或词汇外（OOV）单词等问题的困扰。为克服这些挑战，我们提出了一种多模态框架，该框架利用预训练的视觉-语言模型进行自动伪标签生成、CLIP引导的同义词过滤以及基于GPT的描述文本重建。在我们提出的目标词汇辅助伪标签设定下，该框架首先使用Grounded SAM、LLaVA和CLIP构建伪分割掩码、描述性文本和语义对齐的同义词集合，从而在无需人工标注的情况下提供多模态监督。随后，我们通过三个互补的训练目标来增强视觉-文本对齐：一个融合了视觉对齐同义词的扩展对齐损失、一个语义一致性损失，以及一个生成式描述文本重建损失。在COCO数据集上进行的大量实验表明，在该协议下，所提方法一致优于以往最先进的方法，在OVIS和OSPS基准上均取得了显著改进。
cs.AI / 39 / 2608.11847
LookBack: Where and How to Score LVLM Responses via Visual Reference Usage
LookBack：在何处以及如何通过视觉参考使用对 LVLM 响应进行评分
Abstract
Large Vision-Language Models (LVLMs) integrate visual perception with language generation, enabling responses that span image understanding and complex reasoning. However, LVLMs do not just inherit the text-level hallucinations; they also hallucinate against the image, producing fluent responses ungrounded in what they see. This makes LVLM response scoring inherently harder, and our diagnostics show that existing confidence-based metrics adopted from LLMs are insufficient for LVLMs. Specifically, removing the input image barely changes confidence-based selection, suggesting that output-space confidence primarily captures textual plausibility rather than agreement with the image. To address this gap, we propose LookBack, a training-free LVLM response scoring method that augments token likelihood with visual lookback score, a lightweight measure of how strongly each response token refers to image tokens. Across four benchmarks and three models, LookBack consistently improves Best-of-$N$ selection over existing baselines with negligible additional overhead.
Chinese Translation
大型视觉语言模型（LVLMs）将视觉感知与语言生成相结合，使其能够产生涵盖图像理解和复杂推理的响应。然而，LVLMs 不仅继承了文本层面的幻觉；它们还会针对图像产生幻觉，生成流畅但与所见内容脱节的响应。这使得 LVLM 响应评分本质上更加困难，并且我们的诊断表明，从 LLM 借鉴的现有基于置信度的指标对 LVLM 而言并不充分。具体而言，移除输入图像几乎不会改变基于置信度的选择，这表明输出空间中的置信度主要捕捉的是文本合理性，而非与图像的一致性。为了解决这一差距，我们提出了 LookBack，一种免训练的 LVLM 响应评分方法，该方法用视觉回看分数增强标记似然；视觉回看分数是一种轻量级度量，用于衡量每个响应标记对图像标记的引用强度。在四个基准和三个模型上，LookBack 始终以可忽略的额外开销改善了 Best-of-$N$ 选择，优于现有基线。
cs.AI / 40 / 2608.11907
Do You See What You Draw? A Semantic Closed-Loop Framework for Holistic Evaluation of Unified Multimodal Models
你看到你所画的了吗？一种用于统一多模态模型整体评估的语义闭环框架
Abstract
As Large Vision-Language Models increasingly aim to integrate visual generation and understanding within a single parameter space, evaluating such structural unification in a cohesive manner remains a critical challenge. Current evaluation protocols predominantly treat generative and discriminative capabilities as separate tasks, leaving a gap in system-level evaluation for unified multimodal models (UMMs). In this work, we propose Self-Generative-Understanding (SGU), a novel, annotation-free evaluation framework that probes the integrated capabilities of unified models through a semantic closed-loop challenge. Without requiring new annotations, SGU leverages the dual understanding-and-generation abilities of UMMs by asking them to first perceive an image and produce a textual description, subsequently reconstruct a visual context based on that description, and finally perform reasoning over the self-generated output. This pipeline provides a zero-cost testbed that yields an integrated performance score specifically tailored for evaluating UMMs as unified systems. Extensive experiments show that even high-performing UMMs often struggle to reason over their own generated contexts, revealing limitations that are not captured by separate evaluations of understanding or generation alone. Our work provides a complementary holistic evaluation framework and offers a foundation for benchmarking the development of next-generation unified multimodal models.
Chinese Translation
随着大型视觉-语言模型越来越致力于在单一参数空间内整合视觉生成与理解，以统一的方式评估这种结构一体化仍然是一个关键挑战。当前的评估协议主要将生成能力和判别能力视为独立任务，从而在统一多模态模型（UMMs）的系统级评估方面留下了空白。在本文中，我们提出自生成-理解（SGU），一种新颖的、无需标注的评估框架，该框架通过语义闭环挑战来探测统一模型的综合能力。在无需新增标注的情况下，SGU 利用统一多模态模型的双重理解与生成能力，要求它们首先感知一幅图像并生成文本描述，随后基于该描述重建视觉上下文，最后对自生成输出进行推理。该流水线提供了一个零成本测试平台，能够产生一个专门为将 UMMs 作为统一系统进行评估而定制的综合性能得分。大量实验表明，即使是性能优异的 UMMs 也常常难以对自身生成的上下文进行推理，这揭示了仅通过单独评估理解或生成能力所无法捕捉的局限性。我们的工作提供了一个互补的整体评估框架，并为基准测试下一代统一多模态模型的发展奠定了基础。
cs.AI / 41 / 2608.12035
How Far from Clinical Deployment? Evaluating the Complete Unsupervised Domain Adaptation Pipeline in Medical Imaging
距离临床部署还有多远？评估医学影像中完整的无监督领域自适应流程
Abstract
Deploying unsupervised domain adaptation (UDA) in clinical practice requires choosing which algorithm to use and which of its trained models to ship. However, the deployment (target) domain is unlabeled, so models cannot be evaluated directly on it, leaving it unclear which to select. We address this by evaluating the complete UDA pipeline, considering both adaptation and label-free selection together. Our study covers eleven clinically relevant cross-domain scenarios from nine medical imaging datasets, with ten UDA algorithms and 13 label-free selection methods (validators), evaluating over 80,000 trained models in total. By this, we find that a capable adapted model usually exists, but identifying it without target labels is difficult: the validator-selected models leave a large and structural target performance gap to the best available one, with no evaluated validator consistently reliable. Towards closing it, we explore two strategies, ensembling and a small target-labeling budget; both narrow this gap but do not close it entirely. Overall, deployable UDA depends on the complete pipeline; addressing the less explored selection step could bring much of current UDA closer to clinical use.
Chinese Translation
在临床实践中部署无监督领域自适应（UDA）需要选择使用哪种算法以及将其哪个训练好的模型交付使用。然而，部署（目标）域是无标签的，因此无法直接在其上评估模型，导致不清楚该选择哪一个。我们通过评估完整的UDA流程来解决这一问题，同时考虑自适应和无标签选择。我们的研究涵盖来自九个医学影像数据集的十一个临床相关跨域场景，使用十种UDA算法和13种无标签选择方法（验证器），总共评估了超过80,000个训练模型。由此，我们发现通常存在性能良好的自适应模型，但在没有目标标签的情况下识别它很困难：验证器选择的模型与最佳可用模型之间存在较大且结构性的目标性能差距，且没有哪个被评估的验证器始终可靠。为了缩小这一差距，我们探索了两种策略：集成和少量目标标签预算；两者都能缩小这一差距，但无法完全消除。总体而言，可部署的UDA取决于完整流程；解决较少被探索的选择步骤可能使当前许多UDA更接近临床应用。
cs.AI / 42 / 2608.12078
Better Slots, Better Worlds: Representation Quality & Robustness in Object-Centric World Models
更好的槽，更好的世界：以对象为中心的世界模型中的表示质量与鲁棒性
Abstract
Learning world models from offline trajectories enables agents to accomplish different tasks through planning. Object-centric (OC) representations, which decompose a scene into a set of slots that bind to its objects, have been proposed as an inductive bias for world models that are more sample-efficient and generalize better. Yet prior object-centric world models (OCWMs) take the slot encoder as given and evaluate only in-distribution, leaving open whether the object-centric bias actually delivers for planning and what within the OCWM drives it. We conduct a controlled study of OCWMs for visual model-predictive control along two axes: object-centric representation quality and generalization under distribution shift relative to scene-centric models. We find that (i) planning success correlates positively with unsupervised slot-quality metrics (FG-ARI, mBO), though the gains saturate at high slot quality; (ii) with well-bound slots, the auxiliary proprioception inputs and masking inductive bias that prior methods relied on become unnecessary; and (iii) under unseen distribution shifts, the OCWM with well-bound slots is more robust overall than the end-to-end trained scene-centric LeWM, while DINO-WM, built on similar frozen pretrained features, remains competitive -- pointing to pretrained features as a key contributor to robustness.
Chinese Translation
从离线轨迹中学习世界模型使智能体能够通过规划完成不同任务。以对象为中心（OC）的表示将场景分解为一组绑定到其对象的槽，已被提出作为世界模型的一种归纳偏置，使其更具样本效率并具有更好的泛化能力。然而，先前以对象为中心的世界模型（OCWM）将槽编码器视为给定，并且仅在分布内进行评估，因此尚不清楚以对象为中心的偏置是否真的能为规划带来好处，以及 OCWM 中的什么因素驱动了这一效果。我们对用于视觉模型预测控制的 OCWM 进行了受控研究，沿两个轴展开：以对象为中心的表示质量，以及相对于以场景为中心的模型在分布偏移下的泛化能力。我们发现：(i) 规划成功率与无监督槽质量指标（FG-ARI、mBO）呈正相关，尽管在高槽质量下收益趋于饱和；(ii) 在槽绑定良好的情况下，先前方法所依赖的辅助本体感知输入和掩码归纳偏置变得不再必要；(iii) 在未见过的分布偏移下，具有良好绑定槽的 OCWM 总体上比端到端训练的以场景为中心的 LeWM 更鲁棒，而基于类似冻结预训练特征构建的 DINO-WM 仍具有竞争力——这表明预训练特征是鲁棒性的关键贡献因素。
cs.AI / 43 / 2608.12187
HSTGFormer: Hyper Spatial-Temporal Graph Transformer for 3D Human Pose Estimation
HSTGFormer：用于三维人体姿态估计的超时空图Transformer
Abstract
Transformer-based methods have achieved strong performance in monocular 3D human pose estimation, but most existing approaches organise spatial and temporal reasoning as separate stages, which may weaken unified spatial-temporal interdependencies inherent in human motion and compress frame-level structural information before temporal modelling. In this paper, we propose HSTGFormer, a graph-enhanced Transformer framework that reformulates spatial-temporal reasoning as localised coupled graph aggregation over joint-time nodes. Specifically, HSTGFormer introduces a Hyper Spatial-Temporal Graph (HSTG), which decomposes global spatial-temporal reasoning into local spatial-temporal receptive fields around individual joint-time nodes by extending per-frame skeleton graphs into temporal neighbourhoods, thereby enabling structure-aware coupled reasoning while preserving local structural motion information. It further incorporates an Adaptive Dual-Scale Temporal Graph (ADSTG) to capture joint-specific temporal dependencies over complementary short- and long-range windows. A lightweight node-wise fusion module further adaptively integrates the two graph representations for each joint-time node. Experiments on Human3.6M and MPI-INF-3DHP show that HSTGFormer achieves strong accuracy with high computational efficiency.
Chinese Translation
基于Transformer的方法在单目三维人体姿态估计中已取得了较强的性能，但大多数现有方法将空间推理与时间推理组织为相互分离的阶段，这可能削弱人体运动中固有的统一时空相互依赖关系，并在时间建模之前压缩帧级结构信息。本文提出了HSTGFormer，一个图增强的Transformer框架，将时空推理重新表述为在关节-时间节点上的局部化耦合图聚合。具体而言，HSTGFormer引入了一种超时空图（HSTG），通过将逐帧骨架图扩展到时间邻域，将全局时空推理分解为围绕各个关节-时间节点的局部时空感受野，从而在保留局部结构运动信息的同时实现结构感知的耦合推理。它还进一步结合了一个自适应双尺度时间图（ADSTG），以在互补的短程和长程窗口上捕获关节特定的时间依赖关系。一个轻量级的逐节点融合模块进一步为每个关节-时间节点自适应地整合这两种图表示。在Human3.6M和MPI-INF-3DHP上的实验表明，HSTGFormer在具有高计算效率的同时取得了较强的准确率。
cs.AI / 44 / 2608.12196
M-Net: Integrating Spectral Features and Physical Field Operators into Deep Learning for Medical Image Segmentation
M-Net：将谱特征与物理场算子融入深度学习用于医学图像分割
Abstract
Purpose: Deep learning-based medical image segmentation has achieved remarkable success, yet purely data-driven approaches often fail to exploit the rich mathematical structure inherent in medical images. We investigate whether explicit mathematical inductive biases, specifically matrix spectral analysis and vector calculus operators, can enhance segmentation beyond data-driven learning alone. Methods: We propose M-Net (Math-Augmented Network), which integrates three complementary mathematical priors into U-Net: (1) continuous spectral features derived from the condition number of centered local pixel matrices, providing a differentiable measure of texture ill-conditioning; (2) physical field operators (divergence and a discrete curl-like boundary irregularity operator) computed from image gradient fields, capturing focal intensity extrema and edge non-smoothness; and (3) a Math-Attention Gate (MAG) that adaptively fuses mathematical features with CNN-extracted deep features at skip connections. Results: Experiments on three benchmarks (LiTS, KiTS, and BraTS) show that M-Net achieves Dice scores of 78.42%, 76.15%, and 83.67%, outperforming baseline U-Net by 12.37%, 3.52%, and 5.55% on liver, kidney, and brain tumor segmentation, respectively. Ablations reveal that the condition-number feature contributes a 2.14% gain over binary invertibility features, while MAG adds 1.45% over simple concatenation. Conclusion: M-Net establishes that mathematical inductive biases provide effective complementary information for medical image segmentation. The continuous condition-number feature offers superior gradient information over discrete alternatives, and MAG preserves these priors throughout the network. This work opens avenues for integrating linear algebra and vector calculus into deep architectures for medical imaging.
Chinese Translation
目的：基于深度学习的医学图像分割已取得显著成功，然而纯数据驱动的方法往往未能利用医学图像中固有的丰富数学结构。我们研究显式的数学归纳偏置，特别是矩阵谱分析和向量微积分算子，是否能够在单纯的数据驱动学习之外提升分割性能。方法：我们提出 M-Net（数学增强网络），将三个互补的数学先验集成到 U-Net 中：（1）从中心化局部像素矩阵的条件数导出的连续谱特征，提供纹理病态性的可微度量；（2）从图像梯度场计算的物理场算子（散度和离散类旋度边界不规则性算子），捕获局部强度极值和边缘非平滑性；（3）数学注意力门（MAG），在跳跃连接处自适应地融合数学特征与 CNN 提取的深度特征。结果：在三个基准（LiTS、KiTS 和 BraTS）上的实验表明，M-Net 的 Dice 分数分别达到 78.42%、76.15% 和 83.67%，在肝脏、肾脏和脑肿瘤分割上分别比基线 U-Net 高出 12.37%、3.52% 和 5.55%。消融实验显示，条件数特征相比二值可逆性特征带来 2.14% 的提升，而 MAG 相比简单拼接带来 1.45% 的提升。结论：M-Net 证明了数学归纳偏置为医学图像分割提供了有效的补充信息。连续条件数特征相较于离散替代方案提供了更优的梯度信息，并且 MAG 在整个网络中保留这些先验。这项工作为将线性代数和向量微积分融入医学影像深度架构开辟了途径。
cs.AI / 45 / 2608.12220
SCOUT: Unlocking Enhanced Spatial Reasoning via Structured Chain-of-Thought and Multi-Objective Process Reward
SCOUT：通过结构化思维链与多目标过程奖励释放增强的空间推理能力
Abstract
Existing Vision-Language Models (VLMs) exhibits a critical bottleneck in robust spatial reasoning. Recent reinforcement learning (RL) methods aim to close this gap with verifiable outcomes, yet they suffer from poor credit assignment across intermediate reasoning steps. Concurrently, structured reasoning approaches overlook the critical depth perception necessary for comprehensive 3D understanding. To address these challenges, we propose SCOUT (Structured Chain-Of-Thought Utilizing Process-Supervised RL Training). Specifically, we design a structured Chain-of-Thought (CoT) framework that explicitly models 3D environmental perception to ensure robust spatial understanding and reasoning. Furthermore, we introduce a novel RL algorithm featuring multi-objective process rewards and a tailored advantage estimation method, facilitating fine-grained credit assignment across distinct segments of the reasoning trajectory. To support our framework, we develop SCOUT-24k, a structured spatial reasoning CoT dataset synthesized through a customized pipeline. Extensive evaluations demonstrate that SCOUT-3B improves upon baseline models by 16.85% and 6.3% on general spatial benchmarks and complex spatial reasoning tasks respectively. Notably, our larger SCOUT-7B even outperforms GPT-4o by a margin of 4.28%. Moreover, despite being trained exclusively on single image, SCOUT-7B exhibits robust out-of-domain generalization to multi-image and video scenarios. These empirical results render SCOUT as a critical step towards next generation of spatially-aware VLMs.
Chinese Translation
现有视觉语言模型（VLM）在鲁棒空间推理方面存在关键瓶颈。最近的强化学习（RL）方法旨在通过可验证结果来缩小这一差距，但它们在各中间推理步骤之间的功劳分配上表现不佳。同时，结构化推理方法忽视了全面三维理解所必需的关键深度感知。为应对这些挑战，我们提出 SCOUT（采用过程监督强化学习训练的结构化思维链）。具体来说，我们设计了一个结构化思维链（CoT）框架，显式建模三维环境感知，以确保稳健的空间理解与推理。此外，我们引入了一种新的强化学习算法，具有多目标过程奖励和专门设计的优势估计方法，有助于在推理轨迹的不同部分之间进行细粒度功劳分配。为支持该框架，我们开发了 SCOUT-24k，一个通过定制流程合成的结构化空间推理 CoT 数据集。大量评估表明，SCOUT-3B 在通用空间基准和复杂空间推理任务上分别比基线模型提高了 16.85% 和 6.3%。值得注意的是，我们更大的 SCOUT-7B 甚至比 GPT-4o 高出 4.28%。此外，尽管仅在单图像上训练，SCOUT-7B 对多图像和视频场景表现出鲁棒的分布外泛化能力。这些实证结果使 SCOUT 成为迈向下一代空间感知视觉语言模型的关键一步。
cs.AI / 46 / 2608.12230
Few-Shot Ordinal Learning for Day-Wise Freshness Estimation with Hyperspectral Fish Images
基于高光谱鱼类图像的逐日新鲜度估计的少样本序数学习
Abstract
Non-destructive food quality assessment has increasingly benefited from hyperspectral imaging (HSI), which captures spectral signatures linked to biochemical changes during storage. Estimating day-wise freshness, however, remains challenging owing to strong inter-fillet variability and scarce labelled data per product. All existing deep learning approaches for HSI-based freshness prediction operate under full supervision, requiring densely annotated training sets that are costly to obtain at the individual-product level. We introduce, to the best of our knowledge, the first few-shot learning framework for HSI-based food quality estimation. Each fillet defines a distinct episodic task, and a CORAL-style ordinal prediction head captures the ranked nature of freshness progression through cumulative threshold modelling. Biologically grounded monotonicity and embedding smoothness constraints further guide predictions toward plausible trajectories. On a 16-day salmon HSI dataset under a strict unseen-fillet protocol, our method achieves a mean absolute error of 1.58 days and 2-day accuracy of 72.3% with only three labelled days per fillet, substantially outperforming scalar regression and label-distribution baselines under an identical unseen-fillet protocol.
Chinese Translation
非破坏性食品质量评估已越来越多地受益于高光谱成像（HSI），高光谱成像能够捕捉与储存期间生化变化相关的光谱特征。然而，由于鱼片之间较大的个体差异以及每种产品标注数据稀缺，逐日新鲜度估计仍具有挑战性。现有的所有基于 HSI 的新鲜度预测深度学习方法均在全监督条件下运行，需要密集标注的训练集，而在单个产品层面获取这些标注成本高昂。据我们所知，我们提出了首个面向基于 HSI 的食品质量估计的少样本学习框架。每个鱼片定义一个不同的情景任务，CORAL 风格的序数预测头通过累积阈值建模捕捉新鲜度进展的排序特性。具有生物学依据的单调性约束和嵌入平滑性约束进一步引导预测朝着合理的变化轨迹发展。在严格的未见鱼片协议下的 16 天三文鱼 HSI 数据集上，我们的方法在每片鱼片仅标注三天的情况下实现了 1.58 天的平均绝对误差和 72.3% 的 2 天准确率，显著优于相同未见鱼片协议下的标量回归和标签分布基线。
cs.AI / 47 / 2608.12239
HAMP-LIC: Hessian-Aware Mixed-Precision Post-Training Quantization for Learned Image Compression
HAMP-LIC：面向学习图像压缩的 Hessian 感知混合精度训练后量化
Abstract
Use this plain-text version for the arXiv abstract field: Learned image compression (LIC) models achieve strong rate-distortion performance but are hindered by high computational complexity and encoding-decoding mismatches across heterogeneous hardware platforms. Uniform fixed-precision quantization alleviates these issues but suffers severe quality degradation at low bit widths because it ignores differences in the quantization sensitivities of individual layers. To enable efficient and accurate low-bit deployment of pretrained LIC models, we propose HAMP-LIC, a Hessian-aware mixed-precision post-training quantization (PTQ) framework with a four-stage optimization strategy. First, block-wise sensitivity is estimated from the Hessian trace to capture second-order importance. Second, a task-aware refinement module adjusts these sensitivities by jointly considering quantization distortion and rate-distortion performance. Third, guided by the refined sensitivity profile, bit widths are allocated under a global model-size constraint to balance efficiency and reconstruction quality. Finally, block-wise reconstruction using a small calibration set further suppresses quantization error. Experiments on representative LIC models, including Minnen2018 and Cheng2020, demonstrate that HAMP-LIC achieves up to 4.85x model compression with as little as 0.59% BD-rate loss. It consistently outperforms existing fixed- and mixed-precision PTQ methods across multiple datasets while completely eliminating cross-platform encoding-decoding errors.
Chinese Translation
学习图像压缩（LIC）模型实现了较强的率失真性能，但受到高计算复杂度以及跨异构硬件平台的编码-解码不匹配问题的制约。统一固定精度量化缓解了这些问题，但在低位宽下会出现严重的质量下降，因为它忽略了各个层之间量化敏感度的差异。为了实现预训练 LIC 模型的高效且精确的低比特部署，我们提出了 HAMP-LIC，一个具有四阶段优化策略的 Hessian 感知混合精度训练后量化（PTQ）框架。首先，通过 Hessian 迹估计逐块敏感度，以捕捉二阶重要性。其次，一个任务感知的精炼模块通过联合考虑量化失真和率失真性能来调整这些敏感度。第三，在精炼后的敏感度分布指导下，在全局模型大小约束下分配位宽，以平衡效率和重建质量。最后，使用小型校准集进行逐块重建，进一步抑制量化误差。在代表性 LIC 模型（包括 Minnen2018 和 Cheng2020）上的实验表明，HAMP-LIC 实现了高达 4.85 倍的模型压缩，而 BD-rate 损失仅为 0.59%。它在多个数据集上始终优于现有的固定精度和混合精度 PTQ 方法，同时完全消除了跨平台编码-解码误差。
cs.AI / 48 / 2608.12274
A Neighborhood Attention Transformer Network for Enhanced 3D Segmentation of the Left Anterior Descending Artery
一种用于增强左前降支动脉三维分割的邻域注意力Transformer网络
Abstract
Background: Accurate segmentation of the Left Anterior Descending (LAD) artery in 3D free-breathing, non-contrast CT is critical for cardiac dose sparing in thoracic radiotherapy. The LAD is extremely small, has poor soft-tissue contrast, and varies substantially across patients; even manual contours show limited inter-observer agreement, underscoring the ambiguity of the vessel boundaries. Purpose: To develop a transformer-based framework that improves LAD delineation in low-contrast, imbalanced CT through local-global context modeling and uncertainty-guided optimization. Methods: We propose NA-UNETR, a 3D transformer-based segmentation model whose Neighborhood Attention (NA) and Dilated NA (DiNA) blocks jointly capture fine structural detail and long-range context. Given the scarcity of annotated LAD data, the model is pretrained on 1,000 CTA volumes of general coronary anatomy and fine-tuned with LoRA-based parameter-efficient adaptation on 20 free-breathing institutional CT scans. A composite Dice-Focal and Hausdorff loss, dynamically balanced via homoscedastic uncertainty, improves overlap and boundary accuracy. Results: NA-UNETR reached 45.64% Dice, 38.16 mm HD95, and 10.01 mm ASD, improving Dice by 3.10 percentage points over nnU-Net and reducing HD95 by 2.96 mm relative to Swin UNETR, with the strongest boundary accuracy among all models and improved centerline stability. On ImageCAS it achieved 79.49% Dice, 8.89 mm HD95, and 1.02 mm ASD. Ablations confirmed that residual blocks, variable kernels, and uncertainty-weighted loss each contributed. Conclusions: NA-UNETR balances local precision and global context for thin, low-contrast LAD structures, offering a computationally efficient framework for substructure-level cardiac segmentation in radiotherapy planning.
Chinese Translation
背景：在三维自由呼吸、非对比增强CT中准确分割左前降支（LAD）动脉对于胸部放射治疗中的心脏剂量保护至关重要。LAD极小，软组织对比度差，并且在患者之间差异显著；即便是人工勾画，观察者间一致性也有限，这突显了血管边界的模糊性。
目的：开发一种基于Transformer的框架，通过局部-全局上下文建模和不确定性引导优化，改善低对比度、类别不平衡CT中的LAD勾画。
方法：我们提出NA-UNETR，一种基于三维Transformer的分割模型，其邻域注意力（NA）模块和扩张邻域注意力（DiNA）模块联合捕获精细结构细节和长程上下文。鉴于有标注的LAD数据稀缺，该模型在1,000个一般冠状动脉解剖的CTA容积上进行预训练，并在20例自由呼吸的机构CT扫描上采用基于LoRA的参数高效适配进行微调。由同方差不确定性动态平衡的Dice-Focal与Hausdorff复合损失函数，提高了重叠度和边界精度。
结果：NA-UNETR达到45.64%的Dice、38.16 mm的HD95和10.01 mm的ASD，Dice较nnU-Net提高3.10个百分点，HD95相较Swin UNETR降低2.96 mm，在所有模型中边界精度最高，且中心线稳定性有所改善。在ImageCAS数据集上，其Dice为79.49%，HD95为8.89 mm，ASD为1.02 mm。消融实验证实，残差模块、可变核以及不确定性加权损失均有所贡献。
结论：NA-UNETR在细小、低对比度的LAD结构上平衡了局部精度与全局上下文，为放射治疗计划中的子结构级心脏分割提供了一种计算高效的框架。
cs.AI / 49 / 2608.12299
Class Activation Mapping in Explainable Computer Vision: A Method-Centered Review of CNN, Transformer, and Foundation-Model-Era Visual Explanations
可解释计算机视觉中的类激活映射：以方法为中心的 CNN、Transformer 与基础模型时代视觉解释综述
Abstract
Class activation mapping (CAM) is one of the most widely used visual explanation families in explainable artificial intelligence. Its purpose is intuitive: it converts internal model evidence into a heatmap that highlights the image regions, convolutional channels, tokens, or patches that support a target class or concept. Since the first CAM formulation in 2016, the field has moved far beyond global-average-pooled CNN classifiers. CAM-style methods now include gradient-based post-hoc explanations, gradient-free score and ablation methods, high-resolution upscaling, weakly supervised localization and segmentation, transformer token attribution, causal and debiasing methods, and foundation-model-era approaches that use CLIP, DINO, SAM, or feature-distribution comparisons. This review synthesizes a strict corpus of 57 method-centered papers published from 2016 onward. The paper develops a taxonomy that separates methods by attribution mechanism, architectural dependence, and evaluation objective. It then reviews gradient-based CAMs, recent and hybrid CAM-style methods, and model-based or architecture-aware methods. Across the corpus, the main trend is clear: the field is shifting from explaining one class score in one low-resolution CNN layer toward comparative, multi-layer, probabilistic, token-aware, and foundation-model-aware explanations. At the same time, evaluation remains fragmented. Faithfulness, localization, robustness, computational cost, and human trust are often measured with different protocols. The review therefore emphasizes not only what each method contributes, but also which gap it leaves open and which later methods attempt to close that gap.
Chinese Translation
类激活映射（CAM）是可解释人工智能中使用最广泛的视觉解释家族之一。其目的十分直观：它将模型内部证据转化为热力图，突出支持目标类别或概念的图像区域、卷积通道、token 或图像块。自2016年首次提出CAM以来，该领域已经远远超出了全局平均池化CNN分类器的范畴。CAM风格的方法如今包括基于梯度的事后解释、无梯度的分数与消融方法、高分辨率上采样、弱监督定位与分割、Transformer token归因、因果与去偏方法，以及使用CLIP、DINO、SAM或特征分布比较的基础模型时代方法。本综述综合了自2016年以来发表的57篇以方法为中心的严格论文语料。本文提出了一种分类法，根据归因机制、架构依赖性和评估目标对方法进行划分。随后，本文回顾了基于梯度的CAM、近期及混合CAM风格方法，以及基于模型或架构感知的方法。纵观整个语料库，主要趋势十分清晰：该领域正从解释单个低分辨率CNN层中的单一类别分数，转向比较性、多层、概率性、token感知和基础模型感知的解释。与此同时，评估仍然碎片化。忠实性、定位能力、鲁棒性、计算成本和人类信任往往通过不同的协议来衡量。因此，本综述不仅强调每种方法的贡献，还强调它留下了哪些空白，以及后续哪些方法试图填补这些空白。
cs.AI / 50 / 2608.11344
Governing Agentic AI in FinTech
治理金融科技中的智能体人工智能
Abstract
Financial institutions are delegating consequential decisions to agentic AI systems that decompose goals, coordinate models and tools, and act with little oversight. Yet agentic AI governance in FinTech is under-investigated. We argue the binding governance constraint is not capability but verifiability. We define the Verifiability Gap as the shortfall between the verification delegated authority demands and the explainability and reproducibility retained after a decision. It is indexed to a verifier, evidentiary standard, and audit lag. We develop a multilevel governance theory for agentic AI and test its mechanisms in three studies over nine model versions, from a three-billion-parameter local model to a commercial frontier system. Study 1 shows that provider releases alter historical financial actions, and that the controls replay needs belong to the provider: the frontier model rejects temperature, top_p and top_k outright and exposes no random seed. Under the tightest controls each endpoint allows, a local model reproduced 320 of 320 executions, hosted models 319 of 320 and 959 of 960. Study 2 shows that orchestration is a latent policy layer. Architecture changes final actions, and no execution record repeated in any configuration at any scale. The frontier model reproduces its own actions more often than the local ones, its record no better, and loses a comparable share of its differentiation. Capability buys a higher starting point, not auditability. Study 3 shows two deterministic credit-model versions each reproduce their current action perfectly, yet the current cannot recover a historical one. We conceptualize reproducibility as a governance profile, not a scalar, yielding evidence-contingent delegation: authority is defensible only while retained evidence substantiates its exercise. Beyond finance, the framework extends to other high-stakes domains requiring auditability.
Chinese Translation
金融机构正在将后果重大的决策委托给智能体人工智能系统，这些系统分解目标、协调模型和工具，并在极少监督下采取行动。然而，金融科技中的智能体人工智能治理尚未得到充分研究。我们认为，治理的关键约束不是能力，而是可验证性。我们将可验证性缺口定义为受委托权限所要求的验证与决策后保留的可解释性和可复现性之间的差距。它由验证者、证据标准和审计滞后加以标定。我们为智能体人工智能发展了一个多层次治理理论，并在三项研究中检验其机制，这些研究涵盖九个模型版本，从具有三十亿参数的本地模型到商业前沿系统。研究1显示，提供商的发布版本会改变历史金融行为，而重放所需的控制项属于提供商：前沿模型直接拒绝 temperature、top_p 和 top_k，并且不暴露任何随机种子。在每个端点允许的最严格控制下，一个本地模型复现了320次执行中的320次，托管模型分别复现了320次中的319次和960次中的959次。研究2显示，编排是一个潜在政策层。架构会改变最终行为，并且在任何配置、任何规模下都没有执行记录重复出现。前沿模型比本地模型更经常地复现自身的行为，但其记录并不更好，并且损失了相当份额的差异化。能力换来的是更高的起点，而不是可审计性。研究3显示，两个确定性信用模型版本各自都能完美复现其当前行为，但当前版本无法恢复历史行为。我们将可复现性概念化为一种治理画像，而非标量，由此产生以证据为条件的委托：只有当留存证据能够证实权限行使时，权限才具有可辩护性。在金融之外，该框架还扩展到其他需要可审计性的高风险领域。
cs.AI / 51 / 2608.11794
Toward Meaningful Transparency for AI Chatbots: Disclosing Persuasive Intent Reduces Persuasion
迈向有意义的AI聊天机器人透明度：披露说服意图会降低说服效果
Abstract
The growing role of AI-generated content and AI-enabled systems in public communication has led regulators to demand clear disclosure of content provenance and AI involvement. But the effects of such disclosures remain uncertain. We test two disclosure approaches in their impact on an AI chatbot's persuasive appeal. In a preregistered experiment, 1,500 UK adults held a short conversation with a persuasive chatbot about one of 60 policy issues. The chatbot was identical for everyone. We randomized the disclosure that people received: nothing (control), a prominent disclosure that they were interacting with an AI (T1), or that disclosure plus the chatbot's persuasive intent and instructions (T2). The chatbot shifted attitudes by 12.6 points on a 100-point scale in the control group. The AI-identity disclosure was practically equivalent to no disclosure, with a 13.1-point shift, whereas the additional intent disclosure cut the persuasive effect roughly in half to 6.3 points. It also made participants view the campaign's methods as less acceptable and support stronger penalties against it. For direct chatbot interactions, transparency about AI identity alone does not meaningfully impact its influence. While current rules emphasize what a system is, our results show why the regulation of persuasive AI must also address what the system is trying to do.
Chinese Translation
AI生成内容和AI赋能系统在公共传播中日益重要的作用，已促使监管机构要求明确披露内容来源和AI参与情况。但此类披露的效果仍不确定。我们检验了两种披露方式对AI聊天机器人说服吸引力的影响。在一项预注册实验中，1500名英国成年人与一个具有说服性的聊天机器人围绕60个政策议题之一进行了简短对话。该聊天机器人对所有人完全相同。我们随机分配参与者接收到的披露信息：无披露（对照组）、一种显著披露他们正在与AI互动（T1），或在该披露基础上加上聊天机器人的说服意图和指令（T2）。在对照组中，聊天机器人使态度在100分制量表上改变了12.6分。AI身份披露实际上与无披露相当，态度变化为13.1分，而额外的意图披露则将说服效果削减约一半，降至6.3分。它还使参与者认为该宣传活动的方法更不可接受，并支持对其采取更严厉的处罚。对于直接的聊天机器人互动，仅披露AI身份并不会对其影响力产生有意义的影响。虽然当前规则强调系统是什么，但我们的结果说明了为什么对说服性AI的监管还必须关注系统试图做什么。
cs.AI / 52 / 2608.11891
Benchmark-Based Comparative Assessment of Publicly Benchmarked Indian Foundation Models: A Capability and Evaluation-Maturity Framework
基于基准的公开基准评测印度基础模型的比较评估：能力与评估成熟度框架
Abstract
Governments increasingly fund indigenous foundation models to strengthen national AI capability, digital sovereignty, and multilingual computing. Assessing the progress of such national ecosystems is complicated by inconsistent benchmark reporting, proprietary evaluation methodologies, and rapidly evolving model releases. This paper presents a structured, benchmark-based comparative assessment of publicly benchmarked Indian foundation models against global frontier and comparable-scale models, across eight capability domains: general-purpose reasoning, coding and software engineering, agentic AI and computer use, cybersecurity, vision and image understanding, video and multimodal understanding, scientific research, and Indic language capability. Using only publicly reported benchmark results, we find that Indian models achieve strong scores on established benchmarks such as MMLU and MATH-500. However, these benchmarks are now widely regarded as saturated, and frontier developers no longer report them. Indian models participate far less frequently in newer, agentic, and domain-specialized evaluations. Benchmark participation is also highly uneven across Indian organizations. Among the models surveyed, Sarvam AI reports the broadest benchmark coverage by a substantial margin. We propose an exploratory four-dimension Benchmark Maturity Index (BMI), scoring each capability domain on standardization, participation, independent verification, and national coverage. We show that the BMI refines, and in some cases revises, the maturity judgments that a purely descriptive review would produce. We argue that many apparent capability gaps in the public record cannot be distinguished, on available evidence, from evaluation-ecosystem gaps. This has direct implications for how national AI programs should design monitoring and funding criteria.
Chinese Translation
各国政府日益资助本土基础模型，以增强国家人工智能能力、数字主权和多语言计算能力。评估此类国家生态系统的进展因基准报告不一致、专有评估方法以及模型版本的快速迭代而变得复杂。本文提出一种结构化的、基于基准的比较评估，对公开基准评测的印度基础模型与全球前沿及相当规模的模型进行比较，涵盖八个能力领域：通用推理、编码与软件工程、智能体式人工智能与计算机使用、网络安全、视觉与图像理解、视频与多模态理解、科学研究以及印度语言能力。仅使用公开报告的基准结果，我们发现印度模型在 MMLU 和 MATH-500 等既定基准上取得了较强的分数。然而，这些基准如今被广泛认为已经饱和，前沿开发者已不再报告它们。印度模型在较新的、智能体式的和领域专门化的评估中参与频率远低。基准参与情况在印度各组织之间也高度不均衡。在调查的模型中，Sarvam AI 报告的基准覆盖范围明显最广。我们提出一个探索性的四维基准成熟度指数（BMI），从标准化、参与度、独立验证和国家覆盖度四个维度对每个能力领域进行评分。我们表明，BMI 细化了，并在某些情况下修正了纯粹描述性综述所得出的成熟度判断。我们认为，公开记录中许多明显的能力差距，在现有证据下无法与评估生态系统的差距区分开来。这对国家人工智能项目应如何设计监测和资助标准具有直接影响。
cs.AI / 53 / 2608.12104
No One to Blame: A Framework of Constitutive AI Unaccountability
无人可责：构成性人工智能不可问责性框架
Abstract
The increasing deployment of autonomous, agentic AI systems challenges traditional accountability mechanisms. Existing research predominantly frames AI accountability gaps as barriers that can be overcome through better standards, transparency, and institutional reform. We argue that this framing is insufficient: certain configurations of actors, systems, and institutions render AI accountability conceptually unachievable regardless of effort. We introduce the concept of constitutive AI unaccountability to capture these configurations. Through a three-stage qualitative study comprising a concept-centric literature analysis, a secondary analysis of 27 expert interviews with AI professionals from technical, legal, and sociotechnical backgrounds, and an illustrative framework application to the open-source agentic AI system OpenClaw, we identify nine categories and 20 themes of constitutive AI unaccountability. These are organized across structural, technological, and normative clusters and reinforce one another through eight directed interdependencies. Our framework is operationalized as a diagnostic instrument of 20 questions, which detected 17 of 20 conditions when applied to OpenClaw, including an inverted anthropomorphism configuration in which the AI agent was the only identifiable actor. We contribute a reframing of AI unaccountability as a constitutive property of sociotechnical systems, an extension of the four barriers to accountability, and a practical instrument for identifying accountability voids in specific AI deployments.
Chinese Translation
自主、智能体式人工智能系统的日益增多的部署对传统问责机制构成挑战。现有研究主要将人工智能问责差距框定为可以通过更好的标准、透明度和制度性改革来克服的障碍。我们认为，这种框架是不充分的：某些行动者、系统和制度的配置使人工智能问责在概念上无论付出多少努力都无法实现。我们引入构成性人工智能不可问责性的概念来刻画这些配置。通过一项包含以概念为中心的文献分析、对来自技术、法律和社会技术背景的人工智能专业人士的27次专家访谈的二次分析，以及对开源智能体式人工智能系统OpenClaw的说明性框架应用的三阶段定性研究，我们识别出构成性人工智能不可问责性的9个类别和20个主题。这些类别和主题被组织在结构性、技术性和规范性聚类中，并通过八种有向相互依赖关系相互强化。我们的框架被操作化为一个包含20个问题的诊断工具，应用于OpenClaw时检测出20个条件中的17个，其中包括一种倒置的拟人化配置，在这种配置中，人工智能代理是唯一可识别的行动者。我们贡献了将人工智能不可问责性重新框定为社会技术系统的一种构成性属性、对问责四大障碍的扩展，以及一种用于识别特定人工智能部署中问责真空的实用工具。
cs.AI / 54 / 2608.12166
Co-constructing sociotechnical AI governance: participatory system mapping using algorithm registers
共同建构社会技术人工智能治理：利用算法登记册进行参与式系统映射
Abstract
Algorithm registers have been championed as a means of providing transparency on the use of algorithms in public services. Yet potential publics differ in their expectations of what should be made transparent and how, as well as in their interest in and ability to parse the information currently published in the registers. Moreover, it remains unclear how these instruments can represent the sociotechnical systems in which these algorithms are embedded, and how system-level transparency can facilitate accountability. In this paper, we ask, what do algorithm registers reveal (and occlude) about the sociotechnical systems governing algorithmic systems, and how can diverse stakeholder perspectives inform a more pluralistic system-theoretic safety analysis? To do this, we probe the municipal algorithm register of a Dutch city through a case study of a decision-support tool for caseworkers' assessment of citizens' welfare benefits eligibility based on legal automation through a business rule engine. Through interviews, surveys, and participatory system mapping workshops (with municipal staff, civil society organisations, and ombudsmen, N=8), we seek to understand to what extent the register allows stakeholders to map the algorithmic system in question. These maps inform a System-Theoretic Process Analysis (STPA) that situates the register within a wider sociotechnical governance structure. Participants' contributions allow us to identify potential safety hazards which would not have been possible to see using the algorithm register alone, including benefits eligibility denial, system performance deterioration, and inability to contest wrongful decisions. By engaging both direct and indirect stakeholders, we reflect on the normative dimensions of algorithm governance efforts and how politics shape the practice of system safety analysis.
Chinese Translation
算法登记册一直被提倡作为在公共服务中提供算法使用透明度的一种手段。然而，潜在公众在对什么应当被透明化以及如何透明化的期望上存在差异，在解析登记册当前所发布信息的兴趣和能力方面也存在差异。此外，这些工具如何能够表征这些算法所嵌入的社会技术系统，以及系统层面的透明度如何促进问责，仍不清楚。在本文中，我们提出如下问题：算法登记册揭示（和遮蔽）了关于治理算法系统的社会技术系统的哪些内容，以及多元利益相关者的视角如何能够为一种更具多元性的系统理论安全分析提供信息？为此，我们通过一个案例研究探究了荷兰某城市的市政算法登记册，该案例是一个用于个案工作者基于通过业务规则引擎实现的法律自动化来评估公民福利资格的决策支持工具。通过访谈、调查和参与式系统映射工作坊（参与者包括市政工作人员、公民社会组织和监察员，N=8），我们试图理解该登记册在多大程度上允许利益相关者绘制所涉算法系统的地图。这些地图为系统理论过程分析（STPA）提供信息，该分析将登记册置于更广泛的社会技术治理结构之中。参与者的贡献使我们能够识别仅凭算法登记册不可能看到的潜在安全危害，包括福利资格被拒绝、系统性能退化以及无法对错误决定提出异议。通过让直接和间接利益相关者共同参与，我们反思了算法治理努力的规范维度，以及政治如何塑造系统安全分析的实践。
cs.AI / 55 / 2608.11840
User-Assisted Collaborative Distributed Inference for Efficient QoS-Aware Autoscaling
面向高效QoS感知自动伸缩的用户辅助协作分布式推理
Abstract
Growing demand for artificial intelligence (AI) inference services requires scalable infrastructure, yet centralized serving costs rise with demand. We propose a collaborative distributed inference system combining dedicated infrastructure with resources contributed by service users. Dedicated resources provide baseline capacity for maintaining quality of service (QoS), while volunteered resources absorb increasing demand without proportional growth in centralized infrastructure. To capture stochastic and dynamic interactions among users, resources, tasks, and policies, we develop a high-dimensional generative Markov model with structured temporal factorization. The model supports simulation and provides a foundation for task scheduling and QoS-aware resource allocation optimization. We evaluate the system across user populations, resource capacities, and centralized and distributed scheduling policies. Simulations show that distributed scheduling becomes increasingly advantageous as the user population grows, improving request completion and P99 latency while substantially reducing dedicated resource consumption. These results demonstrate the feasibility of user-assisted collaborative inference for infrastructure-efficient autoscaling.
Chinese Translation
对人工智能（AI）推理服务日益增长的需求需要可扩展的基础设施，然而集中式服务的成本随需求上升。我们提出一种协作式分布式推理系统，将专用基础设施与服务用户贡献的资源相结合。专用资源提供维持服务质量（QoS）的基线容量，而志愿资源吸收不断增长的需求，而无需集中式基础设施成比例增长。为了刻画用户、资源、任务和策略之间随机且动态的交互，我们开发了一个具有结构化时间分解的高维生成式马尔可夫模型。该模型支持仿真，并为任务调度和QoS感知的资源分配优化提供基础。我们在不同用户规模、资源容量以及集中式和分布式调度策略下评估该系统。仿真结果表明，随着用户规模增长，分布式调度变得越来越有优势，提高请求完成率并改善P99延迟，同时大幅减少专用资源消耗。这些结果证明了用户辅助协作推理在基础设施高效的自动伸缩方面的可行性。
cs.AI / 56 / 2608.12123
Ready Cohorts: Bounding GPU Opportunity and Avoiding Host Round Trips in LLM-Agent Control
就绪批次：在 LLM 智能体控制中界定 GPU 机会并避免主机往返
Abstract
LLM-agent services repeatedly execute small deterministic transitions between model and tool calls: route an outcome, update state, and emit the next effect. We ask when this control path exposes enough concurrent work for GPU execution, and what changes when a GPU-computed route decision remains on device. We formalize the ready-cohort boundary using fixed-partition share F, exact offline share P*, local upper bound U, and online achieved share A. Under zero service time, unlimited capacity, and equal relative launch deadlines, a specialized dynamic program computes P* exactly. In a stationary Poisson replay of one pinned 851-session public trace panel, the primary condition at 100,000 target active sessions, K=256, and a 50 ms launch deadline gives F=30.19%, P*=43.00%, and U=45.85%. Exact packing recovers 81.83% of the opportunity lost at fixed window boundaries. The outcome-derived route key is a conditioning proxy, not proof of executable identity. A separate mechanism study keeps a GPU-computed binary decision on device instead of returning four bytes to the host and redispatching. Across four named GPU placements, the device-resident path is faster in all 36 configurations; within-placement row-median ratios range from 1.19x to 2.39x. Across both admissible mechanisms, all 14,557,440 tested batched invocations match a separately implemented host oracle. A fixed nested device graph that removes no host decision is slower in all 60 configurations across five placements. Together, the studies establish two measurable gates for GPU agent control: deadline-feasible cohort supply and observation placement. A joined finite online runtime is required to measure A, CPU displacement, and service-level benefit.
Chinese Translation
LLM 智能体服务反复执行模型调用与工具调用之间的小型确定性转移：路由结果、更新状态并发出下一效果。我们探究该控制路径何时为 GPU 执行暴露出足够的并发工作，以及当 GPU 计算的路由决策留在设备上时会发生什么变化。我们使用固定分区份额 F、精确离线份额 P*、局部上界 U 和在线实现份额 A 来形式化就绪批次边界。在零服务时间、无限容量和相等相对启动截止期条件下，专门的动态规划精确计算 P*。在对一个固定的 851 会话公开跟踪面板进行平稳泊松重放时，在 100,000 个目标活跃会话、K=256 和 50 毫秒启动截止期的主要条件下，得到 F=30.19%、P*=43.00% 和 U=45.85%。精确装箱恢复了在固定窗口边界处损失机会的 81.83%。由结果导出的路由键是一个条件化代理，而非可执行身份的证明。一项独立的机制研究将 GPU 计算的二值决策保留在设备上，而不是将四个字节返回主机并重新派发。在四种命名的 GPU 放置方案中，设备驻留路径在全部 36 种配置下都更快；放置内行中位数比率范围为 1.19 倍到 2.39 倍。在两种可接受机制中，所有 14,557,440 次受测批量调用均与单独实现的主机 oracle 匹配。一个未移除任何主机决策的固定嵌套设备图在五种放置方案的全部 60 种配置中都更慢。这些研究共同为 GPU 智能体控制确立了两个可测量的门限：截止期可行的批次供给和观测放置。需要一个联合的有限在线运行时来测量 A、CPU 转移和服务水平收益。
cs.AI / 57 / 2608.11500
Strengthening Full Justified Representation: Efficient Verification and Computation
强化完全正当代表：高效验证与计算
Abstract
Full justified representation (FJR) is among the strongest known satisfiable proportionality axioms for approval-based committee elections. Recent work has shown that an FJR committee can be found in polynomial time, but verifying whether a given committee satisfies FJR remains coNP-complete. We introduce FJR+, a strict strengthening of FJR and EJR+ that can be verified and satisfied in polynomial time. We then analyze the Residual-Budget Greedy (RBG) algorithm and prove that it selects a partial committee such that every size-$k$ completion satisfies FJR+. This freedom allows us to use sequential Phragmén to obtain a priceable completion. The resulting rule always satisfies FJR+ and the sub-core, and it is priceable whenever at least $k$ candidates receive an approval. We also obtain a Droop-quota version of FJR+. Finally, we extend FJR+ to approval-based participatory budgeting with arbitrary project costs. A project-specific version of RBG computes this property in polynomial time and can be continued to a priceable outcome satisfying a cost-based version of the sub-core.
Chinese Translation
完全正当代表（FJR）是基于赞同投票的委员会选举中已知最强的可满足比例性公理之一。近期研究表明，可以在多项式时间内找到一个 FJR 委员会，但验证给定委员会是否满足 FJR 仍是 coNP-完全问题。我们引入 FJR+，它是 FJR 和 EJR+ 的一个严格加强，并且可以在多项式时间内被验证和满足。然后，我们分析了剩余预算贪心（RBG）算法，并证明它选出一个部分委员会，使得每个规模为 $k$ 的完成方案都满足 FJR+。这一自由度使我们能够使用顺序 Phragmén 方法得到一个可定价的完成方案。由此得到的规则始终满足 FJR+ 和子核心（sub-core），并且只要至少 $k$ 名候选人获得赞同，它就是可定价的。我们还得到了 FJR+ 的 Droop 配额版本。最后，我们将 FJR+ 扩展到具有任意项目成本的基于赞同投票的参与式预算中。RBG 的一个针对项目的版本可在多项式时间内计算该性质，并可继续运行以获得一个可定价的结果，该结果满足基于成本的子核心版本。
cs.AI / 58 / 2608.12125
Do LLMs Take Care of Their Own? Similarity Signals Can Induce Cooperation
LLM会照顾自己的同类吗？相似性信号能够诱导合作
Abstract
As LLM-based agents with user-instructed goals are becoming widely deployed, they increasingly encounter each other in strategic interactions, and face challenges of finding mutually beneficial outcomes. Prior literature has argued that cooperation problems such as the Prisoner's Dilemma are resolvable in settings where agents know they follow very similar decision making patterns, as for example in monocultural AI ecosystems. Following that line of work, this paper introduces the first framework for evaluating LLM decision making when agents are provided with graded similarity signals. Among our findings, we establish that different LLM models vary drastically in how they navigate similarity signals, with some modern models showing consistent behavior across cooperation problems, payoff structures, and prompt framing. Perhaps surprisingly, our experiments also show that the dataset based on which the similarity signal is computed has small to no impact on induced cooperation, and that LLM models systematically self-identify as highly similar when asked to evaluate another model's chain-of-thought reasoning by themselves. Finally, we develop an LLM-behavioral-game-theoretic model that captures some of their reasoning rationale, and show that it can support cooperative outcomes in equilibrium under sufficiently high similarity scores.
Chinese Translation
随着基于LLM的具有用户指定目标的智能体被广泛部署，它们在策略互动中越来越多地相互遭遇，并面临寻找互利结果的挑战。先前文献认为，诸如囚徒困境之类的合作问题在智能体知道它们遵循非常相似的决策模式的环境中是可以解决的，例如在单一文化的AI生态系统中。沿着这一研究方向，本文提出了第一个框架，用于在智能体获得分级相似性信号时评估LLM的决策。在我们的发现中，我们证实不同的LLM模型在处理相似性信号的方式上差异巨大，其中一些现代模型在合作问题、收益结构和提示框架上表现出一致的行为。也许令人惊讶的是，我们的实验还表明，用于计算相似性信号的数据集对诱导的合作影响很小甚至没有影响，并且当LLM模型被要求自行评估另一个模型的思维链推理时，它们会系统性地将自身识别为高度相似。最后，我们开发了一个LLM行为博弈论模型，该模型捕捉了它们的一些推理逻辑，并表明在足够高的相似性得分下，它可以在均衡中支持合作结果。
cs.AI / 59 / 2608.11322
Socioduality: A Relational Process Framework for Human-AI Interaction
社会二元性（Socioduality）：一种面向人类-AI交互的关系过程框架
Abstract
Human-AI research often evaluates individual capabilities, combined performance, or final outputs, but these approaches do not preserve how one party's response becomes part of the conditions under which the other party's next contribution is formed. This article introduces socioduality, a sequential, reciprocal, and history-carrying relational process between two distinguishable parties in which a response from one party becomes part of the observable conditions under which the other party's subsequent contribution, judgement, decision, or action is formed. Specified for human-AI dyads, the construct uses nested units: moves, confirmed sociodual episodes, linked pathways, and the broader interaction container. A minimum episode A1-B1-A2 requires evidence of response contingency and return contingency; candidate episodes are classified as confirmed, non-sociodual, or indeterminate before secondary coding of response orientation and substantive contribution re-formation. Three propositions address history-conditioned formation, pathway divergence, and robustness differences among endpoint-equivalent pathways. A frozen operational protocol was calibrated on three previously unseen natural human-AI records through two separately executed model-based evaluator series. Move and candidate reconstruction converged exactly in two cases and differed by one local multimodal unitisation decision in the third; remaining disagreement was concentrated at return-contingency boundaries. Socioduality therefore provides a bounded and empirically tractable process construct for analysing how human and AI contributions are formed through interaction while preserving pathway information that endpoint-centred analysis cannot recover.
Chinese Translation
人类-AI研究通常评估个体能力、联合表现或最终输出，但这些方法未能保留一方的回应如何成为另一方下一次贡献得以形成的条件的一部分。本文引入了社会二元性（socioduality），它是两个可区分方之间的一种顺序性、相互性且承载历史的关系过程，在该过程中，一方的回应成为另一方随后的贡献、判断、决策或行动得以形成的可观察条件的一部分。在针对人类-AI二元体进行具体化时，该构念使用嵌套单元：动作（moves）、确认的社会二元片段（confirmed sociodual episodes）、联结路径（linked pathways）以及更广泛的交互容器（interaction container）。一个最小片段A1-B1-A2要求具备回应相依性（response contingency）和返回相依性（return contingency）的证据；候选片段在对回应取向和实质性贡献再形成进行二级编码之前，被分类为已确认、非社会二元或不确定。三个命题涉及历史条件化形成、路径分化和终点等价路径之间的稳健性差异。一个冻结的操作化协议经由两个分别执行的基于模型的评估者系列，在三条先前未见过的自然人类-AI记录上进行了校准。动作与候选片段的重建在两个案例中完全一致，在第三个案例中仅因一个局部的多模态单元划分决定而不同；其余分歧集中在返回相依性边界处。因此，社会二元性为分析人类与AI的贡献如何通过互动形成提供了一个有边界且经验上可处理的过程构念，同时保留了以终点为中心的分析无法恢复的路径信息。
cs.AI / 60 / 2608.11980
HCGRec: Hint-Conditioned Generative Recommendation with Semantic IDs
HCGRec：基于语义ID的提示条件生成式推荐
Abstract
Semantic-ID generative recommenders represent each item as a short sequence of discrete semantic tokens and predict the next item by autoregressively generating this token sequence. This paradigm enables a unified generation interface for item IDs, histories, and item text, but it also creates a structured optimization bottleneck during reward-based post-training: when an early semantic token enters the wrong branch of the item-token space, finite rollout groups rarely reach the ground-truth item, so group-relative optimization receives identical zero rewards and produces no useful advantage. We propose Hint-Conditioned Generative Recommendation (HCGRec), a semantic-ID generative recommendation framework that recovers learning signal for such hard training instances. HCGRec diagnoses each instance with checkpoint rollouts and supplies a minimal target-prefix hint only when the current generator cannot reach the correct item. The model then generates the unhinted suffix under the hinted semantic branch, turning zero-reward groups into informative comparisons over item-token completions. Hinting also changes token identity: hinted prefix tokens are oracle-provided item context, while unhinted suffix tokens are sampled generation actions. We therefore introduce hint-aware credit decomposition, using supervised learning to preserve item-semantic and prefix-structure alignment for hinted tokens and GRPO to optimize the sampled suffix. Experiments on sequential recommendation benchmarks show that HCGRec substantially improves over supervised fine-tuning and vanilla reward-based post-training, while reducing zero-advantage training samples from over 70% to below 20%. The code is accessible at https://github.com/WncFht/GRec.
Chinese Translation
语义ID生成式推荐器将每个物品表示为一段由离散语义token组成的短序列，并通过自回归地生成该token序列来预测下一个物品。该范式为物品ID、历史记录和物品文本提供了统一的生成接口，但在基于奖励的后训练过程中也产生了结构化的优化瓶颈：当较早的语义token进入物品-token空间的错误分支时，有限的rollout组很少能达到真实物品，因此组相对优化只能获得相同的零奖励，无法产生有用的优势。我们提出了提示条件生成式推荐（HCGRec），一个面向语义ID的生成式推荐框架，可为此类困难训练实例恢复学习信号。HCGRec通过检查点rollout对每个实例进行诊断，并仅在当前生成器无法达到正确物品时，提供一个最小的目标前缀提示。随后，模型在已提示的语义分支下生成未被提示的后缀，从而将零奖励组转变为关于物品-token补全的有信息量比较。提示还改变了token身份：被提示的前缀token是oracle提供的物品上下文，而未被提示的后缀token是采样得到的生成动作。因此，我们引入了提示感知的信用分解：使用监督学习来保持被提示token的物品语义和前缀结构对齐，并使用GRPO来优化采样得到的后缀。在序列推荐基准上的实验表明，HCGRec相较于监督微调和普通的基于奖励的后训练有显著提升，同时将零优势训练样本从70%以上降至20%以下。代码可在 https://github.com/WncFht/GRec 获取。
cs.AI / 61 / 2608.11340
Self-evolving network verifiers
自演化网络验证器
Abstract
Symbolic network verifiers can reason about correctness across vast spaces of routing inputs and failures, but only for the protocols and features an expert has encoded by hand. Creating and maintaining a faithful model of the control plane is both difficult and never-ending, since no written source specifies perfectly what a network does: vendor implementations deviate from the RFCs, and behaviour shifts with releases. The burden of constant upkeep ultimately keeps verification out of many networks that need it. We argue that the model should instead evolve automatically to faithfully capture the actual network behaviour. To achieve that, we leverage the only source that specifies it unambiguously: the router software itself. In a counterexample-guided loop, a coding agent proposes extensions to the verifier's symbolic encoding, while a trusted oracle (e.g., emulated routers) supplies the ground-truth routing state. The agent iteratively refines the network model using each disagreement with the oracle. As early evidence, a prototype of this system taught a 3,000-line SMT-based verifier three features it did not support: OSPF areas, BGP route reflection, and L3VPN over EVPN, converging autonomously on models that match the oracle, even noticing vendor-specific behaviour. Automating model growth shifts the hard problem from writing verification systems to systematically testing them; we propose a research agenda for trusting and harnessing automatically evolved verifiers.
Chinese Translation
符号化网络验证器能够在巨大的路由输入和故障空间上推理正确性，但仅限于专家手工编码的协议和特性。创建并维护一个忠实的控制平面模型既困难又永无止境，因为没有任何书面来源能完美规定网络的行为：厂商实现偏离 RFC，行为也随版本发布而变化。持续维护的负担最终使验证无法进入许多需要它的网络。我们认为，模型应当反过来自动演化，以忠实刻画实际网络行为。为实现这一目标，我们利用唯一能无歧义规定该行为的来源：路由器软件本身。在反例引导的循环中，编码代理提出对验证器符号化编码的扩展，而可信预言机（例如仿真路由器）提供真实路由状态。该代理利用每次与预言机的分歧迭代地细化网络模型。作为早期证据，该系统的一个原型教会了一个基于 SMT 的 3,000 行验证器三项它原本不支持的特性：OSPF 区域、BGP 路由反射以及 EVPN 上的 L3VPN，自主收敛到与预言机匹配的模型，甚至注意到厂商特定行为。自动化模型增长将难题从编写验证系统转变为系统地测试它们；我们提出了一个研究议程，以信任并利用自动演化的验证器。
cs.AI / 62 / 2608.11451
Herding End-to-End Autonomous Driving via Neuro-Symbolic Safety Guards
通过神经符号安全守卫引导端到端自动驾驶
Abstract
Modern end-to-end driving agents can achieve high average performance yet still violate basic traffic rules that a human driver would never miss. The reason is structural: they learn statistical patterns rather than the physical conditions that guarantee safe driving, leaving their decision-making process opaque and safety constraints unenforced. We introduce a neuro-symbolic safety guard, a lightweight module that attaches to the final command interface of an already-trained agent. Immediately before a command reaches the vehicle, it checks the command against explicit safety rules and, only when necessary, replaces it with the nearest safe alternative. Each intervention is directly executable and traceable to the rule that triggered it, while the guard itself requires no retraining and adds no learned component. Evaluated on the long-tail benchmarks Fail2Drive and Bench2Drive using the state-of-the-art TransFuser v6 (TFv6) as a case study, the guard improves Success Rate by 15% and reduces safety-critical collisions by up to 53%, while preserving the original Driving Score.
Chinese Translation
现代端到端驾驶智能体可以实现较高的平均性能，但仍然会违反人类驾驶员绝不会忽视的基本交通规则。原因是结构性的：它们学习的是统计模式，而不是保证安全驾驶的物理条件，这使得其决策过程不透明，安全约束也未被执行。我们提出了一种神经符号安全守卫，这是一个轻量级模块，附加在已训练智能体的最终命令接口上。在命令到达车辆之前，它会立即根据显式安全规则检查该命令，并且仅在必要时将其替换为最接近的安全替代命令。每次干预都是可直接执行的，并且可追溯到触发该干预的规则，而守卫本身不需要重新训练，也不添加任何学习组件。在使用最先进的 TransFuser v6（TFv6）作为案例研究的长尾基准 Fail2Drive 和 Bench2Drive 上进行评估时，该守卫将成功率提高了15%，并将安全关键碰撞减少了最多53%，同时保持了原有的驾驶得分。
cs.AI / 63 / 2608.11521
Keep the Future, Drop the Rollout: RIFT for World Action Models
Abstract
World action models (WAMs) condition robot actions on predicted futures, but iterative video rollout increases deployment latency. We ask whether action generation requires the evolving rollout trajectory or only its future representation. Across four WAMs on all 40 LIBERO tasks, paired closed-loop interventions show that masking or reassigning future-cache values changes execution and reduces success, indicating sensitivity to future values and their assigned positions. For Joint and Cosmos-2, however, replaying one fixed final-clean key/value (K/V) cache nearly preserves unmodified execution, with $1.7$ to $1.9$~cm end-effector average displacement error and $97.9\%$ to $98.2\%$ success. This separates cache consumption from production: these models can reuse a fixed cache but still require iterative rollout to construct it. We therefore propose RIFT (\emph{Rollout-free Imagination via Future Tokens}), which uses learned anticipation tokens to construct a complete future K/V cache in one backbone pass while retaining the original future-read interface. On LIBERO, RIFT achieves $98.8\%$ success, close to rollout-based Joint, IDM, and LingBot-VA at $98.4\%$ to $98.6\%$, while reducing action-chunk latency by $68.2\%$ to $89.1\%$. On RoboTwin~2.0, RIFT reaches $92.9/92.6\%$ on clean/randomized scenes, the highest observed among the evaluated methods. These results support rollout-free future conditioning without iterative video generation at deployment.
cs.AI / 64 / 2608.11580
RoadWeaver: Large-Scale Lane-Level HD Map Generation from Scratch for Autonomous Driving Simulation
RoadWeaver：面向自动驾驶仿真的大规模车道级高精地图从零开始生成
Abstract
Autonomous driving simulation requires diverse and scalable lane-level HD maps to support long-horizon evaluation across complex road networks. Existing approaches either rely on handcrafted or reconstructed real-world maps, which limits scalability, or generate only local road structures rather than complete HD maps. We present RoadWeaver, a coarse-to-fine framework for from-scratch generation of diverse, large-scale HD maps. RoadWeaver first synthesizes a global road layout, expands it into a connected road network, and then constructs lane-level geometry with topologically consistent lane connectivity. Experimental results show that RoadWeaver achieves a 99.8\% reachability, a 10.7\% dead-end ratio, and an endpoint alignment error of 0.24 m. Compared with SOTA generation methods, it reduces endpoint alignment error by 94.4\% while generating complete HD maps in 1.39--3.50 s. The generated maps can be directly deployed in driving simulators, providing scalable simulation environments for future closed-loop evaluation of autonomous driving systems. The training code and an out-of-the-box implementation of RoadWeaver will be released upon acceptance.
Chinese Translation
自动驾驶仿真需要多样化且可扩展的车道级高精地图，以支持跨复杂道路网络的长时域评估。现有方法要么依赖手工制作或重建的真实世界地图，这限制了可扩展性；要么只生成局部道路结构，而不是完整的高精地图。我们提出了 RoadWeaver，一个从零生成多样化、大规模高精地图的由粗到细框架。RoadWeaver 首先合成全局道路布局，将其扩展为连通的道路网络，然后构建具有拓扑一致车道连接关系的车道级几何。实验结果表明，RoadWeaver 实现了 99.8% 的可达率、10.7% 的断头路比例以及 0.24 m 的端点对齐误差。与最先进（SOTA）的生成方法相比，它在 1.39--3.50 秒内生成完整高精地图的同时，将端点对齐误差降低了 94.4%。生成的地图可直接部署于驾驶仿真器中，为未来自动驾驶系统的闭环评估提供可扩展的仿真环境。RoadWeaver 的训练代码和开箱即用实现将在论文录用后发布。
cs.AI / 65 / 2608.11739
G0.5: One Autoregressive Stream for Robot Reasoning and Action
G0.5：面向机器人推理与动作的单条自回归流
Abstract
The prevailing recipe for Vision-Language-Action (VLA) models couples a pretrained VLM with a separately trained flow-matching action expert. This makes the VLM a context encoder rather than a decision-maker. We introduce G0.5, a pretrained autoregressive VLA in which a single transformer decoder emits reasoning and action tokens under a single objective. Three components make this tractable at foundation-model scale: a learnable cross-embodiment action tokenizer that maps heterogeneous robot actions into a shared vocabulary; a native chain-of-thought stream interleaving task decomposition, object grounding, and action hints with action tokens; and a visual memory module that injects multi-second history through the vision encoder. Because reasoning and action share a single set of weights, the pretrained VLM's capabilities carry over to physical behavior: the model follows instructions closely, and prompts directly steer action granularity, task horizon, and out-of-distribution scene handling without further training. Pretrained on a large collection of robot datasets together with VQA samples, G0.5 surpasses state-of-the-art models across 7 independent regimes: real-world fine-tuning on R1lite and R1pro robots (76.7\% vs.\ 53.3\% for $π_{0.5}$ and 24.4\% for GR00T-N1.7), the 2025 BEHAVIOR Challenge on 50 long-horizon household mobile manipulation tasks using a generalist policy (31.4\% vs.\ 26.3\% for $π_{0.5}$ and 26.1\% for the challenge winner), DROID post-training followed by zero-shot transfer to an unseen environment and objects (82.5\%), a language-following Pick-and-Place benchmark, LIBERO (98.9\%), RoboTwin 2.0 (93.3\%), and SimplerEnv-Bridge (87.3\%).
Chinese Translation
当前主流的视觉-语言-动作（VLA）模型方案将预训练的 VLM 与一个单独训练的流匹配动作专家耦合在一起。这使得 VLM 成为一个上下文编码器，而不是决策器。我们提出 G0.5，一个预训练的自回归 VLA，其中单个 Transformer 解码器在单一目标下输出推理 token 和动作 token。三个组件使这一方法在基础模型规模上可行：一个可学习的跨具身动作分词器，将异构的机器人动作映射到共享词表；一条原生思维链流，将任务分解、物体定位和动作提示与动作 token 交错输出；以及一个视觉记忆模块，通过视觉编码器注入多秒历史信息。由于推理和动作共享同一组权重，预训练 VLM 的能力会迁移到物理行为上：模型能紧密遵循指令，并且提示词可以直接控制动作粒度、任务时域和分布外场景处理，而无需进一步训练。G0.5 在大量机器人数据集以及 VQA 样本上预训练后，在 7 个独立评测体系中超越了最先进模型：在 R1lite 和 R1pro 机器人上的真实世界微调（76.7% 对比 $π_{0.5}$ 的 53.3% 和 GR00T-N1.7 的 24.4%）；在使用通用策略的 50 个长程家庭移动操作任务上的 2025 BEHAVIOR 挑战赛（31.4% 对比 $π_{0.5}$ 的 26.3% 和挑战赛冠军的 26.1%）；DROID 后训练，随后零样本迁移到未见环境和物体（82.5%）；一个遵循语言的抓取-放置基准；LIBERO（98.9%）；RoboTwin 2.0（93.3%）以及 SimplerEnv-Bridge（87.3%）。
cs.AI / 66 / 2608.12063
Learning Loco-Manipulation From SMPC Demonstrations With Sparse Offline-to-Online RL
通过稀疏离线到在线强化学习从SMPC示范中学习移动操作
Abstract
Integrating locomotion and manipulation is essential for robot autonomy, but scaling standard Reinforcement Learning (RL) to complex tasks is severely bottlenecked by the slow, manual process of dense reward shaping. To bypass this limitation, we leverage Sample-based Model Predictive Control (SMPC) entirely in simulation as an automated, rapidly tunable expert to generate massive offline datasets. Because this data solves the fundamental exploration problem, we can train an off-policy RL agent using purely sparse task rewards, drastically reducing the time required to learn new skills and eliminating the need for manual tuning. Integrating this high-level agent with a low-level dynamic stability controller yields more optimal behaviors that strictly align with true task objectives, ultimately allowing the learned policies to surpass the original optimal control teacher. We validate the robustness of this sim-to-real framework by successfully deploying complex loco-manipulation skills across different morphologies, including an arm-equipped Spot quadruped and a G1 humanoid.
Chinese Translation
将运动与操作相结合对机器人自主性至关重要，但将标准强化学习（RL）扩展到复杂任务受到密集奖励塑形缓慢且需人工手动调试的严重制约。为了绕过这一限制，我们完全在仿真中采用基于采样的模型预测控制（SMPC）作为自动化、可快速调节的专家，以生成大规模离线数据集。由于这些数据解决了根本的探索问题，我们可以仅使用稀疏任务奖励训练离策略RL智能体，大幅减少学习新技能所需时间，并消除人工调参需求。将该高层智能体与底层动态稳定控制器相结合，能产生更优的行为，并严格符合真实任务目标，最终使学习到的策略超越最初的优化控制教师。我们通过在不同形态机器人上成功部署复杂的移动操作技能，验证了这一从仿真到现实框架的鲁棒性，这些机器人包括配备手臂的Spot四足机器人和G1人形机器人。
cs.AI / 67 / 2608.12198
Learning-Based Behavior Planning for Automated Driving: Real-World Integration and Deployment
基于学习的自动驾驶行为规划：真实世界集成与部署
Abstract
Recent research in machine and deep learning has shown the potential of learningbased motion planning approaches to improve the driving behavior of automated vehicles, especially in complex environments. However, their complex nature and lack of transparency can hinder explainability and trustworthiness and complicate safety assurance. Motivated by these challenges, we propose a hybrid planning architecture that combines the advantages of machine learning with the verifiability and the determinism of classical approaches. Specifically, we developed a deep neural network to interpret complex traffic scenes and propose driving behavior, while an optimization-based supervision layer validates this proposal and enforces explicit drivability and safety constraints. We evaluate the learned planner's driving behavior in open-loop studies on real-world urban data, discuss system integration aspects for stable closed-loop operation, and report results from real-world deployment on our research vehicle karl..
Chinese Translation
近期机器学习和深度学习研究已经表明，基于学习的运动规划方法在改善自动驾驶车辆驾驶行为方面具有潜力，尤其是在复杂环境中。然而，其复杂性和缺乏透明性可能阻碍可解释性和可信度，并使安全保证复杂化。受这些挑战的驱动，我们提出了一种混合规划架构，该架构将机器学习的优势与经典方法的可验证性和确定性相结合。具体来说，我们开发了一个深度神经网络来解读复杂交通场景并提出驾驶行为，同时，一个基于优化的监督层对该提议进行验证，并强制执行明确的可驾驶性与安全约束。我们基于真实世界城市数据，通过开环研究评估了学习得到的规划器的驾驶行为，讨论了实现稳定闭环运行的系统集成方面，并报告了在研究车辆 karl.. 上进行真实世界部署的结果。
cs.AI / 68 / 2608.12236
How Organizations Use AI: Evidence from ChatGPT
组织如何使用人工智能：来自 ChatGPT 的证据
Abstract
We study how organizations use frontier generative AI by linking ChatGPT Enterprise account records to usage, worker roles, task classifications, and public-company financial data through March 2026. These linked data enable a privacy-preserving analysis of adoption, worker roles, and message-level tasks at scale: for instance, the worker-level sample we analyze at the six-month adoption horizon includes over 1,500 organizations and over 17 million messages. We document four facts about enterprise AI adoption and use. First, ChatGPT Enterprise usage has grown rapidly due to a combination of new firm adoption and growing intensity among existing adopters. Second, U.S.-based public company adoption is concentrated among larger, more valuable, and more R&D- and SG&A-intensive firms. Third, active use within adopting firms spans job functions and seniority levels, with especially high usage intensity among early-career workers. Fourth, ChatGPT Enterprise usage encompasses a broad range of knowledge work tasks, including writing, technical work, communication, and information synthesis. In aggregate, these results suggest that firms differ widely in the speed, breadth and purpose of their enterprise AI adoption, and that they are still actively learning how to integrate AI into organizational workflows.
Chinese Translation
我们通过将 ChatGPT Enterprise 账户记录与截至 2026 年 3 月的使用情况、员工角色、任务分类以及上市公司财务数据相关联，研究组织如何使用前沿生成式人工智能。这些关联数据使得能够在保护隐私的前提下大规模分析采用情况、员工角色和消息级任务：例如，我们在六个月采用时间窗内分析的员工层面样本包括 1,500 多个组织和超过 1,700 万条消息。我们记录了关于企业人工智能采用和使用的四个事实。第一，由于新企业的采用以及现有采用者使用强度的不断提高，ChatGPT Enterprise 的使用量迅速增长。第二，美国上市公司的采用集中在规模更大、价值更高、研发和销售及管理费用强度更高的企业。第三，采用企业内部的积极使用跨越不同岗位职能和资历层级，其中早期职业员工的使用强度尤其高。第四，ChatGPT Enterprise 的使用涵盖广泛的知识工作任务，包括写作、技术工作、沟通和信息综合。总体而言，这些结果表明，企业在采用企业人工智能的速度、广度和目的上差异很大，并且它们仍在积极学习如何将人工智能融入组织工作流程。
cs.AI / 69 / 2608.11627
Deep Learning Based Relative Transfer Matrix Estimation for Multiple Sources and Multiple Microphones
基于深度学习的多声源多麦克风相对传递矩阵估计
Abstract
The Relative Transfer Matrix (ReTM), recently introduced as a generalization of the relative transfer function for multiple receivers and sources, shows promising performance when applied to speech enhancement in noisy environments. Estimating the ReTM of sound sources by exploiting the covariance matrices of multichannel recordings is highly beneficial for practical applications and, to date, remains the only proposed approach. This paper investigates deep learning-based ReTM estimation. We propose three novel supervised learning frameworks using time and short-time frequency transform domain convolutional networks, and a Long Short-Term Memory-based recurrent neural network. Experimental results demonstrate that the proposed models achieve more accurate estimation of the ReTM using five objective metrics compared to the covariance-based method. We also show the effectiveness of the proposed frameworks for speech enhancement, achieving performance on par with the baseline method.
Chinese Translation
相对传递矩阵（ReTM）最近被提出，作为针对多个接收器和多个声源的相对传递函数的推广，在应用于噪声环境中的语音增强时表现出有前景的性能。通过利用多通道录音的协方差矩阵来估计声源的 ReTM 对实际应用非常有益，并且迄今为止仍是唯一被提出的方法。本文研究了基于深度学习的 ReTM 估计。我们提出了三种新颖的有监督学习框架，分别使用时间域卷积网络、短时频率变换域卷积网络和基于长短期记忆的循环神经网络。实验结果表明，与基于协方差的方法相比，所提出的模型在五个客观指标上实现了更准确的 ReTM 估计。我们还展示了所提出的框架在语音增强方面的有效性，取得了与基线方法相当的性能。
cs.AI / 70 / 2608.12227
Domain-Aware Lightweight Spectral-Grouped Convolutions for Hyperspectral Fish Freshness Classification
面向高光谱鱼类新鲜度分类的域感知轻量级光谱分组卷积
Abstract
Hyperspectral imaging (HSI) offers nondestructive assessment of fish freshness by detecting biochemical alterations across spectral bands. However, conventional deep learning approaches do not fully address the particular characteristics of HSI data, such as spectral dominance over spatial textures, ordinal label structure, and a small number of training samples. We propose SGNet (Spectral-Grouped Network), a lightweight architecture that separates spectral and spatial feature extraction using grouped convolutions and a depthwise spatial pathway. A dual attention mechanism that couples channel-wise squeeze-and-excitation with spatial gating adaptively highlights informative features. SGNet achieves 97.8% classification accuracy and 0.64 days mean absolute error (MAE) with just 4.75M parameters when tested on our newly developed 16-day refrigerator-stored salmon fillet dataset. Ablation studies validate the contribution of each component, while comparisons demonstrate a five- to eighteen-fold parameter reduction relative to ResNet-50 and Vision Transformers. Our findings indicate that domain-aware design supports precise, real-time freshness prediction for industrial implementation.
Chinese Translation
高光谱成像（HSI）通过检测各光谱波段中的生化变化，提供对鱼类新鲜度的无损评估。然而，传统深度学习方法并未充分应对高光谱数据特有的特性，例如光谱信息相对于空间纹理的主导地位、有序标签结构以及训练样本数量少。我们提出SGNet（光谱分组网络），这是一种轻量级架构，通过分组卷积和逐深度空间通路将光谱特征提取与空间特征提取分离。一种将逐通道压缩-激励与空间门控相结合的双重注意力机制，能够自适应地突出信息性特征。在我们新构建的16天冰箱储存三文鱼片数据集上测试时，SGNet仅以4.75M参数实现了97.8%的分类准确率和0.64天的平均绝对误差（MAE）。消融研究验证了每个组件的贡献，同时比较结果表明，相对于ResNet-50和视觉Transformer，参数量减少了五到十八倍。我们的研究结果表明，域感知设计有助于实现面向工业应用的精确、实时的新鲜度预测。
cs.AI / 71 / 2608.11540
A Conceptual Framework for Enhancing Workforce Readiness for Smart Manufacturing in the AI Era
AI时代提升智能制造劳动力就绪度的概念框架
Abstract
The convergence of artificial intelligence (AI), Industrial Internet of Things, cyber-physical systems, and advanced robotics is reshaping manufacturing faster than engineering curricula can adapt, widening the gap between the competencies required on the shop floor and those delivered by traditional engineering and technology education. This paper proposes a Workforce Readiness Level (WRL) framework, which adapts the Technology Readiness Level scale into nine progressive competency stages and a four-pillar rubric, digital and AI literacy, cyber-physical systems fluency, human-machine collaboration, and data-driven decision making, aggregated through a composite stage score and a cohort-level workforce-readiness index under a ``no-thin-pillar'' rule. The framework is instantiated at a university smart-manufacturing teaching laboratory and draws on 89 sponsored capstone projects delivered over four semesters, four of which are analyzed in depth. Four pillars jointly span the relevant ABET student outcomes. Across the highlighted cohorts the workforce-readiness index ranged from 5.2 to 6.4, and the no-thin-pillar rule was diagnostically informative in three of the four cases and the binding certification constraint in one, repeatedly surfacing cyber-physical and data-driven-decision gaps concealed behind strong analytics profiles; advancement to the highest stages was gated by industry-embedded experience rather than additional coursework. WRL offers educators, accreditation bodies, and regional workforce systems a common, evidence-based instrument for diagnosing and advancing workforce readiness; future work will calibrate pillar weights and test reliability and predictive validity.
Chinese Translation
人工智能（AI）、工业物联网、信息物理系统与先进机器人技术的融合正在以快于工程课程调整速度的方式重塑制造业，扩大了车间所需能力与传统工程和技术教育所培养能力之间的差距。本文提出一个劳动力就绪度水平（WRL）框架，该框架将技术就绪度水平量表改编为九个递进的胜任力阶段和一个四支柱评估标准，即数字与AI素养、信息物理系统流畅性、人机协作和数据驱动决策，并通过一个综合阶段得分和群体层面的劳动力就绪度指数在“无薄弱支柱”规则下进行汇总。该框架在一所大学的智能制造教学实验室中得到实例化，并利用了四个学期内交付的89个受资助的顶点项目，其中四个项目被深入分析。四大支柱共同覆盖了相关的ABET学生成果。在所突出的各群体中，劳动力就绪度指数范围为5.2至6.4，且“无薄弱支柱”规则在四个案例中的三个中具有诊断信息价值，在其中一个案例中成为具有约束力的认证限制条件，反复揭示出隐藏在强大分析能力背后的信息物理和数据驱动决策方面的差距；向最高阶段的进阶受到行业嵌入式经验而非额外课程学习的制约。WRL为教育工作者、认证机构和区域劳动力系统提供了一种通用的、基于证据的工具，用于诊断和提升劳动力就绪度；未来工作将校准支柱权重，并检验信度和预测效度。
cs.AI / 72 / 2608.11860
Two-Stage Deformable-Convolutional Inverse Design of Nanophotonic Absorbers from Optical Spectra
基于光谱的纳米光子吸收器两阶段可变形卷积逆向设计
Abstract
Data-driven inverse design enables efficient generation of nanophotonic structures with prescribed optical responses, but spectrum-to-geometry mapping remains challenging due to non-uniqueness and fine geometric features. This work presents a two-stage deformable-convolutional framework for reconstructing metal--insulator--metal resonator geometries from 80-dimensional absorption spectra. The spectrum is projected to a $150\times4\times4$ latent representation and decoded into a $64\times64$ resonator mask. Training combines supervised reconstruction with least-squares adversarial refinement initialized from the best supervised checkpoint. A three-run ablation compares deformable convolution with plain convolution, involution, Dynamic Conv, and ODConv under the same architecture. The proposed model achieves $20.79\pm0.31$~dB PSNR and $0.8501\pm0.0082$ SSIM, improving over plain convolution by 2.16~dB and 0.0831, respectively. It further achieves Dice $0.9623\pm0.0027$, IoU $0.9342\pm0.0038$, and boundary F-score $0.9550\pm0.0027$. Spectral consistency evaluated using a frozen forward surrogate yields RMSE $0.0805\pm0.0013$ and $R^2=0.7923\pm0.0065$. Learned offsets show stronger adaptive sampling at coarse and intermediate decoder stages. Overall, deformable sampling with supervised initialization and adversarial refinement improves spectrum-conditioned geometry reconstruction.
Chinese Translation
数据驱动的逆向设计能够高效生成具有指定光学响应的纳米光子结构，但由于非唯一性和精细几何特征，光谱到几何结构的映射仍然具有挑战性。本工作提出了一种两阶段可变形卷积框架，用于从80维吸收光谱重建金属-绝缘体-金属谐振器几何结构。光谱被投影为 $150\times4\times4$ 潜在表示，并解码为 $64\times64$ 谐振器掩模。训练将监督重建与从最佳监督检查点初始化的最小二乘对抗细化相结合。三次运行的消融实验在同一架构下比较了可变形卷积与普通卷积、内卷、动态卷积和ODConv。所提模型达到 $20.79\pm0.31$~dB PSNR 和 $0.8501\pm0.0082$ SSIM，分别比普通卷积提高2.16~dB和0.0831。它还进一步达到 Dice $0.9623\pm0.0027$、IoU $0.9342\pm0.0038$ 和边界F-score $0.9550\pm0.0027$。使用冻结的正向代理模型评估的光谱一致性得到 RMSE $0.0805\pm0.0013$ 和 $R^2=0.7923\pm0.0065$。学习到的偏移量在粗糙和中间解码器阶段表现出更强的自适应采样。总体而言，具有监督初始化和对抗细化的可变形采样改善了以光谱为条件的几何结构重建。
cs.AI / 73 / 2608.11733
A 12-CNOT Double Qubit Excitation Gate
一种 12-CNOT 双量子比特激发门
Abstract
Effective implementation of high-level quantum gates is essential for practical quantum computing. To the best of our knowledge, we present the first reported 12-CNOT decomposition of the double qubit excitation operator, improving upon state-of-the-art (SOTA) implementations with 13 CNOTs. Our new circuit has the lowest CNOT count (12), lowest CNOT depth (10), and lowest total circuit depth (16) among all the previous SOTA circuits. Further, we only added 2 extra one-qubit gates compared to the lowest one-qubit gate count (11) among the previous SOTA circuits.
Chinese Translation
高效实现高级量子门对于实用量子计算至关重要。据我们所知，我们首次报告了双量子比特激发算符的 12-CNOT 分解，优于具有 13 个 CNOT 的最新（SOTA）实现。我们的新电路在所有先前的 SOTA 电路中具有最低的 CNOT 数量（12）、最低的 CNOT 深度（10）和最低的总电路深度（16）。此外，与先前 SOTA 电路中最低的单量子比特门数量（11）相比，我们仅增加了 2 个额外的单量子比特门。
cs.AI / 74 / 2608.11884
CoQui: A Coordinate-Conditioned Quantum Implicit Generative Adversarial Network for End-to-End Image Generation
CoQui：一种用于端到端图像生成的坐标条件量子隐式生成对抗网络
Abstract
Quantum generative adversarial networks (QGANs) have attracted increasing attention for image generation using parameterized quantum circuits. Existing amplitude-based approaches face two key limitations: pixel locations are typically encoded by computational-basis indices or address qubits, causing quantum resources to grow with image resolution; meanwhile, jointly decoding many pixels from normalized quantum states introduces probability competition among pixels and limits precise pixel-wise control. To address these issues, we reformulate quantum image generation as coordinate-conditioned implicit function learning. Our method takes spatial coordinates and latent variables as inputs, uses a classical embedding network to generate input-dependent circuit parameters, and evaluates a variational quantum circuit at each coordinate. Pixel intensities are directly obtained from the expectation value of a dedicated color qubit, and a complete image is generated by querying all spatial coordinates. This design decouples image resolution from address-qubit requirements and avoids shared probability-normalization constraints across pixels. We further design a specialized variational quantum circuit to provide structural inductive bias for coordinate-conditioned generation. Simulated experiments on two benchmark datasets show that our method outperforms FRQI-based generation and PQWGAN in visual and quantitative quality while using fewer qubits, and also achieves better generation quality than the corresponding classical baseline.
Chinese Translation
量子生成对抗网络（QGANs）因使用参数化量子电路进行图像生成而受到越来越多的关注。现有的基于振幅的方法面临两个关键局限：像素位置通常通过计算基索引或地址量子比特进行编码，导致量子资源随图像分辨率增长；同时，从归一化量子态中联合解码多个像素会引入像素之间的概率竞争，并限制精确的逐像素控制。为了解决这些问题，我们将量子图像生成重新表述为坐标条件隐式函数学习。我们的方法以空间坐标和潜在变量作为输入，使用经典嵌入网络生成依赖于输入的电路参数，并在每个坐标处评估变分量子电路。像素强度直接从一个专用颜色量子比特的期望值获得，完整图像通过查询所有空间坐标生成。这种设计将图像分辨率与地址量子比特需求解耦，并避免了跨像素共享概率归一化约束。我们进一步设计了一个专门的变分量子电路，为坐标条件生成提供结构性归纳偏置。在两个基准数据集上的仿真实验表明，我们的方法在使用更少量子比特的情况下，在视觉质量和定量质量上优于基于FRQI的生成方法和PQWGAN，并且比相应的经典基线获得了更好的生成质量。
cs.AI / 75 / 2608.11954
Causal inference for group-contaminated structured outcomes: observable quotients, lossless reduction and exact randomization inference
群污染结构化结果的因果推断：可观测商、无损约简与精确随机化推断
Abstract
Structured potential outcomes such as microscopy images may be recorded after an unknown, unit-specific transformation. If that transformation can depend on treatment, covariates or the intrinsic outcome, raw-coordinate analyses may mix biological effects with acquisition geometry. We study the unrestricted observation model X = Γ . Y(A) and characterize its observable information: a target is uniformly recoverable exactly when it is constant on group orbits, while a Borel maximal invariant retains every measurable invariant target. We then distinguish observability from statistical losslessness. A quotient-faithful reconstruction theorem shows that quotient reduction is sufficient for the full transformed experiment exactly when the conditional law of the raw observation given treatment, covariates and the quotient has a parameter-free version. Conditional Haar contamination on a compact group yields Blackwell equivalence as a special case; it is not imposed in the main model. We also separate independent site-specific product actions from shared diagonal actions and show why componentwise canonicalization can discard relative cross-site information. Under explicit metric and kernel regularity, an approximate-contamination theorem bounds quotient-law Wasserstein error and the induced perturbation of population maximum mean discrepancy. For finite-support multichannel lattice images, we construct a maximal invariant under integer translations and quarter turns, combine its characteristic Gaussian kernel with a complete paired-swap test, and retain the original simulations and RxRx1 HUVEC study. Under the sharp null, the quotient test rejected in 0.052 of simulation replicates; at unit effect strength its power was 0.992. The primary RxRx1 contrast had an enumerated paired-swap p-value of 0.0078.
Chinese Translation
诸如显微镜图像之类的结构化潜在结果可能在某种未知的、单元特定的变换之后被记录。如果该变换可能依赖于处理、协变量或内在结果，那么原始坐标分析可能会将生物学效应与采集几何混在一起。我们研究不受限制的观测模型 $X = Γ . Y(A)$，并刻画其可观测信息：一个目标当且仅当它在群轨道上恒定时是一致可恢复的，而一个 Borel 极大不变量保留了每个可测不变目标。随后，我们区分可观测性与统计无损失性。一个商忠实重构定理表明，商约简对完全变换后的实验是充分的，当且仅当原始观测在给定处理、协变量和商时的条件分布具有一个无参数版本。紧群上的条件 Haar 污染作为特例产生 Blackwell 等价；在主模型中并未施加这一条件。我们还区分独立的位点特定乘积作用与共享对角作用，并说明为什么逐分量规范化可能丢弃相对跨位点信息。在明确的度量与核正则性条件下，近似污染定理界定了商律 Wasserstein 误差以及所诱导的总体最大均值差异的扰动。对于有限支撑多通道格点图像，我们构造了一个在整数平移和四分之一旋转下的极大不变量，将其特征高斯核与完全成对交换检验相结合，并保留了原始的模拟实验和 RxRx1 HUVEC 研究。在尖锐零假设下，商检验在模拟重复中拒绝的比例为 0.052；在单位效应强度下其功效为 0.992。主要 RxRx1 对比的枚举成对交换 p 值为 0.0078。
机器学习 (cs.LG)
110
cs.LG / 1 / 2608.11585
Unifying Physical Backpropagation
统一物理反向传播
Abstract
Physical computing systems exploit device dynamics for computation, but their gradient-based optimization is challenging: backpropagation through a digital twin suffers from model-reality gap. On-device gradient computation could resolve this issue, and a handful of theoretical and experimental studies have proposed ways to achieve it. Yet a unifying theory identifying when a physical system can compute the gradient of its own performance has been missing. Here we develop such a unification, based on the adjoint method: we identify sufficient conditions under which the adjoint field required for formally exact gradients can be generated on the same hardware that performs the computation. Linear and nonlinear systems obey fundamentally different conditions: for linear systems damping or gain is admissible provided reciprocity is preserved. For nonlinear trajectory systems the sufficient conditions are reciprocity of the linearized system and the existence of a time-reversal mirror. Algorithmically, the nonlinear case requires infinitesimal nudging, whereas linear systems admit a finite-amplitude experiment. We recover Equilibrium Propagation, Hamiltonian echo backpropagation, fully forward mode training and in situ gradient methods in integrated-photonic and free-space-optical systems. We further show that reciprocity is only the simplest instance of a more general intertwining condition, which extends exact on-device gradient computation to a class of non-Hermitian, non-reciprocal systems. Further generalizations include time-dependent parameters, Onsager-reciprocal dynamics and nonlinear, PT-symmetric Schrödinger equations. Our work provides a unified theoretical basis for formally exact physical learning algorithms and a template for constructing them across a range of physical systems.
Chinese Translation
物理计算系统利用器件动力学进行计算，但其基于梯度的优化具有挑战性：通过数字孪生进行反向传播受到模型-现实差距的困扰。片上梯度计算可以解决这一问题，并且已有少量理论和实验研究提出了实现它的方法。然而，至今仍缺乏一种统一理论来判定物理系统何时能够计算自身性能的梯度。在此，我们基于伴随方法建立了这样一种统一理论：我们确定了充分条件，在这些条件下，形式精确梯度所需的伴随场可以在执行计算的同一硬件上生成。线性系统与非线性系统遵循根本不同的条件：对于线性系统，只要保持互易性，阻尼或增益都是允许的。对于非线性轨迹系统，充分条件是线性化系统的互易性以及时间反演镜的存在。在算法上，非线性情形需要无穷小微扰，而线性系统则允许有限幅度实验。我们恢复了平衡传播、哈密顿回波反向传播、全前向模式训练以及集成光子与自由空间光学系统中的原位梯度方法。我们进一步表明，互易性只是更一般的交织条件的最简单实例，该条件将精确片上梯度计算推广到一类非厄米、非互易系统。进一步的推广包括含时参数、昂萨格互易动力学以及非线性PT对称薛定谔方程。我们的工作为形式精确的物理学习算法提供了统一的理论基础，并为在一系列物理系统中构建这些算法提供了模板。
cs.LG / 2 / 2608.11472
Gaussian Meta-Space Augmentation for Stacking Ensembles in Multimodal IPMN Risk Stratification
用于多模态 IPMN 风险分层的堆叠集成的高斯元空间增强
Abstract
Pancreatic cancer is among the most lethal malignancies; risk stratification of intraductal papillary mucinous neoplasms (IPMNs) offers a crucial opportunity for early intervention but typically requires invasive tissue biopsy. Dominant vision-based approaches, including radiomics and deep learning, provide promising but initially separate discrimination opportunities. Similarly, multisequence MRI (T1W/T2W) and anatomically decomposed (head, body and tail) analysis of the pancreas provide additional and potentially complementary signals. Effective fusion of this information is crucial in ordinal IPMN dysplasia risk prediction and can be accomplished via a meticulously regularized and calibrated ensemble stacking combiner. We present cUPMI, a class-conditional Gaussian augmentation of a combiner's log-probability meta-features, and test it on various prediction paradigms. In our multi-center analysis, we find cUPMI adds limited value to properly regularized L2-logistic binary classification stacks, but consistently regularizes higher-capacity tree combiners in the binary and radiomics-only setting (RF +0.015 and XGBoost +0.024 binary AUC, positive in all seeds). Its cleanest ordinal benefit appears for XGBoost on an 8-stream radiomics task (3-class no < low < high, +0.022 QWK in all seeds). Separately, fold-locked fusion of radiomics and 2.5D CNN streams yields the strongest overall model, an RF stack reaching QWK 0.595 (95% CI [0.54, 0.64]) and binary AUC 0.839, surpassing radiomics, 2.5D ResNet, and 3D DenseNet-121 baselines.
Chinese Translation
胰腺癌是最致命的恶性肿瘤之一；导管内乳头状黏液性肿瘤（IPMNs）的风险分层为早期干预提供了关键机会，但通常需要进行有创组织活检。主流的基于视觉的方法，包括影像组学和深度学习，提供了有前景但最初相互独立的判别机会。类似地，多序列 MRI（T1W/T2W）以及对胰腺按解剖部位（头、体和尾）的分解分析提供了额外且可能互补的信号。对这些信息进行有效融合对于有序 IPMN 异型增生风险预测至关重要，并且可以通过一个经过精细正则化和校准的集成堆叠组合器来实现。我们提出了 cUPMI，一种对组合器对数概率元特征进行类别条件高斯增强的方法，并在多种预测范式上对其进行了测试。在我们的多中心分析中，我们发现 cUPMI 对经过适当正则化的 L2-逻辑二分类堆叠的增益有限，但在二分类和仅影像组学设置中，它能够持续地对容量更高的树组合器进行正则化（RF 二分类 AUC +0.015，XGBoost +0.024，在所有种子中均为正增益）。其最清晰的有序增益出现在 8 流影像组学任务中的 XGBoost 上（3 分类：无 < 低 < 高，所有种子中 QWK +0.022）。另外，影像组学和 2.5D CNN 流的折叠锁定融合得到了最强的整体模型，一个 RF 堆叠达到了 QWK 0.595（95% CI [0.54, 0.64]）和二分类 AUC 0.839，超过了影像组学、2.5D ResNet 和 3D DenseNet-121 基线。
cs.LG / 3 / 2608.11498
Language-Structured Relational Q-Learning for Threat-Aware Control in Safety-Critical Driving
面向安全关键驾驶中威胁感知控制的语言结构化关系Q学习
Abstract
Natural-language-based scenario generation offers an intuitive means of describing rare and complex driving interactions, yet it is still uncertain whether training with language-structured data leads to truly adaptive control policies. We propose Language-Structured Relational Q-Learning, instantiated through an Ego-Centric Relational Q-Network (ERQ-Net), which jointly learns inter-vehicle relevance and action values from dynamic traffic graphs. Language descriptions define surrounding-vehicle behaviours during training, while prompts and semantic actor roles are hidden from the policy. ERQ-Net must therefore infer threat relevance solely from observable kinematics and interactions. Across 2,500 safety-critical scenarios, language-structured training improves test success from 49-52% to 55-58% and increases adversary-focused attention from 1.2x to 2.1x, demonstrating emergent threat awareness. However, this representational gain does not consistently translate into adaptive control: trained policies perform similarly to the best constant action, while a portfolio of simple policies solves 76% of scenarios. We formalise this discrepancy as a recognition-control gap and show that reward reweighting and margin shaping do not eliminate the resulting policy collapse. Evaluations of realism, criticality, semantic accuracy, and transfer of state-interface representations to CARLA further highlight both the strengths and the constraints of language-structured relational policy learning in safety-critical driving scenarios.
Chinese Translation
基于自然语言的场景生成为描述罕见且复杂的驾驶交互提供了一种直观方式，然而，使用语言结构化数据进行训练是否能够产生真正自适应的控制策略仍不确定。我们提出了语言结构化关系Q学习，通过以自车为中心的关系Q网络（ERQ-Net）进行实例化，该网络从动态交通图中联合学习车辆间的相关性和动作价值。语言描述在训练期间定义周围车辆的行为，而提示和语义参与者角色对策略不可见。因此，ERQ-Net 必须仅从可观测的运动学信息和交互中推断威胁相关性。在 2,500 个安全关键场景中，语言结构化训练将测试成功率从 49–52% 提高到 55–58%，并将以对手为中心的注意力从 1.2 倍增加到 2.1 倍，显示出涌现的威胁感知能力。然而，这种表征上的增益并未一致地转化为自适应控制：训练得到的策略表现与最佳恒定动作相似，而一组简单策略组合可以解决 76% 的场景。我们将这种不一致形式化为识别与控制之间的差距，并表明奖励重新加权和间隔塑造并不能消除由此产生的策略坍缩。对真实性、关键性、语义准确性以及状态接口表示迁移到 CARLA 的评估，进一步凸显了语言结构化关系策略学习在安全关键驾驶场景中的优势与局限。
cs.LG / 4 / 2608.11634
CAM-Guided Saliency Cutout and Image-Based Malware Classification
CAM 引导的显著性裁剪与基于图像的恶意软件分类
Abstract
Dropout regularization is commonly used to reduce overfitting by removing parts of a neural network during training. For Convolutional Neural Networks (CNN), cutouts serve a somewhat analogous purpose. Cutouts can be implemented as data augmentation: the original training image is retained, and additional copies are created with regions removed. In this chapter, we test whether cutout placement can be improved by using High-Resolution Class Activation Mapping (HiResCAM). We compare four controlled training conditions: no cutout, standard random cutout, low-saliency cutout, and high-saliency cutout. We experiment using grayscale malware images from the RawMal-TF dataset (17 families with~1,000 samples per family), and for comparison to natural images, we experiment with the well-known CIFAR-100 dataset. All experiments are based on ResNet18 with~100 training epochs. For the cutout experiments, we test cutout areas of~5\%, 10\%, 20\%, and~30\%, and we consider~$M\in\{4,8}$ augmented copies per original training image. The RawMal-TF results are slightly worse for all three cutout cases (random, high and low saliency) as compared to no cutouts. In contrast, our CIFAR-100 experimental results improve slightly under low-saliency cutout. These results suggest that the value of saliency-guided cutout is domain dependent, and that malware images should not be treated as equivalent to natural images.
Chinese Translation
Dropout 正则化通常用于通过在训练期间移除神经网络的若干部分来减少过拟合。对于卷积神经网络（CNN），裁剪（cutout）起着某种类似的作用。裁剪可以作为数据增强来实现：保留原始训练图像，并创建移除区域后的额外副本。在本章中，我们测试通过使用高分辨率类别激活映射（HiResCAM）是否能够改进裁剪位置的选择。我们比较四种受控训练条件：无裁剪、标准随机裁剪、低显著性裁剪和高显著性裁剪。我们使用 RawMal-TF 数据集中的灰度恶意软件图像（17 个家族，每个家族约 1,000 个样本）进行实验，并且为了与自然图像进行比较，我们在著名的 CIFAR-100 数据集上进行实验。所有实验均基于 ResNet18 并训练约 100 个 epoch。对于裁剪实验，我们测试约 5%、10%、20% 和 30% 的裁剪面积，并考虑每张原始训练图像 $M\in\{4,8\}$ 个增强副本。与无裁剪相比，RawMal-TF 的结果在所有三种裁剪情形（随机、高显著性和低显著性）下都略差。相比之下，我们的 CIFAR-100 实验结果在低显著性裁剪下略有改善。这些结果表明，显著性引导的裁剪的价值依赖于具体领域，并且恶意软件图像不应被视为等同于自然图像。
cs.LG / 5 / 2608.11759
Automated binary classification of hazelnut X-ray images: A deep-learning benchmark for quality assessment
榛子X射线图像的自动二分类：面向质量评估的深度学习基准
Abstract
Non-destructive X-ray imaging can reveal internal hazelnut defects that are difficult to detect by external inspection alone; however, automated interpretation remains challenging because of subtle radiographic differences among classes, marked class imbalance, and limited annotated data. Here, we present a benchmark for binary hazelnut quality classification (healthy versus defective) based on 799 segmented single-kernel X-ray images (224 x 224 pixels, grayscale), grouped into 101 acquisition units. Seven single-model configurations and ten probability-aggregation ensembles were evaluated using a group-wise split-rotation protocol across five data splits generated using different random seeds. Decision thresholds were selected on the validation set, and performance was assessed deterministically on validation and test sets. Under the expert-reassessed annotation condition, the average-probability ensemble of the binary cross-entropy-trained convolutional neural network and frozen Swin Transformer achieved the highest mean balanced accuracy (86.3% +/- 1.8%, five seeds), with several other ensembles providing comparable performance. Across methods, substantial split-to-split variability was observed, indicating that multi-split evaluation is essential for reliable model comparison at this dataset scale. Expert reassessment of ambiguous samples improved the performance of all 17 evaluated methods by 2.8-8.1 percentage points, while having only a limited effect on cross-split variance. The results highlight both the potential of deep learning for automated X-ray-based hazelnut quality assessment and the importance of rigorous evaluation and label curation in small, imbalanced agricultural imaging datasets.
Chinese Translation
无损X射线成像能够揭示仅靠外部检查难以发现的榛子内部缺陷；然而，由于类别之间射线影像差异细微、类别不平衡明显以及标注数据有限，自动判读仍然具有挑战性。在此，我们基于799幅分割后的单粒榛子X射线图像（224×224像素，灰度）构建了一个榛子质量二分类（健康与缺陷）基准，这些图像被分为101个采集单元。使用分组划分-轮换协议，在由不同随机种子生成的五种数据划分上，对七种单模型配置和十种概率聚合集成方法进行了评估。决策阈值在验证集上选取，性能在验证集和测试集上以确定性方式评估。在专家复核标注条件下，由二分类交叉熵训练的卷积神经网络与冻结的Swin Transformer组成的平均概率集成取得了最高的平均平衡准确率（86.3% ± 1.8%，五个种子），其他多个集成方法也提供了相当的性能。在所有方法中，均观察到显著的划分间变异性，这表明在该数据集规模下，多划分评估对于可靠的模型比较至关重要。对模糊样本的专家复核将所有17种被评估方法的性能提高了2.8–8.1个百分点，同时对跨划分方差只有有限的影响。这些结果既凸显了深度学习在基于X射线的自动化榛子质量评估中的潜力，也凸显了在小规模、类别不平衡的农业图像数据集中进行严格评估和标签整理的重要性。
cs.LG / 6 / 2608.11770
Achieving Near-Zero-Overhead Multi-Model Hierarchical Classification in Real-Time Detection Pipelines
在实时检测流水线中实现近零开销的多模型层次分类
Abstract
Edge-deployed vision systems in target recognition, surveillance, autonomous vehicles, and drone domains require hierarchical inference pipelines where a detection model identifies objects of interest and downstream classifiers provide fine-grained attribute analysis. Running all models on the GPU creates a serial bottleneck that limits real-time throughput as pipeline stages grow. Modern edge SoCs pair GPUs with dedicated neural accelerators (NPUs, DLAs) capable of concurrent execution, yet deploying custom models on these accelerators remains impractical due to strict operator constraints, quantization incompatibilities, and an undocumented end-to-end pipeline. We target NVIDIA Jetson DLA cores as the representative platform. We present a five-step methodology for zero GPU fallback DLA INT8 deployment of classification backbones, comprising architecture adaptation, manual dynamic range workaround to rescue TensorRT's implicit quantization (recovering 94.0% accuracy from implicit quantization's 75%) for rapid pipeline validation before explicit quantization, quantization-aware training, ONNX graph surgery for DLA compilation, and a concurrent GPU-detection/DLA-classification inference pipeline. We document nine engineering constraints with root-cause analysis and generalizable solutions. Validation on a dual-head person attribute classifier running on DLA alongside a GPU object detector on a Jetson Orin NX demonstrates near-zero pipeline overhead (12.5 vs. 13.3~FPS detector-only at 1080p), with dual-DLA scaling at no additional cost. The methodology is backbone-agnostic and generalizes to any detection-classification edge pipeline.
Chinese Translation
在目标识别、监控、自动驾驶车辆和无人机领域，边缘部署的视觉系统需要层次化推理流水线，其中检测模型识别感兴趣目标，下游分类器提供细粒度属性分析。在 GPU 上运行所有模型会形成串行瓶颈，随着流水线阶段增多而限制实时吞吐量。现代边缘 SoC 将 GPU 与能够并发执行的专用神经加速器（NPU、DLA）配对，然而由于严格的算子约束、量化不兼容以及缺乏文档化的端到端流水线，在这些加速器上部署自定义模型仍然不切实际。我们以 NVIDIA Jetson DLA 核心作为代表性平台。我们提出一种用于分类骨干网络的无 GPU 回退 DLA INT8 部署的五步方法，包括架构适配、手动动态范围变通方案（用于挽救 TensorRT 的隐式量化，将隐式量化的 75% 准确率恢复至 94.0%，以便在显式量化前快速验证流水线）、量化感知训练、面向 DLA 编译的 ONNX 图手术，以及并发的 GPU 检测/DLA 分类推理流水线。我们记录了九个工程约束，并给出根因分析和可推广的解决方案。在 Jetson Orin NX 上，与 GPU 目标检测器并行运行在 DLA 上的双头行人属性分类器验证表明，流水线开销接近零（1080p 下，仅检测器为 13.3 FPS，而流水线为 12.5 FPS），并且双 DLA 扩展无需额外成本。该方法与骨干网络无关，并可推广到任何检测-分类边缘流水线。
cs.LG / 7 / 2608.11810
Can Vision Models Read the Radar Display? On the Feasibility of Radar Imagery for Air Traffic Complexity Estimation
视觉模型能读懂雷达显示吗？关于雷达图像用于空中交通复杂度估计的可行性
Abstract
Air traffic controllers perceive traffic complexity through the radar display, suggesting that a computer vision model operating on the same imagery may provide a natural architecture for modeling controller-perceived complexity; however, whether radar imagery is a viable input format for deep learning vision models remains unclear. Unlike natural images, radar images are extremely sparse and self-similar, consisting primarily of a black background and a few visually identical aircraft blobs, while small changes in aircraft positions can substantially alter sector-level complexity. To test whether a vision model can capture these operationally important differences, we encode each traffic situation as a position image supplemented by five channels representing aircraft state variables, including heading, speed, and altitude, and train a Vision Transformer (ViT) to regress four intrinsic complexity components derived from pairwise geometric relations among aircraft. The model achieves $R^2 > 0.96$ for all four components, and a one-aircraft-removal perturbation study shows that its response changes proportionally to how much the removed aircraft contributed to sector complexity rather than treating every removal as equivalent. These results demonstrate that, despite its atypical visual characteristics, radar imagery is a viable input format for air traffic complexity modeling.
Chinese Translation
空中交通管制员通过雷达显示感知交通复杂度，这表明基于相同图像运行的计算机视觉模型可能为建模管制员感知的复杂度提供一种自然的架构；然而，雷达图像作为深度学习视觉模型的输入格式是否可行仍不清楚。与自然图像不同，雷达图像极其稀疏且自相似，主要由黑色背景和少量视觉上相同的航空器斑点组成，而航空器位置的微小变化就可能显著改变扇区级复杂度。为了检验视觉模型能否捕捉这些运行上重要的差异，我们将每个交通态势编码为一个位置图像，并辅以五个表示航空器状态变量（包括航向、速度和高度）的通道，并训练一个视觉Transformer（ViT）来回归四个源自航空器之间成对几何关系的内在复杂度分量。该模型在所有四个分量上均达到 $R^2 > 0.96$，且一项移除单架航空器的扰动研究表明，其响应变化与所移除航空器对扇区复杂度的贡献成比例，而不是把每次移除都视为等效。这些结果表明，尽管雷达图像具有非典型的视觉特征，但其仍是空中交通复杂度建模的一种可行输入格式。
cs.LG / 8 / 2608.11996
A Remote Approach to Cashew Orchard Detection: Leveraging Active Learning with Satellite Imagery in Guinea-Bissau
一种用于腰果果园检测的遥感方法：在几内亚比绍利用卫星影像与主动学习
Abstract
Cashew production is a widespread economic activity in Guinea-Bissau, as well as other countries in West Africa. However, unregulated cashew production can be directly associated with increasing regionwide deforestation rates, biodiversity losses, and a fragile economic structure. There is no nationwide database for listing or georeferencing cashew orchards, so there is a clear need to remotely map their locations. In recent years, multiple methods for detecting orchards have been developed, though they have only been applied on a regional level. This work expands regional analyses to a nationwide scale. It develops a scalable and cost-effective remote approach, based on Sentinel-2 satellite imagery, using Machine Learning techniques to detect cashew orchards automatically. Margin-based Active Learning techniques were employed to develop an optimal training set in terms of the number of points and their informativeness, leading to a cashew map with 94.0% balanced accuracy obtained entirely off-site. We created two datasets and a 2021 cashew map with 10m spatial resolution that are openly accessible through GitHub. The results demonstrate the possibility of a broader cashew orchard mapping, creating a new stepping stone for this environmental application.
Chinese Translation
腰果生产是几内亚比绍以及西非其他国家广泛开展的经济活动。然而，不受监管的腰果生产可能与整个区域森林砍伐率的上升、生物多样性丧失以及脆弱的经济结构直接相关。目前没有用于列出腰果果园或为其添加地理参考的全国性数据库，因此显然需要通过遥感方式绘制其位置图。近年来，已经开发了多种果园检测方法，但这些方法仅在区域层面上得到应用。本研究将区域分析扩展到全国范围。本研究基于 Sentinel-2 卫星影像，开发了一种可扩展且具有成本效益的遥感方法，利用机器学习技术自动检测腰果果园。采用基于间隔的主动学习技术，从点数量及其信息量角度构建最优训练集，从而获得了一张平衡准确率为 94.0% 的腰果分布图，该图完全在非实地条件下获得。我们创建了两个数据集和一幅 2021 年空间分辨率为 10 米的腰果分布图，这些数据可通过 GitHub 公开获取。结果表明，更大范围的腰果果园制图是可行的，为这一环境应用奠定了新的基础。
cs.LG / 9 / 2608.12064
Draw This First
先画这个
Abstract
We invert the typical formulation of sketch generation: instead of drawing strokes in order, we predict a 2D field that defines the order in which strokes are drawn. We use a pretrained latent flow-matching transformer to supply the image prior to predict an intermediate representation, while training the VAE's decoder to predict the order field, stroke mask, and stroke segmentation. We vectorize the predicted segmentation into polylines and sort them by the field, producing an ordered vector sketch. Our model can predict an ordered vector sketch from a text description or derender an image into ordered vectors; for either, it follows text instructions specifying the order of drawing.
Chinese Translation
我们反转了草图生成的典型形式化方式：不是按顺序绘制笔画，而是预测一个二维场，该场定义了笔画被绘制的顺序。我们使用预训练的潜在流匹配Transformer来提供图像先验，以预测中间表示，同时训练VAE的解码器来预测顺序场、笔画掩码和笔画分割。我们将预测的分割向量化为折线，并根据该场对其进行排序，从而生成有序矢量草图。我们的模型可以根据文本描述预测有序矢量草图，或将图像反渲染为有序矢量；无论哪种情况，它都遵循指定绘制顺序的文本指令。
cs.LG / 10 / 2608.12086
Look What the Probes Dragged In! Real-World Chest X-ray Shortcuts in MedCLIP
看看探针拖来了什么！MedCLIP中的真实世界胸部X射线捷径
Abstract
Vision-language models, such as contrastive language-image pre-training (CLIP)-based approaches, have reached state-of-the-art (SOTA) results in medical artificial intelligence. However, recent work reveals that CLIP-based models remain vulnerable to shortcuts. We investigate how real-world shortcuts manifest across different layers of the medical CLIP-based model, MedCLIP, and its vision encoder, a frozen ResNet-50. We attach 17 linear classification probes to the intermediate layers of the ResNet-50 and train them on three different dataset configurations and targets: NIH-CXR14 (pneumothorax) and PadChest (cardiomegaly and pneumothorax). This setup allows us to observe model behaviour during evaluation using subgroup-based calibration and layer-wise confidence curves. We find that the final linear probes achieve a high AUROC but poor calibration in the models. The layer-wise confidence analyses suggest that shortcuts emerge at different depths. Patterns consistent with localised shortcuts, such as drains, appear at later layers, while patterns consistent with diffuse shortcuts, such as scanner-specific noise patterns, emerge earlier, aligning with previous work. Finally, we conduct a manual analysis of the images, which reveals data quality issues in both NIH-CXR14 and PadChest. Our findings underscore that even SOTA models remain vulnerable to shortcuts, and the need for high-quality and well-annotated datasets to draw solid conclusions. Code can be found on our GitHub: https://github.com/nikodice4/MedCLIP_shortcuts.
Chinese Translation
视觉-语言模型，例如基于对比语言-图像预训练（CLIP）的方法，已在医学人工智能中取得最先进（SOTA）的结果。然而，最近的研究表明，基于CLIP的模型仍然容易受到捷径的影响。我们研究了现实世界中的捷径如何在基于医学CLIP的模型MedCLIP及其视觉编码器（一个冻结的ResNet-50）的不同层中表现出来。我们将17个线性分类探针连接到ResNet-50的中间层，并在三种不同的数据集配置和目标上训练它们：NIH-CXR14（气胸）和PadChest（心脏肥大和气胸）。这一设置使我们能够使用基于亚组的校准和逐层置信度曲线，在评估期间观察模型行为。我们发现，最终线性探针取得了较高的AUROC，但在模型中的校准性较差。逐层置信度分析表明，捷径在不同深度出现。与局部捷径（如引流管）一致的模式出现在较后的层，而与弥散捷径（如扫描仪特异性噪声模式）一致的模式出现得更早，这与先前的工作一致。最后，我们对图像进行了人工分析，发现NIH-CXR14和PadChest中均存在数据质量问题。我们的发现强调，即使是最先进的模型仍然容易受到捷径的影响，并且需要高质量、标注良好的数据集才能得出可靠的结论。代码可在我们的GitHub上找到：https://github.com/nikodice4/MedCLIP_shortcuts。
cs.LG / 11 / 2608.12145
Autonomous Telerehabilitation via Skeletal Motion Prediction and Joint-Level Performance Assessment
基于骨骼运动预测与关节级表现评估的自主远程康复
Abstract
Autonomous rehabilitation systems must not only recognize human motion but also provide structured feedback to support users without continuous therapist supervision. This paper presents a telerehabilitation pipeline that integrates skeleton-based exercise quality assessment and short-term motion prediction into a two-module system operating on marker-free RGB video. A self-attentive Bidirectional LSTM performs exercise quality classification using MMD-NCA metric learning, while a graph-based motion prediction module computes per-joint position errors between predicted and observed poses, generating spatially localized deviation signals. Each module is evaluated independently on established benchmarks: the classifier achieves 96.45% mean-class accuracy on squat sequences from the PROZIS dataset, and the adopted STARS predictor achieves a mean MPJPE of 75.8 mm at 560 ms on Human3.6M, outperforming graph and recurrent baselines across all prediction horizons. The framework is designed for eventual deployment in assistive robotics and home-based rehabilitation contexts; end-to-end integration and clinical validation are important directions for future work. By combining motion recognition and prediction in a single system, this work contributes a step toward autonomous, feedback-driven telerehabilitation, for more accessible and scalable rehabilitation solutions.
Chinese Translation
自主康复系统不仅必须识别人体运动，还必须提供结构化反馈，以在无需治疗师持续监督的情况下支持用户。本文提出了一种远程康复流程，将基于骨骼的锻炼质量评估和短期运动预测集成到一个在无标记RGB视频上运行的双模块系统中。一个自注意力双向LSTM使用MMD-NCA度量学习执行锻炼质量分类，而一个基于图的运动预测模块计算预测姿态与观测姿态之间的逐关节位置误差，生成空间局部化的偏差信号。每个模块在公认基准上独立评估：分类器在PROZIS数据集的深蹲序列上实现了96.45%的平均类别准确率，所采用的STARS预测器在Human3.6M上560毫秒时实现了75.8毫米的平均MPJPE，在所有预测时长上均优于图基线和循环基线。该框架旨在最终部署于辅助机器人和家庭康复场景中；端到端集成和临床验证是未来工作的重要方向。通过将运动识别和预测结合在单一系统中，本工作为迈向自主、反馈驱动的远程康复迈出了一步，从而实现更易获取和可扩展的康复解决方案。
cs.LG / 12 / 2608.11318
Terminal Symmetry as a Decision Resource: Statewise Refinement for Anytime Verified Construction
终端对称性作为决策资源：面向随时可验证构造的状态级精化
Abstract
Many sequential construction tasks exhibit exact symmetry at completion while their execution remains directed and history-dependent. We develop a decision-resource view of terminal symmetry: process evidence supplies directionality, terminal correspondence transports that structure across equivalent outcomes, realized-state evidence refines its current decision relevance after transitions, and a fixed verifier certifies execution. This decomposition yields transport--refine--certify. \method{} instantiates the principle with an episode-fixed transported process structure, its state-restricted process rank, a state-dependent residual rank refreshed after accepted transitions, and an ordinal rank meet whose top-$k$ set is exactly the union of the two proposal prefixes. The meet provides a completion guarantee under prefix coverage and attains the tight worst-case verifier-query bound under the corresponding prefix information model; a two-state construction predicts a strict post-transition dynamic--static separation. Across CAD assembly, Mini-Programs, and exact-fill packing, statewise refresh improves anytime AUC by up to $6.77$, $21.75$, and $8.68$ points, respectively. On 1,135 target-removal episodes from the official GRN OOD scenes, \method{} attains the lowest mean capped verifier cost at all three scales among the compared GRN and CDGS-style planners. The statewise signal also transfers across aggregation and scheduler organizations. Terminal symmetry thereby becomes a reusable decision resource for directed construction.
Chinese Translation
许多顺序构造任务在完成时表现出精确对称性，而它们的执行过程仍具有方向性并依赖历史。我们提出了一种关于终端对称性的决策资源视角：过程证据提供方向性，终端对应将该结构在等价结果之间传输，已实现状态证据在转移后细化其当前决策相关性，固定验证器则对执行进行验证。这种分解产生传输--精化--验证。\method{} 通过回合固定的传输过程结构、其状态受限过程秩、在接受转移后刷新的状态依赖残差秩，以及一个序数秩交来实例化该原则；该序数秩交的 top-$k$ 集合恰好是两个提案前缀的并集。该交在前缀覆盖下提供完成保证，并在相应前缀信息模型下达到紧的最坏情况验证器查询界；一个两状态构造预测了转移后严格的动态--静态分离。在 CAD 装配、小程序和精确填充装箱中，状态级刷新将随时 AUC 分别提高至多 $6.77$、$21.75$ 和 $8.68$ 个点。在来自官方 GRN OOD 场景的 1,135 个目标移除回合上，\method{} 在三个尺度上均获得与所比较的 GRN 和 CDGS 风格规划器相比最低的平均截断验证器代价。状态级信号还能跨聚合和调度器组织结构进行迁移。终端对称性由此成为用于有向构造的可重用决策资源。
cs.LG / 13 / 2608.11324
Contextual Quality-Diversity Evolutionary Reinforcement Learning for HVAC Control in Tropical Commercial Buildings
用于热带商业建筑 HVAC 控制的情境质量多样性进化强化学习
Abstract
This paper proposes a contextual quality-diversity evolutionary reinforcement-learning controller, CQD-ERL, for the supervisory control of a tropical, water-cooled chiller plant and its associated air side. Rather than converging to a single scalarised policy, the controller maintains a product archive of specialised policies indexed jointly by a data- driven operating context, a cluster of daily weather and load regime, and a context-invariant behaviour descriptor, filled by a gradient-free evolutionary operator and a soft-actor-critic policy-gradient operator that share one replay buffer. Every action is filtered through a deterministic safety shield before execution. The controller is trained on a two-tier reduced-order environment representing the latent load, cooling-tower approach and humidity constraints of a Singapore commercial building, and is evaluated over a full annual backtest against an ASHRAE Guideline 36 baseline.
Chinese Translation
本文提出了一种情境质量多样性进化强化学习控制器 CQD-ERL，用于热带水冷式冷水机组及其相关空气侧的监督控制。该控制器不是收敛到单一标量化策略，而是维护一个由专用策略组成的乘积档案，这些策略由数据驱动的运行情境、每日天气和负荷状态的聚类以及情境不变行为描述符联合索引，并由共享同一回放缓冲区的无梯度进化算子和软演员-评论家策略梯度算子填充。每个动作在执行前都经过一个确定性安全护盾过滤。该控制器在一个两层降阶环境中训练，该环境表示一座新加坡商业建筑的潜热负荷、冷却塔逼近度和湿度约束，并在全年度回测中与 ASHRAE 指南 36 基线进行评估。
cs.LG / 14 / 2608.11342
Weightless Fine-Tuning: Personalizing LLMs via Logit-Space Transport
无权重微调：通过 Logit 空间传输实现大语言模型的个性化
Abstract
Supervised fine-tuning (SFT) is a standard approach for adapting LLMs to a target distribution, but in settings such as personalization, where each author requires separate weight access, optimization, storage, and retraining, its costs become prohibitive. We propose Weightless Fine-Tuning (WFT), a training-free decoding-time method that approximates the distributional effect of SFT without weight updates. WFT computes supervised residuals on an author's training sequence and transports them to the current prompt through a cross-prefix transport operator estimated from dropout-induced cross-covariance. The operator captures how a perturbation at one context propagates to predictions at another, replacing gradient-based parameter updates with logit-space corrections. On three LaMP personalization benchmarks, WFT achieves the best average performance across datasets, matches or exceeds SFT on individual tasks, and outperforms other lightweight baselines on average. In a budget-controlled comparison, WFT approaches SFT performance using less than 7% of the effective computation. Logit-level analysis shows a cosine similarity of 0.875 between the logit shifts induced by WFT and SFT over 95% of the next-token probability mass, suggesting that WFT captures the distributional effect of supervised adaptation without modifying model weights.
Chinese Translation
监督微调（SFT）是使大语言模型（LLMs）适应目标分布的标准方法，但在个性化等场景中，每位作者都需要单独的权重访问、优化、存储和重新训练，其成本变得难以承受。我们提出无权重微调（Weightless Fine-Tuning，WFT），一种无需训练、在解码时进行的方法，在不更新权重的情况下近似监督微调的分布效应。WFT 在作者的训练序列上计算有监督残差，并通过一个由 dropout 诱导的交叉协方差估计得到的跨前缀传输算子，将它们传输到当前提示。该算子刻画了一个上下文处的扰动如何传播到另一个上下文的预测，从而用 logit 空间校正取代基于梯度的参数更新。在三个 LaMP 个性化基准上，WFT 在跨数据集的平均性能上取得最佳结果，在单个任务上达到或超过 SFT，并在平均意义上优于其他轻量基线。在预算受控的比较中，WFT 以不到 7% 的有效计算量接近 SFT 的性能。Logit 层面的分析显示，在 95% 的下一个 token 概率质量上，WFT 与 SFT 所诱导的 logit 偏移之间的余弦相似度为 0.875，这表明 WFT 在不修改模型权重的情况下捕捉到了有监督适应的分布效应。
cs.LG / 15 / 2608.11349
Dynamics Models for Offline Hyperparameter Selection in Real-World RL
真实世界强化学习中离线超参数选择的动力学模型
Abstract
A key obstacle to deploying reinforcement learning in real-world systems is hyperparameter selection, particularly when simulators are unavailable and online experimentation is costly. Prior work has proposed calibration models trained on offline data to approximate environment dynamics and enable offline hyperparameter selection, but these methods have so far been evaluated only in simple simulated settings. In this paper, we present the first application of calibration models in a real-world industrial setting: a municipal water treatment plant. We evaluate several calibration model approaches, including a k-nearest neighbors model with a Laplacian distance metric, on high-dimensional, non-stationary sensor data for nexting prediction tasks. Our results show that these models can generate realistic long-horizon rollouts and recover meaningful hyperparameter sensitivity trends. We further examine how calibration models scale to year-long datasets, how they support the selection of fine-tuning learning rates for pre-trained agents, and how robust they are under distribution shift. Overall, our findings provide a proof of concept for using offline dynamics models to support RL deployment in real-world environments, while highlighting important practical challenges for future work.
Chinese Translation
在真实世界系统中部署强化学习的一个关键障碍是超参数选择，尤其是在模拟器不可用且在线实验成本高昂的情况下。先前的工作提出了基于离线数据训练的校准模型，用于近似环境动力学并支持离线超参数选择，但这些方法迄今仅在简单的仿真环境中得到评估。在本文中，我们展示了校准模型在真实世界工业场景中的首次应用：一座市政水处理厂。我们在高维、非平稳的传感器数据上评估了多种校准模型方法，包括带有拉普拉斯距离度量的 k 近邻模型，用于 nexting 预测任务。我们的结果表明，这些模型能够生成逼真的长时程 rollout，并恢复有意义的超参数敏感性趋势。我们进一步考察了校准模型如何扩展到长达一年的数据集、如何支持为预训练智能体选择微调学习率，以及它们在分布偏移下的鲁棒性。总体而言，我们的发现为使用离线动力学模型支持真实世界环境中的强化学习部署提供了概念验证，同时指出了未来工作面临的重要实践挑战。
cs.LG / 16 / 2608.11359
Market-Information-Aware Gated-LoRA of Foundation Models for Transferable Day-Ahead Electricity Price Forecasting
用于可迁移日前电价预测的基础模型市场信息感知门控 LoRA
Abstract
Electricity price forecasting is crucial for market participants but remains difficult because prices are volatile, market-specific, and closely tied to anticipated system conditions. Existing supervised methods depend largely on market-specific historical data, limiting their use in newly established or data-scarce markets. This paper proposes a market-information-aware adaptation framework that transfers the Chronos-2 time-series foundation model to day-ahead electricity price forecasting. It first constructs a multi-source market information (MSMI) interface aligning 7-day price context with pre-clearing supply--demand, reserve, maintenance, generator-capacity, and intertie variables, and then trains a source-domain gated low-rank adapter (LoRA), updating about $1\%$ of model parameters without target-market labels. The gate scales the frozen source adapter according to reserve-tightness and operating-state signals. A leave-one-market-out protocol is adopted for evaluating cross-market transferability. Experiments on four Chinese provincial day-ahead spot markets show that the proposed framework reduces the average MAE/RMSE by $6.24\%/7.99\%$ relative to market-information-aware zero-shot Chronos-2 and by $3.05\%/3.52\%$ relative to vanilla Source-LoRA. Experiments show that the gain is not reproduced by a learned global scalar or by random gate initialization, while the additional improvement over Source-LoRA is limited. These results suggest that market-structured inputs and state-dependent gated LoRA can provide a practical transfer path for data-scarce electricity markets.
Chinese Translation
电价预测对市场参与者至关重要，但由于价格波动大、具有市场特异性，并且与预期系统条件密切相关，因此仍然困难。现有监督方法在很大程度上依赖特定市场的历史数据，这限制了它们在新建市场或数据稀缺市场中的应用。本文提出一种市场信息感知适配框架，将 Chronos-2 时间序列基础模型迁移到日前电价预测。它首先构建一个多源市场信息（MSMI）接口，将 7 天价格上下文与预出清供需、备用、检修、发电容量和联络线变量对齐，然后训练一个源域门控低秩适配器（LoRA），在没有目标市场标签的情况下更新约 $1\%$ 的模型参数。门控根据备用紧缺程度和运行状态信号对冻结的源适配器进行缩放。采用留一市场协议评估跨市场可迁移性。在四个中国省级日前现货市场上的实验表明，相对于市场信息感知的零样本 Chronos-2，所提框架将平均 MAE/RMSE 降低了 $6.24\%/7.99\%$；相对于普通 Source-LoRA，则降低了 $3.05\%/3.52\%$。实验表明，这种增益无法通过学习到的全局标量或随机门控初始化来复现，而相对于 Source-LoRA 的额外改进有限。这些结果表明，市场结构化输入和依赖状态的 gated LoRA 可以为数据稀缺的电力市场提供实用的迁移路径。
cs.LG / 17 / 2608.11368
PAIR: Pairwise-Aware Inclusion Reweighting for Adaptive Rollout Allocation in RLVR
PAIR：用于RLVR中自适应展开分配的成对感知包含重加权
Abstract
Reinforcement learning with verifiable rewards (RLVR) spends most of its compute generating groups of long reasoning trajectories. Recent allocators reduce this cost by assigning budgets to prompts, rollouts, or tokens according to a pointwise notion of difficulty or utility. We identify a statistical mismatch: the unclipped leave-one-out group-relative score gradient is not a sum of independent point contributions, but a second-order U-statistic over pairs of rollouts. Completing one rollout therefore reveals contrast with every other completed rollout, and adaptive endpoint selection changes which pair terms are observable. We introduce PAIR (Pairwise-Aware Inclusion Reweighting), which treats short rollout prefixes as vertices and pair-gradient terms as edges of a contrast graph. A prefix-only predictor estimates correctness and remaining token cost; a convex design chooses positive continuation probabilities under an expected suffix-token budget; and each edge induced by completed vertices is inverse-weighted by its logged joint inclusion probability. Under conditionally independent on-policy rollouts and an unclipped, unstandardized objective, the resulting estimator is design-unbiased for the complete candidate-pair gradient. Across compute-matched RLVR runs on Qwen3-1.7B/4B, PAIR improves average accuracy by +1.2 and +1.4 over the strongest pointwise allocator while using 51% and 52% fewer generated tokens than full-group GRPO. A frozen-population estimator audit confirms that unweighted adaptive selection is biased, whereas pair-inclusion correction recovers the complete-pair target at matched suffix cost.
Chinese Translation
使用可验证奖励的强化学习（RLVR）将大部分计算用于生成一组组长推理轨迹。近期的分配器通过根据逐点的难度或效用概念为提示、展开或 token 分配预算来降低这一成本。我们发现一个统计上的不匹配：未裁剪的留一法群体相对得分梯度不是独立逐点贡献之和，而是关于成对展开的二阶 U 统计量。因此，完成一个展开会揭示它与每一个其他已完成展开之间的对比，而自适应终点选择会改变哪些成对项是可观测的。我们引入了 PAIR（成对感知包含重加权），它将短展开前缀视为顶点，将成对梯度项视为对比图的边。仅使用前缀的预测器估计正确性和剩余 token 成本；一个凸设计在期望后缀 token 预算下选择正的继续概率；并且由已完成顶点诱导的每条边都按其对数联合包含概率进行逆加权。在条件独立的同策略展开以及未裁剪、未标准化的目标下，所得估计量对于完整候选对梯度是设计无偏的。在 Qwen3-1.7B/4B 上计算量匹配的 RLVR 运行中，与最强的逐点分配器相比，PAIR 将平均准确率分别提高了 +1.2 和 +1.4，同时生成的 token 数量比全组 GRPO 分别少 51% 和 52%。一项冻结总体估计量审计证实，未加权的自适应选择是有偏的，而配对包含校正在匹配的后缀成本下恢复了完整配对目标。
cs.LG / 18 / 2608.11375
Towards an approach to multivariate outlier detection for District Heating System data
迈向一种针对区域供热系统数据的多元异常值检测方法
Abstract
In this paper, we test different methods for multivariate detection of outliers in the data of transmitted heat energy in the selected substation of local District Heating System, by also considering outside ambient temperature, namely Z-score (univariate, as a benchmark), Mahalanobis distances, Principal Component Analysis (PCA), Isolation Forest and Hotelling's T-squared test. The overall research aims at uncovering irregular plant operation, with a wider objective of identifying the opportunities for reducing the consumption of gas in central heating plants as well as the CO2 emission. The proposed approach considers specific domain circumstances, such as irrelevance of zero transmit-ted energy timepoints as indication of off-grid plant. The outcomes of the different methods are discussed with domain experts. It was concluded that PCA, Isolation Forest and Hotelling method provide relevant results. Finally, we adopt the ensemble method (selection based on the agreement of all three methods on the detected outliers) as the final approach.
Chinese Translation
在本文中，我们在本地区域供热系统选定子站的传输热能数据中，同时考虑室外环境温度，测试了不同的多元异常值检测方法，即Z分数（单变量，作为基准）、马氏距离、主成分分析（PCA）、孤立森林和霍特林T平方检验。整体研究旨在揭示供热站的不正常运行，更广泛的目标是识别减少集中供热厂燃气消耗以及CO2排放的机会。所提出的方法考虑了特定的领域情况，例如零传输能量时间点作为离网供热站指示的无关性。不同方法的结果与领域专家进行了讨论。得出的结论是，PCA、孤立森林和霍特林方法提供了相关的结果。最后，我们采用集成方法（基于所有三种方法对检测到的异常值达成一致的筛选）作为最终方法。
cs.LG / 19 / 2608.11383
Reoptimization Algorithms for Contextual Bandits with Knapsack Constraints
带背包约束的上下文赌博机的重优化算法
Abstract
We study new algorithms for Contextual Bandits with Knapsack. In these problems, there are finitely many types of customers, products, and resources. Each product is made from a fixed combination of resources, and resources have finite capacity. A decision maker must assign each arriving customer one out of a set of multiple possible products. Every assignment of a customer to a product will generate a random reward, which equals an unknown linear function of customer and product features, plus a noise term. The objective is to jointly learn the mean reward function, and to make online assignments to minimize the expected revenue loss relative to an optimal policy that knows the reward function. We propose a natural and simple extension of the Upper-Confidence-Bound (UCB) family of algorithms and apply re-optimization techniques. We show that by taking advantage of re-optimization, our algorithm achieves an average regret of $O(\frac{(\ln T)^3}{T})$ where $T$ is the horizon length. Our bound significantly reduces the $O(\frac{1}{\sqrt{T}})$ bound in the literature for closely related dynamic-pricing problems that are based on re-optimization.
Chinese Translation
我们研究带背包约束的上下文赌博机的新算法。在这些问题中，顾客、产品和资源的类型都是有限的。每种产品由固定的资源组合制成，且资源具有有限容量。决策者必须为每位到达的顾客从一组多个可能产品中分配一种。每次将顾客分配给一种产品都会产生一个随机奖励，该奖励等于顾客和产品特征的未知线性函数加上一个噪声项。目标是同时学习平均奖励函数，并进行在线分配，以最小化相对于知道该奖励函数的最优策略的期望收益损失。我们提出了上置信界（UCB）算法族的一种自然且简单的扩展，并应用了重优化技术。我们证明，通过利用重优化，我们的算法实现了 $O(\frac{(\ln T)^3}{T})$ 的平均遗憾，其中 $T$ 是时间范围长度。我们的界显著降低了文献中基于重优化的密切相关动态定价问题的 $O(\frac{1}{\sqrt{T}})$ 界。
cs.LG / 20 / 2608.11390
Mechanism Design for Generative Engines: From Exploitation toward Win-Win Outcomes
生成式引擎的机制设计：从利用走向双赢结果
Abstract
Generative engines are reshaping the web ecosystem by making citations a key mechanism for allocating attention, attribution, and downstream value. This creates a strategic tension: content providers are incentivized to optimize for model citation, while platforms must preserve answer quality and trustworthy attribution. We show that this tension can escalate into citation wars. In repeated simulations, state-of-the-art generative engine optimization (GEO) attacks adapt to conventional defenses by producing citation-seeking rewrites that degrade document quality and introduce unsupported claims. To study this problem, we formulate the supplier--platform interaction as a repeated Stackelberg game with partial monitoring. A local best-response analysis identifies when citation competition approaches an inert stationary outcome. Motivated by this finding, we propose a platform--creator mechanism called VCR based on verifiable-content rewards. Rather than only penalizing suspicious rewrites, the platform also credits rewrites that surface checkable factual substance, aligning creator incentives with answer trustworthiness. Experiments on three benchmarks show that VCR consistently achieves the largest Net defense-utility score, outperforming the strongest baseline by an average of 12.1 percentage points, and produces a win--win outcome under our empirical equivalence criterion.
Chinese Translation
生成式引擎正在重塑网络生态系统，使引用成为分配关注度、署名归属和下游价值的关键机制。这造成了一种策略性张力：内容提供者有激励针对模型引用进行优化，而平台必须保持答案质量和可信的署名归属。我们表明，这种张力可能升级为“引用战争”。在重复模拟中，最先进的生成引擎优化（GEO）攻击会适应传统防御，生成以获取引用为目的的改写，这些改写降低文档质量并引入无据主张。为了研究这一问题，我们将供应方—平台互动建模为具有部分监测的重复斯塔克尔伯格博弈。局部最优响应分析确定了引用竞争何时趋近于一种惰性稳态结果。受这一发现启发，我们提出了一种基于可验证内容奖励的平台—创作者机制，称为 VCR。该平台不仅惩罚可疑改写，还对能够呈现可核查事实内容的改写给予奖励，从而使创作者激励与答案可信度保持一致。在三个基准上的实验表明，VCR 持续取得最高的净防御效用分数，比最强基线平均高出 12.1 个百分点，并在我们的经验等价标准下产生双赢结果。
cs.LG / 21 / 2608.11410
Unmasking Toxic Mimicry in Medical Offline Reinforcement Learning for ICU Sepsis Management via Counterfactual Clinical Audits
通过反事实临床审计揭示用于ICU脓毒症管理的医疗离线强化学习中的有毒模仿
Abstract
Offline reinforcement learning (RL) offers considerable promise for optimizing ICU treatment decisions, yet standard evaluation metrics Mean Squared Error (MSE) and Fitted Q-Evaluation (FQE) assess only behavioral imitation and cannot detect Toxic Mimicry, a failure mode in which agents replicate harmful patterns such as treatment withdrawal during comfort-care transitions. Using the MIMIC-III database, we propose the Counterfactual Clinical Audit (CCA) framework, which stress-tests RL agents through physiological perturbations anchored in Surviving Sepsis Campaign (SSC) guidelines. We audit a Medical Decision Transformer (MedDT) and a Historical Causal Transformer (HCT-RL), the latter employing Causal Action Shielding, propensity-based importance weighting, and Conservative Q-Learning. CCA reveals that MedDT paradoxically reduces vasopressor dosage as lactate escalates, contradicting resuscitation guidelines, while HCT-RL maintains physiologically consistent responses. These findings expose a systemic misalignment between statistical fit and clinical safety, supporting counterfactual audits as a necessary evaluation standard for medical RL.
Chinese Translation
离线强化学习（RL）在优化ICU治疗决策方面具有相当大的前景，然而标准评估指标均方误差（MSE）和拟合Q评估（FQE）仅评估行为模仿，无法检测有毒模仿——这是一种失败模式，即智能体复制有害模式，例如在舒适护理过渡期间撤回治疗。使用MIMIC-III数据库，我们提出了反事实临床审计（CCA）框架，该框架通过基于拯救脓毒症运动（SSC）指南的生理扰动对强化学习智能体进行压力测试。我们审计了一个医疗决策Transformer（MedDT）和一个历史因果Transformer（HCT-RL），后者采用了因果动作屏蔽、基于倾向性的重要性加权和保守Q学习。CCA揭示，MedDT在乳酸升高时反而减少血管升压药剂量，这与复苏指南相矛盾，而HCT-RL保持生理上一致的反应。这些发现暴露了统计拟合与临床安全之间的系统性错位，支持将反事实审计作为医疗RL的必要评估标准。
cs.LG / 22 / 2608.11423
Analysis of Federated Aggregation under Model Poisoning and Backdoor Attacks: A Reconstructed Cross-Dataset and Cross-Architecture Benchmark
模型投毒与后门攻击下联邦聚合的分析：一个重建的跨数据集与跨架构基准
Abstract
Robust comparisons of federated aggregation methods require joint consideration of predictive performance, threat definitions, metric semantics, and execution provenance. A 500-cell seed-1 evaluation matrix was reconstructed across five aggregation methods, five datasets, five architectures, and four recorded conditions: clean, sign-flipping, Gaussian, and BadNets. Successful execution logs were identified for 454 original runs and 36 repaired or rerun executions, whereas 10 clean SVHN cells were supported by summary-only provenance. Trimmed Mean achieved the highest clean macro-mean accuracy (76.02%) and the lowest mean within-task rank (1.70). Krum attained the highest recorded accuracy under both sign-flipping and Gaussian configurations. These relative rankings remained unchanged when analysis was restricted to 21 task pairs for which original successful logs were available for every method-condition combination. Audit of the supplied BadNets metric implementation established that every test input is triggered prior to target-label counting; consequently, the retained metric represents Triggered Target-Label Rate (TTLR) rather than a conventional target-excluding attack success rate. An audit of the supplied FedPARETO scaffold further identified a pathway in which predictive summaries may characterize an uncorrupted local model while the aggregation weight is applied to a separately corrupted update, introducing a potential discrepancy between reported predictive outcomes and the updates used for aggregation. The canonical matrix contains a single identified seed for each cell, and exact attack and configuration lineage is incomplete. Accordingly, the findings should be interpreted as descriptive comparisons within the recorded configurations and not as statistical estimates or universal claims regarding robustness.
Chinese Translation
对联邦聚合方法的稳健比较需要同时考虑预测性能、威胁定义、度量语义和执行来源。重建了一个500单元格的种子1评估矩阵，涵盖五种聚合方法、五个数据集、五种架构和四种记录条件：干净、符号翻转、高斯和BadNets。成功执行日志被识别为454次原始运行和36次修复或重新运行的执行，而10个干净的SVHN单元格仅由摘要来源支持。修剪均值实现了最高的干净宏观平均准确率（76.02%）和最低的平均任务内排名（1.70）。Krum在符号翻转和高斯配置下均取得了最高的记录准确率。当分析限制在21个任务对上时——对于这些任务对，每种方法-条件组合都有原始成功日志——这些相对排名保持不变。对所提供的BadNets度量实现的审计确认，每个测试输入在目标标签计数之前都被触发；因此，保留的度量表示触发目标标签率（TTLR），而不是常规的排除目标标签的攻击成功率。对所提供的FedPARETO脚手架的审计进一步发现了一条路径，其中预测摘要可能描述一个未损坏的局部模型，而聚合权重却被应用于一个单独损坏的更新，从而在报告的预测结果与用于聚合的更新之间引入潜在的不一致。规范矩阵中每个单元格仅包含一个已识别的种子，并且确切的攻击和配置谱系不完整。因此，研究结果应解释为在记录配置内的描述性比较，而不是关于鲁棒性的统计估计或普遍性声明。
cs.LG / 23 / 2608.11424
Click2Poly: A VLM for vector mapping buildings and walls
Click2Poly：一种用于建筑物和墙体矢量制图的 VLM
Abstract
Accurate vector mapping of buildings and walls is critical for geospatial applications but remains a labor-intensive process. While recent deep learning methods have improved automatic extraction, in order to meet cartographic standards they always require a human to perform quality control and fix complex cases in the extraction. We present Click2Poly, a human-in-the-loop AI assistant designed to speed up this manual step. Extending the Florence-2 Vision Language Model (VLM), Click2Poly responds to user clicks by editing the building or wall vector layer directly. Implemented as a QGIS plugin, Click2Poly speeds up the manual editing of building and wall vector layers in a real-world production environment.
Chinese Translation
建筑物和墙体的精确矢量制图对地理空间应用至关重要，但仍然是劳动密集型过程。尽管最近的深度学习方法改进了自动提取，但为了满足制图标准，它们始终需要人工进行质量控制并修复提取中的复杂情况。我们提出了 Click2Poly，一种人在环路的 AI 助手，旨在加快这一手动步骤。通过扩展 Florence-2 视觉语言模型（VLM），Click2Poly 通过直接编辑建筑物或墙体矢量图层来响应用户点击。Click2Poly 以 QGIS 插件形式实现，在实际生产环境中加快了建筑物和墙体矢量图层的人工编辑速度。
cs.LG / 24 / 2608.11427
Three Tokens Force Exponential Feature Rank in Nonnegative Kernel Attention
三个 Token 迫使非负核注意力具有指数级特征秩
Abstract
Full attention exposes every token pair, whereas kernel attention compresses a sequence into a fixed-dimensional sketch. We show that this distinction becomes exponential at the first context length containing two competing candidates. On Min-IP over Boolean inputs, rank-one normalized kernel attention solves every sequence of length at most two exactly. In contrast, any single normalized nonnegative kernel-attention head that succeeds on all three-token sequences with error strictly below $1/2$ requires $2^{Ω(m)}$ features, even with arbitrary finite-dimensional tokenwise values and an arbitrary query-dependent affine readout. Dense softmax solves the same task with $m$-dimensional scores and constant temperature. The conclusion survives position-dependent token maps and a causal final query. As context length grows, the lower bound approaches the exact $2^m$-feature realization. Separately, for deterministic multihead, multilayer sketch models whose cross-token channels have finite alphabets, we prove a transcript lower bound linear in the number of independent answers and logarithmic in their alphabet size.
Chinese Translation
全注意力会暴露每一对 token，而核注意力会将序列压缩为一个固定维度的草图。我们证明，这种区别在首个包含两个竞争候选者的上下文长度上就会变成指数级。在布尔输入上的 Min-IP 任务中，秩一归一化核注意力可以精确解决所有长度至多为二的序列。相比之下，任何单个归一化非负核注意力头，如果在所有三 token 序列上以严格低于 $1/2$ 的误差成功，都需要 $2^{Ω(m)}$ 个特征，即使允许任意有限维的逐 token 值和任意依赖于查询的仿射读出。稠密 softmax 以 $m$ 维分数和恒定温度即可解决同样的任务。该结论在位置依赖的 token 映射和因果最终查询下仍然成立。随着上下文长度增长，该下界会趋近于精确的 $2^m$ 个特征的实现。另外，对于跨 token 通道具有有限字母表的确定性多头、多层草图模型，我们证明了一个 transcript 下界，该下界与独立答案的数量成线性关系，并与字母表大小成对数关系。
cs.LG / 25 / 2608.11431
AutoGrable: What Is a Good Graph for a Table?
AutoGrable：什么是一张表的好图？
Abstract
Graph learning presupposes a graph, and tables and relational databases do not come with one. Applying a GNN to them requires deciding which entities become nodes, which of them to connect, and through which relations---a decision made by hand, by schema heuristics, or by training a model on every candidate graph and keeping the best. We give a criterion that requires no trained graph model. In the minimal table-to-graph abstraction each row is a node, so a message-passing GNN, bounded by 1-WL, sees a construction only as a partition of the rows into colour-refinement classes: a construction is good for a task when that partition separates rows with different labels and does not split rows that share one. AutoGrable turns this criterion into a construction procedure. For incidence constructions the partition is fixed by the selected columns, so building a graph reduces to choosing them, and we score a candidate subset by a label-alignment risk: the held-out risk of the best predictor constant on its blocks, penalised by an occupancy term measuring how thinly the blocks are populated. The score materialises no graph and trains no GNN, so AutoGrable can search the space of subsets greedily and cheaply, and returns the resulting grable for single tables and for foreign-key schemas alike. Our experiments show that over a space of candidate graphs the score discards a large fraction while retaining the best; that AutoGrable recovers the columns that generate the label on controlled tasks and outperforms fixed, random, and task-aware constructors on real tasks under a fixed predictor; and that it is the only method compared that can decline to build a graph when none helps.
Chinese Translation
图学习以图为先决条件，而表格和关系数据库并不自带图。将 GNN 应用于它们需要决定哪些实体成为节点、连接哪些实体以及通过哪些关系——这一决策可由人工、模式启发式方法做出，也可通过在每一个候选图上训练模型并保留最佳者来做出。我们给出了一个不需要任何已训练图模型的准则。在最小表到图抽象中，每一行是一个节点，因此受 1-WL 约束的消息传递 GNN 只会将一种构造视为把行划分成颜色细化类的分区：当该分区将具有不同标签的行分开并且不拆分共享同一标签的行时，该构造对某个任务而言就是好的。AutoGrable 将这一准则转化为一种构造过程。对于关联构造，分区由所选列确定，因此构建图就简化为选择这些列，并且我们通过标签对齐风险为候选子集评分：该风险是在其块上取常值的最佳预测器的留出风险，并加上一个衡量块的填充稀疏程度的占用项作为惩罚。该评分不实际构造任何图，也不训练任何 GNN，因此 AutoGrable 可以贪婪地且低成本地搜索子集空间，并且对单表和外键模式同样返回所得到的 grable。我们的实验表明，在候选图空间中，该评分会丢弃其中很大一部分并保留最佳者；AutoGrable 在受控任务上恢复了生成标签的列，并在固定预测器下于真实任务上优于固定、随机和任务感知构造器；并且它是所比较的方法中唯一能够在没有图有帮助时拒绝构建图的方法。
cs.LG / 26 / 2608.11435
Variational Parameter Calibration with Physics-Aware Latent-Space Surrogates
基于物理感知潜空间代理模型的变分参数校准
Abstract
Forward and inverse modeling of parametric dynamical systems requires surrogate models that are not only accurate for state prediction, but also informative for parameter calibration. However, a systematic end-to-end differentiable formulation for coupling deep-learning-based reduced-order surrogates with variational parameter estimation remains underdeveloped. In this work, we introduce a physics-aware neural-network-based latent-space framework for reduced-order forward modeling and variational parameter estimation. The proposed autoencoder-based approach yields a differentiable surrogate that maps physical parameters to predicted flow fields through a latent representation. The observable supervision is used during offline training to encourage the latent variables to retain information correlated with system parameters, while the online inverse problem is solved in the parameter space through the surrogate-induced observation operator. The method is evaluated on two computational-fluid-dynamics benchmarks. The results show that reconstruction accuracy alone is insufficient for inverse modeling, owing to the lack of end-to-end differentiability or physics awareness for variational parameter calibration. Quantitative latent-space analysis further shows that observable supervision improves case-level separability and temporal organization of latent representations. Experiments with realistic measurement settings, including noisy, low-resolution, randomly masked, and block-wise partial observations, demonstrate the robustness of the proposed framework and show that it generally reduces calibration error and variability compared with the standard surrogate models.
Chinese Translation
参数化动力系统的正演与反演建模需要代理模型不仅在状态预测上准确，而且对参数校准具有信息价值。然而，将基于深度学习的降阶代理模型与变分参数估计耦合的系统性端到端可微分公式仍然发展不足。在本文中，我们提出了一种基于物理感知神经网络的潜空间框架，用于降阶正演建模和变分参数估计。所提出的基于自编码器的方法产生了一个可微分代理模型，该模型通过潜表示将物理参数映射到预测流场。在离线训练期间使用可观测监督，以促使潜变量保留与系统参数相关的信息，而在线反问题则通过代理诱导的观测算子在参数空间中求解。该方法在两个计算流体力学基准上进行了评估。结果表明，仅凭重建精度不足以支持反演建模，因为缺乏用于变分参数校准的端到端可微分性或物理感知。定量潜空间分析进一步表明，可观测监督改善了潜表示的案例级可分性和时间组织性。在包含噪声、低分辨率、随机掩码和分块部分观测的真实测量设置下的实验证明了所提出框架的鲁棒性，并表明与标准代理模型相比，它通常降低了校准误差和变异性。
cs.LG / 27 / 2608.11446
XGBoost "is all you need": the case of forecasting transmitted heat energy in District Heating Systems
XGBoost“即为你所需要的一切”：以区域供热系统中的传输热能预测为例
Abstract
This paper presents a comparative study of two distinct approaches, XGBoost and Long-Short Term Memory (LSTM), for forecasting transmitted heat energy in District Heating Systems (DHS). The objective is to explore scenarios in which conventional ML algorithms demonstrate better performance over deep learning networks in time series forecasting and the associated benefits in terms of computational cost and environmental impact. The study focuses on a real-world DHS dataset. Through experimentation and analysis, it is demonstrated that XGBoost consistently outperforms LSTM in this specific forecasting task. The difference is explained by the error distribution illustrating that LSTM makes more significant errors in the intervals of less data availability. The reduced computational demands of conventional ML approaches not only result in cost savings but also minimize the carbon footprint associated with data analysis tasks in energy systems.
Chinese Translation
本文针对区域供热系统（DHS）中传输热能的预测，对两种不同方法——XGBoost和长短期记忆（LSTM）——进行了比较研究。其目标是探索在时间序列预测中传统机器学习算法相较于深度学习网络表现更优的场景，以及由此带来的在计算成本和环境影响方面的相关益处。本研究聚焦于一个真实的区域供热系统数据集。通过实验和分析，结果表明，在这一特定的预测任务中，XGBoost始终优于LSTM。这一差异可由误差分布加以解释：误差分布表明，在数据可用性较低的区间内，LSTM会产生更显著的误差。传统机器学习方法计算需求的降低不仅可节省成本，还可最大限度减少能源系统中与数据分析任务相关的碳足迹。
cs.LG / 28 / 2608.11465
PAC-Bayes Beyond Parameter Space: Behavioral Equivalence, Z-Information, and Exact Complexity Decomposition
超越参数空间的 PAC-Bayes：行为等价、Z-信息与精确复杂度分解
Abstract
PAC-Bayes theory provides generalization guarantees by controlling the Kullback--Leibler (KL) divergence between posterior and prior distributions over a chosen hypothesis representation. However, predictive risk depends only on the predictive behavior induced by a hypothesis, not on the particular internal realization that implements that behavior. In over-parameterized systems, many distinct configurations induce identical predictive behavior, yet the classical PAC-Bayes KL divergence does not distinguish uncertainty over predictive behavior from variation among behaviorally equivalent realizations. We show that this distinction induces an exact structural decomposition of classical PAC-Bayes complexity. We formalize behavioral equivalence through a measurable behavior map and use measure disintegration to decompose probability measures on the configuration space into a distribution over predictive behaviors and conditional distributions over behavioral fibers. This yields an exact decomposition of the classical PAC-Bayes KL divergence into a behavior-selection term and a realization-level term given by an expected conditional KL within fibers. We define Z-information as the negative of this realization-level contribution: the exact gap between the KL divergence and the complexity of uncertainty over predictive behavior alone. We further show that the behavior-selection term admits an exact variational characterization: it is the minimum KL divergence among all posteriors inducing the same distribution over predictive behaviors, attained by a canonical fiber-symmetrized representative. Finally, we show that symmetry, behavior-preserving directions, fiber geometry, and invariance under fiber-preserving perturbations arise naturally from the same behavior-map structure. Together, these results identify predictive behavior as the natural object of PAC-Bayes complexity.
Chinese Translation
PAC-Bayes 理论通过控制所选假设表示上后验分布与先验分布之间的 Kullback--Leibler（KL）散度，来提供泛化保证。然而，预测风险仅取决于假设所诱导的预测行为，而不取决于实现该行为的特定内部实现。在过参数化系统中，许多不同的配置会诱导出相同的预测行为，但经典的 PAC-Bayes KL 散度并未将预测行为上的不确定性与行为等价实现之间的变异区分开来。我们证明，这一区分诱导出经典 PAC-Bayes 复杂度的一种精确结构分解。我们通过一个可测行为映射来形式化行为等价，并利用测度分解将配置空间上的概率测度分解为预测行为上的分布和行为纤维上的条件分布。由此，经典 PAC-Bayes KL 散度被精确分解为一个行为选择项和一个实现层项，后者由纤维内期望条件 KL 给出。我们将 Z-信息定义为这一实现层贡献的负值：即 KL 散度与仅关于预测行为的不确定性的复杂度之间的精确差距。我们进一步证明，行为选择项具有精确的变分刻画：它是在预测行为上诱导出相同分布的所有后验中的最小 KL 散度，并由一个典范的纤维对称化代表达到。最后，我们证明，对称性、保持行为的方向、纤维几何以及对保持纤维的扰动的不变性都自然地源于同一个行为映射结构。总之，这些结果表明预测行为是 PAC-Bayes 复杂度的自然对象。
cs.LG / 29 / 2608.11473
Dual-Primal Graph VAEs for Noisy Label Aggregation
用于噪声标签聚合的对偶-原始图VAE
Abstract
Inferring the ground-truth from noisy crowdsourced labels is an important theoretical and practical problem. Neural network-based methods offer an alternative to classical Bayesian models which require specifying a family of generative models used for inference. However, current models either still rely on fairly simple generative models for inference or require pseudo-labels or synthetic data to train the aggregate classifier. We propose a graph VAE architecture in which the decoder and encoder use GAT-based message passing on the adjacency graph of a crowdsourced dataset and its dual, respectively. The ground-truth labels are treated as latent variables, enabling unsupervised representation learning without needing to train a separate classifier. We show our model achieves state of the art performance on crowdsourcing benchmarks. We then demonstrate the generality of our approach by showing how the original crowdsourcing graph can be augmented to incorporate side information such as representations from neural network classifiers trained on the noisy labels to substantially boost their classification performance at test time.
Chinese Translation
从含噪众包标签中推断真实标签是一个重要的理论和实践问题。基于神经网络的方法为经典贝叶斯模型提供了一种替代方案，后者需要指定一族用于推断的生成模型。然而，当前模型要么在推断时仍依赖相当简单的生成模型，要么需要伪标签或合成数据来训练聚合分类器。我们提出了一种图VAE架构，其中解码器和编码器分别在众包数据集的邻接图及其对偶图上使用基于GAT的消息传递。真实标签被视作潜变量，从而无需训练单独的分类器即可进行无监督表示学习。我们表明，我们的模型在众包基准上取得了最先进的性能。随后，我们通过展示如何对原始众包图进行增强以纳入辅助信息（例如在噪声标签上训练的神经网络分类器所产生的表示），从而在测试时显著提升其分类性能，证明了我们方法的通用性。
cs.LG / 30 / 2608.11479
Convergence Guarantees of Gradient Descent for Neural Networks via Generalized Lipschitz Smoothness
通过广义 Lipschitz 光滑性的神经网络梯度下降收敛性保证
Abstract
We establish convergence guarantees of gradient descent for general feedforward neural networks of arbitrary width or depth, with no special requirements on the initialization or dataset. We only assume that the activation functions are Lipschitz smooth, Lipschitz continuous, and linearly bounded--- properties that hold for linear, tanh, softplus, and sigmoid activation functions. For the loss function, we require that it is Lipschitz smooth in the model outputs, which is true for mean-squared error. The key theoretical insight is that the Lipschitz properties of the activation functions are partially preserved even through repeated compositions, leading to a novel generalized Lipschitz smoothness condition where the change in gradient is upper bounded by the change in the parameter space, multiplied by polynomial terms of the parameter norms at both endpoints. This type of condition holds for both the model function and the loss function, enabling a descent lemma where the loss decreases as long as the learning rate is small enough with respect to the parameter norms. By ensuring that the parameter norms do not grow too quickly to infinity, we prove that the minimum squared gradient norm converges to zero in $T$ iterations at rate $O(1/T^{1/L})$ for an $L$-layer neural network.
Chinese Translation
我们为任意宽度或深度的一般前馈神经网络建立了梯度下降的收敛性保证，且对初始化或数据集没有特殊要求。我们仅假设激活函数是 Lipschitz 光滑的、Lipschitz 连续的且线性有界的——这些性质对线性、tanh、softplus 和 sigmoid 激活函数均成立。对于损失函数，我们要求它在模型输出上是 Lipschitz 光滑的，这对均方误差成立。关键的理论洞见在于，即使经过重复复合，激活函数的 Lipschitz 性质也被部分保留，从而产生一个新的广义 Lipschitz 光滑性条件：其中梯度变化的上界由参数空间中的变化乘以两个端点处参数范数的多项式项给出。这类条件对模型函数和损失函数都成立，从而能够建立一个下降引理：只要学习率相对于参数范数足够小，损失就会下降。通过确保参数范数不会过快地增长到无穷大，我们证明，对于一个 $L$ 层神经网络，最小平方梯度范数在 $T$ 次迭代中以 $O(1/T^{1/L})$ 的速率收敛到零。
cs.LG / 31 / 2608.11495
Defending against Model Extraction for GNNs with Model Reprogramming
利用模型重编程防御图神经网络模型提取
Abstract
Graph Neural Networks (GNNs) serve as the backbone for high-stakes applications in Machine-Learning-as-a-Service (MLaaS). Still, their black-box deployment exposes them to Model Extraction (ME) attacks, in which adversaries steal intellectual property by querying APIs. Existing defenses suffer from a critical ''Euclidean bias'': they transfer image-based strategies (e.g., random noise) to graphs, ignoring the complex topological dependencies between nodes, which often results in severe utility degradation. Passive methods like watermarking also fail to prevent theft in real time. To bridge this gap, we propose GraphRP (Graph Reprogramming Protection), a proactive defense framework that repurposes Model Reprogramming for security. Unlike static perturbations, GraphRP introduces a Structure-Aware Gating Mechanism driven by learnable topological prototypes. This creates a dynamic ''structural firewall'' that selectively modulates the model's decision boundary: it preserves fidelity for benign queries residing on the training manifold, while maximizing the Fisher Information along the perturbation direction for adversarial queries. Under standard assumptions (bounded loss, optimal attacker, and local second-order approximation), we prove a lower bound on the attacker's estimation error that increases with the structural sensitivity of the reprogramming noise. Extensive experiments on both hard-label and soft-label ME attacks demonstrate that GraphRP significantly degrades attack effectiveness while preserving benign utility.
Chinese Translation
图神经网络（GNN）是机器学习即服务（MLaaS）中高风险应用的核心支撑。然而，其黑盒部署使其面临模型提取（ME）攻击，攻击者通过查询API窃取知识产权。现有防御存在严重的“欧几里得偏差”：它们将基于图像的策略（如随机噪声）迁移到图上，忽略了节点之间复杂的拓扑依赖关系，这往往导致严重的效用下降。被动方法（如水印）也无法实时阻止窃取。为弥补这一不足，我们提出了GraphRP（图重编程防护），一个将模型重编程重新用于安全目的的主动防御框架。与静态扰动不同，GraphRP引入了一种由可学习拓扑原型驱动的结构感知门控机制。由此形成一个动态的“结构防火墙”，选择性地调节模型的决策边界：对位于训练流形上的良性查询保持保真度，同时对对抗性查询沿扰动方向最大化Fisher信息。在标准假设（有界损失、最优攻击者和局部二阶近似）下，我们证明了攻击者估计误差的一个下界，该下界随重编程噪声的结构敏感性增大而增大。在硬标签和软标签ME攻击上的大量实验表明，GraphRP在保持良性效用的同时显著降低了攻击有效性。
cs.LG / 32 / 2608.11499
HyperFix: Combinatorial Nonlinear Correction for Task Vector Merging
HyperFix：用于任务向量合并的组合非线性校正
Abstract
Task vectors enable model merging without joint retraining. In practice, the subset of task vectors to be merged may vary, but many existing methods use scalar tuning for a particular subset, requiring repeated tuning across subsets and restricting task vector merging to linear rescaling. We therefore formulate merging across varying task subsets as a combinatorial correction problem and introduce HyperFix, a lightweight hypernetwork that predicts subset-conditioned nonlinear corrections in weight space. Trained once on singleton, pair, and triple subsets from a task bank, HyperFix generalizes to larger subsets without per-subset optimization. Our local perturbation analysis bounds the residual correction beyond linear merging and motivates learning it from small task updates. Experiments across diverse benchmarks show that HyperFix outperforms existing task vector merging methods while reducing tuning cost.
Chinese Translation
任务向量使得模型合并无需联合重训练。在实践中，要合并的任务向量子集可能变化，但许多现有方法针对特定子集使用标量调优，这需要在不同子集间重复调优，并将任务向量合并限制为线性缩放。因此，我们将跨不同任务子集的合并形式化为一个组合校正问题，并引入 HyperFix，这是一个轻量级超网络，可在权重空间中预测以子集为条件的非线性校正。HyperFix 在任务库中的单任务、双任务和三任务子集上训练一次后，无需对每个子集进行优化即可泛化到更大的子集。我们的局部扰动分析界定了线性合并之外的残差校正，并激励从小任务更新中学习该校正。跨多个基准的实验表明，HyperFix 优于现有的任务向量合并方法，同时降低了调优成本。
cs.LG / 33 / 2608.11508
RelShap: Relationally Consistent Shapley Explanations
RelShap：关系一致性的 Shapley 解释
Abstract
Machine learning pipelines commonly flatten relational data into single-table representations, discarding structural constraints. Widely used Shapley value-based feature attributions then rely on feature independence, evaluating the model on combinations that could never arise in the underlying data, producing misleading explanations. We propose RelShap, a framework that incorporates relational constraints and data provenance into Shapley value computation, restricting both background data and coalition evaluation to relationally valid configurations. The framework is estimator-agnostic and composes with Kernel SHAP, Monte Carlo, and Leverage SHAP without altering their sampling or weighting properties. Functional dependencies further induce equivalence classes over feature coalitions, which RelShap exploits to reduce runtime without changing Shapley values; we provide a combinatorial characterization of the expected speedup. Experiments across multiple datasets, models, and estimators show that RelShap produces explanations that are more faithful to the data-generating process, correctly identifying the dominant feature in controlled settings where existing methods, including Conditional SHAP and ManifoldShap, do not. Our code is available at: https://github.com/duneag2/relshap.
Chinese Translation
机器学习流水线通常将关系型数据扁平化为单表表示，从而丢弃结构约束。随后广泛使用的基于 Shapley 值的特征归因依赖于特征独立性，在底层数据中根本不可能出现的组合上评估模型，从而产生误导性解释。我们提出 RelShap，这是一个将关系约束和数据溯源纳入 Shapley 值计算的框架，将背景数据和联盟评估都限制在关系上有效的配置中。该框架与估计器无关，并且可以与 Kernel SHAP、Monte Carlo 和 Leverage SHAP 组合，而不会改变它们的采样或加权性质。函数依赖进一步在特征联盟上诱导出等价类，RelShap 利用这一点在不改变 Shapley 值的情况下减少运行时间；我们给出了预期加速的组合刻画。跨多个数据集、模型和估计器的实验表明，RelShap 产生的解释更忠实于数据生成过程，在受控设置中正确识别出主导特征，而包括 Conditional SHAP 和 ManifoldShap 在内的现有方法则无法做到。我们的代码可在 https://github.com/duneag2/relshap 获取。
cs.LG / 34 / 2608.11511
Let it Cook: Learning to Wait in Sequential Decision Making
让它煮：在序贯决策中学习等待
Abstract
In sequential decision making, an agent typically observes its environment and acts at every timestep. However, such active participation may not always be necessary; tasks such as brewing coffee include periods that are served equally well by letting the environment evolve without constant monitoring and control. During such periods, the agent could simply wait to conserve its resources, or redirect its attention to another task. We capitalize on these opportunities by training a "waiting policy" that decides where and how long to wait. This involves forgoing sensing to commit to a wait action, representing a deliberate pause for a set number of timesteps. We formalize "learning to wait" as minimizing the frequency of sensing and decision making without sacrificing task performance (e.g., the total amount of time to complete a task). To train a waiting policy, we propose an approach that employs reinforcement learning with lexicographically ordered objectives. In experiments across 4 discrete-state household tasks and 3 continuous-state environments, we show that our approach successfully learns waiting behaviors, and can adapt pre-trained policies to wait where appropriate. While different tasks permit different amounts of waiting without sacrificing task performance, our approach consistently finds solutions with significant waiting, sometimes waiting for over 50 percent of the task duration.
Chinese Translation
在序贯决策中，智能体通常会在每个时间步观察其环境并执行动作。然而，这种主动参与可能并不总是必要的；诸如煮咖啡等任务包含这样一些时间段：让环境在没有持续监测和控制的情况下自行演化也能同样好地完成。在这些时间段内，智能体可以简单地等待以节省其资源，或将其注意力转向另一项任务。我们通过训练一个“等待策略”来利用这些机会，该策略决定在哪里等待以及等待多长时间。这涉及放弃感知以承诺执行等待动作，表示在设定数量的时间步内有意暂停。我们将“学习等待”形式化为在不牺牲任务性能（例如，完成一项任务所需的总时间）的前提下最小化感知和决策的频率。为了训练等待策略，我们提出一种使用按词典序排序目标的强化学习方法。在4个离散状态家务任务和3个连续状态环境上的实验中，我们表明我们的方法成功地学会了等待行为，并且能够调整预训练策略，使其在适当之处进行等待。虽然不同任务在不牺牲任务性能的情况下允许不同的等待量，但我们的方法始终能找到具有显著等待量的解决方案，有时等待时间超过任务持续时间的50%。
cs.LG / 35 / 2608.11519
FLARE++: Low-rank attention with dynamic attention routing
FLARE++：具有动态注意力路由的低秩注意力
Abstract
Full self-attention is a strong token mixer for PDE surrogates on irregular domains, but its quadratic cost limits its use on high-resolution problems. Efficient latent-attention models such as the Fast Low-rank Attention Routing Engine (FLARE) avoid that cost by routing all N tokens through M << N learned latent queries, but those queries are parameters: once trained, the same learned query templates serve every input. We remove this restriction with FLARE++, a low-rank attention architecture with dynamic token routing. FLARE++ reuses FLARE's own encoder to build its routing queries: learned latent seeds drive one extra encode call that gathers the N input tokens into M input-conditioned queries, and those queries then determine how the same tokens are compressed and redistributed. This preserves FLARE's explicit low-rank factorization and linear O(NM) complexity, and expresses the complete routing operation with standard scaled dot-product attention (SDPA) calls alone. We also provide a multi-GPU context-parallel implementation that shards input tokens across devices without ever gathering the full token sequence on one of them. FLARE++ is competitive across a set of standard PDE surrogate benchmarks, improving on fixed-query FLARE by 24% on average, and it gains 2.3 points of average accuracy on Long Range Arena.
Chinese Translation
完全自注意力对于不规则域上的PDE代理模型是一种强大的token混合器，但其二次代价限制了其在高分辨率问题上的使用。高效的潜在注意力模型，例如快速低秩注意力路由引擎（FLARE），通过将所有N个token经由M << N个学习到的潜在查询进行路由来避免这一代价，但这些查询是参数：一旦训练完成，相同的已学习查询模板会服务于每个输入。我们通过FLARE++（一种具有动态token路由的低秩注意力架构）消除了这一限制。FLARE++复用了FLARE自身的编码器来构建其路由查询：学习到的潜在种子驱动一次额外的编码调用，将N个输入token汇聚为M个以输入为条件的查询，然后这些查询决定相同的token如何被压缩和重新分配。这保留了FLARE的显式低秩分解和线性O(NM)复杂度，并且仅用标准的缩放点积注意力（SDPA）调用即可表达完整的路由操作。我们还提供了一种多GPU上下文并行实现，该实现将输入token分片到多个设备上，而无需在其中任何一个设备上汇聚完整的token序列。FLARE++在一组标准PDE代理模型基准上具有竞争力，相较于固定查询的FLARE平均提升了24%，并且在Long Range Arena上平均准确率提高了2.3个百分点。
cs.LG / 36 / 2608.11532
Hierarchical Federated Transfer Learning in Digital Twin-Based Vehicular Networks
基于数字孪生的车载网络中的分层联邦迁移学习
Abstract
In recent research on the Digital Twin-based Vehicular Ad hoc Network(DT-VANET), Federated Learning (FL) has shown its ability to provide data privacy. However, Federated learning struggles to adequately train a global model when confronted with data heterogeneity and data sparsity among vehicles, which ensure suboptimal accuracy in making precise predictions for different vehicle types. To address these challenges, this paper combines Federated Transfer Learning (FTL) to conduct vehicle clustering related to types of vehicles and proposes a novel Hierarchical Federated Transfer Learning (HFTL). We construct a framework for DT-VANET, along with two algorithms designed for cloud server model updates and intra-cluster federated transfer learning, to improve the accuracy of the global model. In addition, we developed a data quality score-based mechanism to prevent the global model from being affected by malicious vehicles. Lastly, detailed experiments on real-world datasets are conducted, considering different performance metrics that verify the effectiveness and efficiency of our algorithm.
Chinese Translation
近年来，在基于数字孪生的车载自组织网络（DT-VANET）研究中，联邦学习（FL）已展现出提供数据隐私的能力。然而，当面对车辆之间的数据异构性和数据稀疏性时，联邦学习难以充分训练全局模型，这导致在针对不同车辆类型进行精确预测时只能获得次优准确率。为了解决这些挑战，本文结合联邦迁移学习（FTL）来进行与车辆类型相关的车辆聚类，并提出了一种新颖的分层联邦迁移学习（HFTL）。我们构建了一个面向DT-VANET的框架，并设计了两种分别用于云服务器模型更新和簇内联邦迁移学习的算法，以提高全局模型的准确率。此外，我们开发了一种基于数据质量评分的机制，以防止全局模型受到恶意车辆的影响。最后，我们在真实世界数据集上进行了详细实验，考虑了不同的性能指标，验证了所提算法的有效性和效率。
cs.LG / 37 / 2608.11541
Robust Ambiguity Detection (RAD) From Model- and Feature-Space Consistency
基于模型空间与特征空间一致性的鲁棒模糊性检测（RAD）
Abstract
Machine learning models should be robust, in the sense of remaining predictively consistent under permissible variations. A model's predictions should ideally remain unchanged when it is replaced by a functionally equivalent one, or when its inputs are subject to minor, admissible perturbations. If such changes alter a prediction significantly, then the prediction is "ambiguous" with respect to the model. Models should abstain from making such ambiguous predictions and/or should flag them for human inspection, especially in high-stakes decision-making scenarios. However, in practice, such ambiguity is not easy to identify once a model is deployed. Here, the Robust Ambiguity Detection (RAD) framework is advanced for quantifying predictive ambiguity using two complementary metrics: Model-Space Consistency and Feature-Space Consistency. These two scores, the RAD Score-Pair, visualised through the RAD Plot, provide an interpretable characterisation of the sources of ambiguity and the actions a user may consider in response. RAD is evaluated on synthetic datasets with systematically controlled overlap, as well as several real-world datasets where the level of ambiguity cannot be directly inspected. Finally, we demonstrate a downstream application of RAD where samples are ranked by their RAD Pareto-Rank and the most ambiguous are abstained from prediction, achieving performance comparable to existing rejection-based approaches.
Chinese Translation
机器学习模型应当具有鲁棒性，即在可容许的变化下保持预测一致性。理想情况下，当一个模型被功能等价的模型替换，或当其输入受到微小且可容许的扰动时，模型的预测应保持不变。如果此类变化显著改变某个预测，那么该预测相对于该模型就是“模糊的”。模型应当避免做出此类模糊预测，并且/或者应将其标记出来供人工检查，尤其是在高风险决策场景中。然而，在实践中，模型一旦部署，这种模糊性并不容易被识别。本文提出了鲁棒模糊性检测（RAD）框架，利用两个互补指标来量化预测模糊性：模型空间一致性和特征空间一致性。这两个分数，即 RAD 分数对，通过 RAD 图可视化，为模糊性来源以及用户可考虑采取的应对行动提供了可解释的刻画。RAD 在系统性控制重叠程度的合成数据集以及若干模糊性水平无法直接检查的真实世界数据集上进行了评估。最后，我们展示了 RAD 的一个下游应用：样本根据其 RAD Pareto 秩排序，并对最模糊的样本放弃预测，取得了与现有基于拒绝的方法相当的性能。
cs.LG / 38 / 2608.11555
Certifying What Helps Customer-Return Timing: A Screen-and-Confirm Test for Conditioning Signals, and Why Decay Is Nearly Enough
证明什么有助于客户返回时间：针对条件信号的筛选-确认检验，以及为何衰减几乎足够
Abstract
Practitioners enrich customer-return models with ever more signals (lifetime value, category, recency/frequency, calendar, geography), and the temporal-point-process (TPP) literature follows suit with covariate- and external-covariate-conditioned intensities. But does any of it improve the timing, and how would you know? A null ("feature X doesn't help") is only meaningful if the model could have found a signal. We make two contributions--a method and a measurement--to answer this credibly. (i) A screen-and-confirm protocol that certifies whether a candidate signal improves a TPP's event-timing likelihood: a positive control plants a coupling of known strength and confirms the model recovers it, so a real-data null can be read as "no signal" rather than "weak method." The control is validated for categorical and continuous encodings, and on a real clock-driven dataset (NYC taxi hour-of-day). (ii) A model-free ceiling quantifying how little of customer-return timing is point-predictable at all (a single-digit percentage of gap variance from any covariate; returns are near-memoryless). With these we certify a clean result on three public benchmarks (Amazon, Taobao, RetailRocket) and a real marketplace (Thumbtack): the inter-event clock--continuous-time decay, long known to beat frozen-intensity models--is nearly sufficient, and the conditioning the field keeps adding is redundant or harmful on top of it (statistically null on the public benchmarks, at most 0.06 NLL; null to mildly harmful on the marketplace). We do not claim to discover that decay helps; our contribution is the tools that turn "conditioning doesn't help" into a checkable, certified statement--plus an honest-evaluation account of the read-out/leakage pitfalls we hit and retracted.
Chinese Translation
从业者用越来越多的信号（生命周期价值、类别、新近度/频率、日历、地理）来丰富客户返回模型，时间点过程（TPP）文献也紧随其后，采用以协变量和外部协变量为条件的强度函数。但其中任何信号是否改善了时间预测，又如何知道呢？一个零假设（“特征 X 没有帮助”）只有在模型本来能够发现信号时才有意义。我们做出两项贡献——一种方法和一种度量——以可信地回答这个问题。(i) 一种筛选-确认协议，用于证明候选信号是否改善 TPP 的事件时间似然：阳性对照植入已知强度的耦合，并确认模型能够恢复该耦合，因此真实数据上的零结果可以被解读为“没有信号”，而不是“方法太弱”。该对照在分类编码和连续编码上，以及在一个真实时钟驱动数据集（纽约市出租车小时时段）上得到验证。(ii) 一种无模型上限，量化客户返回时间究竟有多少能够被点预测（来自任何协变量的间隔方差中只有个位数百分比；返回事件几乎是近无记忆的）。借助这些工具，我们在三个公开基准（Amazon、Taobao、RetailRocket）和一个真实市场（Thumbtack）上证明了一个干净的结果：事件间时钟——连续时间衰减，早已知道优于冻结强度模型——几乎足够，而该领域不断添加的条件化在其之上是冗余或有害的（在公开基准上统计上为零，至多 0.06 NLL；在市场平台上从零到轻度有害）。我们并不声称发现衰减有帮助；我们的贡献是这些工具，它们将“条件化没有帮助”转变为一种可检验、可证明的陈述，以及一份关于我们遇到并撤回的读出/泄漏陷阱的诚实评估记录。
cs.LG / 39 / 2608.11560
When Offline Evaluation Misleads: A Diagnostic Protocol for Reward and Policy Selection in Delayed-Feedback Contextual Bandits
当离线评估产生误导时：延迟反馈上下文老虎机中奖励与策略选择的诊断协议
Abstract
Personalizing marketing messages with contextual multi-armed bandits (CMABs) drives real business value, yet the objective that ultimately matters - a downstream conversion - is observed only weeks later, too late to drive online learning. Teams therefore train the bandit on a fast proxy reward, and separately must judge whether a contextual bandit is worth its complexity over sending one best message. Settling both decisions with the usual offline checks - a batch off-policy estimate, a marginal arm-discrimination test, a confidence interval - can mislead systematically under delayed feedback. We give an ordered diagnostic protocol that screens a reward-and-policy candidate on two axes, alignment (does optimizing the reward move the north-star?) and learnability (can the bandit identify the reward-optimal policy?), before trusting any reported lift. We validate it where the truth is known - a public off-policy-evaluation benchmark and a controllable synthetic generator - and illustrate it on a deployed large-marketplace push system (where, with five arms and one split, the evidence is directional rather than powered). Two lessons recur. (N1) A single offline number can mis-rank rewards: a denser reward signal gives the bandit more to learn from, so rewards that look tied in a static estimate pull apart once learning happens online. (N2) If you cannot tell in advance which single message is best, a per-user policy partly just avoids betting on the wrong one - that looks like personalization but is really robustness, so a "personalization premium" is easily overstated. Our contribution is methodological rather than algorithmic: the ordered protocol, the two lessons it surfaces, and the end-to-end experience of applying it to a delayed-feedback CMAB.
Chinese Translation
用上下文多臂老虎机（CMAB）个性化营销信息能够带来真实的商业价值，然而最终真正重要的目标——下游转化——要在数周后才被观测到，这太迟了，无法驱动在线学习。因此，团队会用快速的代理奖励来训练老虎机，并且还必须单独判断：与发送一条最佳消息相比，上下文老虎机的复杂性是否值得。在延迟反馈下，用常见的离线检查——批量离策略估计、边际臂区分检验、置信区间——来敲定这两项决策，可能会系统性地产生误导。我们给出一个有序的诊断协议，在信任任何报告的提升之前，从两个维度筛选“奖励-策略”候选：对齐性（优化该奖励会推动北极星指标吗？）和可学习性（老虎机能否识别出该奖励下的最优策略？）。我们在已知真实情况的环境中对其进行了验证——一个公开的离策略评估基准和一个可控的合成生成器——并在一套已部署的大型市场推送系统中对其进行了示例说明（在那里，由于有五个臂和一个分割，证据是方向性的，而不是具有充分功效的）。有两个教训反复出现。（N1）单一的离线数字可能会错误地对奖励排序：更密集的奖励信号给老虎机提供了更多可学习的内容，因此，在静态估计中看起来不相上下的奖励，一旦在线学习发生就会拉开差距。（N2）如果你无法提前判断哪一条单一消息最好，那么按用户制定策略在一定程度上只是避免押注在错误的那一条上——这看起来像个性化，但实际上是稳健性，因此“个性化溢价”很容易被高估。我们的贡献是方法论上的而非算法上的：这一有序协议、它所揭示的两个教训，以及将其应用于延迟反馈CMAB的端到端经验。
cs.LG / 40 / 2608.11567
Sparse and robust geometric twin support vector machine via asymmetric RoBoSS loss function
基于非对称 RoBoSS 损失函数的稀疏鲁棒几何孪生支持向量机
Abstract
In real-world scenarios, the training data usually contains redundant features, label noise and feature noise, which provide severe challenges for the efficiency of machine learning methods. Since standard support vector machine (SVM) adopts $l_2$-norm penalty and hinge loss function, it lacks the ability of selecting significant features and is sensitive to noise. To address these issues, this paper proposes a novel asymmetric, robust, bounded, sparse and smooth (aR) loss function for $l_1$-norm penalized geometric twin SVM (aRSGTSVM) to handle classification and regression tasks. The $l_1$-norm penalty can achieve the feature selection. The proposed aR loss function can not only effectively mitigate the impact of label noise, but also significantly enhance the stability to resampling noise, i.e., the zero-mean feature noise around the boundary hyperplanes. Furthermore, a statistical analysis of the robustness of aRSGTSVM was also conducted using the influence function. Since aRSGTSVM involves nonconvex and nonsmooth optimization, we develop a fast and stable proximal gradient descent based solving algorithm. Compared with related state-of-the-art methods, experimental results demonstrate the superiority of the proposed aRSGTSVM on both synthetic and UCI datasets. Furthermore, we apply aRSGTSVM to index tracking tasks, where results for tracking the different indices in the China stock market show that it can achieve satisfactory performance.
Chinese Translation
在现实场景中，训练数据通常包含冗余特征、标签噪声和特征噪声，这给机器学习方法的效率带来了严峻挑战。由于标准支持向量机（SVM）采用 $l_2$-范数惩罚和铰链损失函数，它缺乏选择重要特征的能力，并且对噪声敏感。为了解决这些问题，本文提出了一种新颖的非对称、鲁棒、有界、稀疏且光滑的（aR）损失函数，用于 $l_1$-范数惩罚的几何孪生支持向量机（aRSGTSVM），以处理分类和回归任务。$l_1$-范数惩罚可以实现特征选择。所提出的 aR 损失函数不仅能有效减轻标签噪声的影响，还能显著增强对重采样噪声（即边界超平面附近的零均值特征噪声）的稳定性。此外，还使用影响函数对 aRSGTSVM 的鲁棒性进行了统计分析。由于 aRSGTSVM 涉及非凸和非光滑优化，我们开发了一种快速且稳定的基于近端梯度下降的求解算法。与相关的最新方法相比，实验结果证明了所提出的 aRSGTSVM 在合成数据集和 UCI 数据集上的优越性。此外，我们将 aRSGTSVM 应用于指数跟踪任务，跟踪中国股票市场中不同指数的结果表明，它能够取得令人满意的性能。
cs.LG / 41 / 2608.11612
Dion3: Full-Stack Orthogonal Updates
Dion3：全栈正交更新
Abstract
The Muon optimizer incurs a significant overhead cost due to its cubic-time Newton-Schulz orthogonalization step. When weights are sharded, communication overhead compounds this computational cost, eroding the benefits of Muon in many settings. We present Dion3, a revision of Muon that targets this overhead at every level of the stack. Our Gram Newton-Schulz algorithm reduces the FLOP cost of orthogonalization, our CuteDSL kernels accelerate it by exploiting symmetry, and our megabatching strategy reduces communication overhead. Moreover, we propose a simple change to the update rule that cuts costs even further: selecting only a fraction of the momentum matrix's rows to orthogonalize at each step. This update rule improves on Dion (another "compressed" version of Muon), in both speed and performance. Overall, Dion3 matches or improves on the loss achieved by Muon but reduces optimizer step time by up to 6x. Dion3 is available via the dion package (https://github.com/microsoft/dion) as a drop-in replacement for Muon.
Chinese Translation
Muon 优化器因其立方时间的 Newton-Schulz 正交化步骤而产生显著的开销成本。当权重被分片时，通信开销会加剧这一计算成本，在许多场景中削弱 Muon 的优势。我们提出 Dion3，这是 Muon 的一个修订版本，它在栈的每个层面针对这一开销进行优化。我们的 Gram Newton-Schulz 算法降低了正交化的 FLOP 成本，我们的 CuteDSL 内核通过利用对称性加速该过程，我们的超大批次策略降低了通信开销。此外，我们提出对更新规则进行一项简单修改，以进一步降低成本：每一步仅选择动量矩阵的一部分行进行正交化。这一更新规则在速度和性能上都优于 Dion（Muon 的另一种“压缩”版本）。总体而言，Dion3 达到或改善了 Muon 所取得的损失，同时将优化器步长时间最多缩短 6 倍。Dion3 可通过 dion 软件包（https://github.com/microsoft/dion）获得，作为 Muon 的直接替代品。
cs.LG / 42 / 2608.11613
A Local Sinkhorn Framework for Conditional Distribution Reconstruction of Multidimensional Random Fields
一种用于多维随机场条件分布重建的局部 Sinkhorn 框架
Abstract
In this paper, we propose a local Sinkhorn divergence framework for conditional distribution reconstruction of multidimensional random fields. By utilizing the debiased Sinkhorn divergence, our proposed approach develops a differentiable and computationally efficient local distribution matching objective to train stochastic neural networks (SNNs). Furthermore, we establish theoretical generalization error estimates for our local Sinkhorn divergence framework, which explicitly characterizes the trade-off between approximation bias and statistical efficiency controlled by the regularization parameter and reveals how our proposed local Sinkhorn divergence loss function can be efficiently applied to learning multidimensional random field models. The proposed framework provides a scalable alternative to exact local optimal transport for conditional distribution reconstruction, offering a practical compromise between geometric fidelity, statistical efficiency, and computational scalability for uncertainty quantification and probabilistic scientific machine learning. Through various numerical examples, we compare our proposed local Sinkhorn divergence framework with other loss functions to train SNNs and with other machine-learning-based uncertainty quantification frameworks, demonstrating that the proposed local Sinkhorn divergence framework achieves an effective balance between reconstruction accuracy and computational efficiency while maintaining good scalability for multidimensional stochastic systems.
Chinese Translation
在本文中，我们提出了一种用于多维随机场条件分布重建的局部 Sinkhorn 散度框架。通过利用去偏 Sinkhorn 散度，我们提出的方法构建了一个可微分且计算高效的局部分布匹配目标，用于训练随机神经网络（SNN）。此外，我们为局部 Sinkhorn 散度框架建立了理论上的泛化误差估计，该估计明确刻画了由正则化参数控制的近似偏差与统计效率之间的权衡，并揭示了我们所提出的局部 Sinkhorn 散度损失函数如何能够有效地应用于学习多维随机场模型。所提出的框架为条件分布重建提供了一种相对于精确局部最优传输的可扩展替代方案，在几何保真度、统计效率和计算可扩展性之间为不确定性量化和概率科学机器学习提供了一种实用的折中方案。通过各种数值示例，我们将所提出的局部 Sinkhorn 散度框架与其他用于训练 SNN 的损失函数以及其他基于机器学习的不确定性量化框架进行了比较，结果表明所提出的局部 Sinkhorn 散度框架在重建精度和计算效率之间实现了有效的平衡，同时对于多维随机系统保持良好的可扩展性。
cs.LG / 43 / 2608.11638
Transferable Above-Ground Biomass (AGB) Estimation Model from Multi-Sensor Data with Sparse Field Calibration
基于多传感器数据与稀疏实地校准的可迁移地上生物量（AGB）估算模型
Abstract
Spatially continuous quantification of forest above-ground biomass (AGB) is what makes carbon accounting credible and mitigation strategies actionable. While field inventories provide high localized accuracy, they are spatially sparse; conversely, spaceborne LiDAR from the Global Ecosystem Dynamics Investigation (GEDI) offers broad biomass samples but lacks spatial continuity and systematic underestimation of high-biomass forests. This paper presents an operational framework centered on a single globally trained convolutional neural network (CNN) that is seamlessly adapted to each new landscape through a lightweight empirical field-calibration workflow. The global model combines optical (Sentinel-2), C-band SAR (Sentinel-1), L-band SAR (ALOS-2 PALSAR-2), and terrain (DEM) data. It is trained once against GEDI Level-4A biomass reference data spanning multiple regions and both wet and dry seasons so that it learns the persistent woody-structure rather than a single-date appearance. To avoid retraining for every landscape, the framework applies a small number of local field plots to fit a scale-and-bias correction that aligns the global prediction with ground truth in each region. The pipeline harmonizes sensor data onto a shared 10 m grid, derives vegetation indices and polarimetric ratios, computes per-band normalization stats, and trains the CNN with a hybrid log-domain SmoothL1 with RMSE loss for skewed biomass distribution. On held-out validation the global GEDI-based model achieved R^2 approximately 0.78 and RMSE approximately 22 Mg/ha. A subsequent field calibration combining Random Forest fine-tuning under a 10-fold cross-validation eliminates localized regional biases. This improves local validation performance to R^2 approximately 0.82 and reduces RMSE to approximately 15 Mg/ha, outperforming both the uncalibrated global model and the ESA CCI Biomass product against field plots.
Chinese Translation
对森林地上生物量（AGB）进行空间连续量化，是使碳核算可信、减缓策略可操作的关键。虽然实地清查能提供较高的局部精度，但它们在空间上稀疏；相反，来自全球生态系统动态调查（GEDI）的星载激光雷达可提供广泛的生物量样本，但缺乏空间连续性，并且对高生物量森林存在系统性低估。本文提出一个可操作的框架，其核心是一个经过全局训练的单一卷积神经网络（CNN），通过轻量级经验实地校准流程可无缝适配每一种新的景观。该全局模型结合了光学（Sentinel-2）、C波段SAR（Sentinel-1）、L波段SAR（ALOS-2 PALSAR-2）和地形（DEM）数据。它仅针对覆盖多个区域以及雨季和旱季的GEDI Level-4A生物量参考数据训练一次，从而学习持久的木质结构，而不是某一日期的表观特征。为了避免对每种景观重新训练，该框架使用少量局部实地样地来拟合尺度和偏差校正，使全局预测与每个区域的地面真值对齐。该流程将传感器数据协调到统一的10 m网格上，推导植被指数和极化比，计算逐波段归一化统计量，并针对偏斜的生物量分布，使用混合对数域SmoothL1与RMSE损失训练CNN。在留出验证中，基于GEDI的全局模型取得了约为0.78的R^2和约为22 Mg/ha的RMSE。随后结合随机森林微调并在10折交叉验证下进行的实地校准，消除了局部区域偏差。这将局部验证性能提升至R^2约为0.82，并将RMSE降低至约为15 Mg/ha，在针对实地样地的对比中优于未校准的全局模型和ESA CCI Biomass产品。
cs.LG / 44 / 2608.11654
Towards a Formal Definition of Agent Memory: Basis, Span, Optimality, and the Sequential Memory Problem
迈向智能体记忆的形式化定义：基、张成、最优性与序贯记忆问题
Abstract
Despite the wide deployment of memory in large-model agents, there is no unified formal account of what a memory is or when it is optimal. This paper takes a first step toward this account. The central idea is that memory is a basis, knowledge is its span, and answerability is a coverage problem: an agent stores events extracted from a material; a generation operator turns any event set into the knowledge it entails; and a query is answerable exactly when some single item in the span covers it. The optimal memory is then the capacity-constrained maximizer of expected coverage, and its value traces a utility--capacity frontier, the common yardstick on which memory systems can be compared. Next, we consider noise in the memory and discuss coverage versus precision under it: a memory may store false claims, so the write policy must infer the truth of what it stores. Drawing an analogy with biological memory, which is formed continuously through ongoing experience, we formalize the continual agent-memory problem in a sequential MDP that covers multiple levels, where memory is the state, writing is the action, and the utility settled at query time is the delayed reward that drives learning. To make the framework concrete, we instantiate it on Homer's \emph{Odyssey}, turning the frontier, the compression zone, and the divergence of coverage from precision into concrete numbers. Finally, we position existing systems within the framework, making ``how good is a memory'' measurable and recasting the open problems of constructing and learning agent memory as concrete research questions.
Chinese Translation
尽管记忆在大模型智能体中被广泛部署，但对于什么是记忆或何时记忆是最优的，尚不存在统一的形式化说明。本文朝着这一说明迈出了第一步。核心思想是：记忆是一个基，知识是其张成，可回答性是一个覆盖问题：智能体存储从材料中提取的事件；一个生成算子将任意事件集合转化为其所蕴含的知识；并且当一个查询恰好被张成中的某个单项覆盖时，该查询是可回答的。最优记忆便是在容量约束下最大化期望覆盖的记忆，其价值描绘出一条效用-容量前沿，这是可用来比较记忆系统的共同标尺。接下来，我们考虑记忆中的噪声，并讨论噪声下的覆盖与精度：记忆可能存储虚假陈述，因此写入策略必须推断其所存储内容的真实性。通过与生物记忆进行类比——生物记忆通过持续的经历不断形成——我们在一个涵盖多个层次的序贯MDP中将持续智能体记忆问题形式化，其中记忆是状态，写入是动作，而在查询时结算的效用是驱动学习的延迟奖励。为了使该框架具体化，我们在荷马的《奥德赛》上实例化它，将前沿、压缩区以及覆盖与精度的偏离转化为具体数字。最后，我们将现有系统置于该框架之中，使“记忆有多好”变得可量化，并将构建和学习智能体记忆的开放问题重塑为具体的研究问题。
cs.LG / 45 / 2608.11658
Is Per-Agent Policy Composition Safe? Rethinking Successor-Feature Transfer in Cooperative Multi-Agent Reinforcement Learning
逐智能体策略组合是否安全？重新思考合作多智能体强化学习中的后继特征迁移
Abstract
Many reinforcement learning systems, from fleet management to traffic signal control, must serve an objective that changes dynamically after deployment, and retraining a policy for each new objective is prohibitively expensive. For a single agent, this problem is well understood: successor features with generalized policy improvement, together with their universal extension, recombine a library of learned policies into a policy for any new objective, with a guarantee that the result is never worse than any policy in the library. However, multi-agent transfer has received far less attention, and the common practice of letting each agent recombine its own library independently inherits the recipe but not the guarantee. We prove that this independent composition can produce joint behavior strictly worse than every policy in the library, because recombining teammates changes the environment each agent faces and invalidates the values it relies on, a failure with no single-agent counterpart. We further show that the only unconditionally safe fixed rule is synchronized composition, which moves the whole team to one jointly trained policy but cannot serve objectives that assign different goals to different agents. To attain safety and flexibility at once, we propose MA-USFA, a hierarchical method with two layers: a lower layer of universal successor feature approximators that predicts each agent's successor features while conditioned on its teammates' objectives, and an upper composer that selects, across agents, which library entry each agent should follow and supplies the cross-agent correction a per-agent value cannot represent. Trained once over the distribution of objectives, it is applied at deployment with no per-task adaptation.
Chinese Translation
许多强化学习系统，从车队管理到交通信号控制，都必须服务于部署后动态变化的目标，而为每个新目标重新训练策略的代价高得难以承受。对于单个智能体，这个问题已得到很好的理解：后继特征与广义策略改进，连同它们的通用扩展，将学习到的策略库重组为适用于任何新目标的策略，并保证结果不会比库中的任何策略更差。然而，多智能体迁移受到的关注要少得多，而让每个智能体独立重组其自身库的常见做法继承了这一方法，却没有继承其保证。我们证明，这种独立组合可能产生比库中每一个策略都严格更差的联合行为，因为重组队友会改变每个智能体所面对的环境，并使其所依赖的价值失效，这是一种在单智能体情形中没有对应物的失败模式。我们进一步表明，唯一无条件安全的固定规则是同步组合，它将整个团队切换到一个联合训练的策略，但无法服务于为不同智能体分配不同目标的目标。为了同时实现安全性与灵活性，我们提出了 MA-USFA，一种具有两层结构的分层方法：下层是通用后继特征逼近器，在以队友目标为条件的情况下预测每个智能体的后继特征；上层是一个组合器，它在智能体之间选择每个智能体应遵循的库条目，并提供单个智能体价值无法表示的跨智能体校正。它在目标分布上只训练一次，部署时无需逐任务适配即可应用。
cs.LG / 46 / 2608.11661
Low-Interaction-Rank Learning: Unifying Multiplicative Dual-Encoder Heads
低交互秩学习：统一乘法双编码器头
Abstract
A multiplicative dual-encoder network computes a real-valued output for a pair of inputs as the inner product of their separate encodings. This architecture has been developed independently in operator learning, bipartite matching, contrastive vision-language models, retrieval, and other areas, yet no unified theory guides the basic design decisions: how many interaction modes to represent, how to normalize the encoders, and when the architecture should be avoided. We provide such a foundation by introducing the class of functions of low interaction rank, a class whose intrinsic complexity is measured by its interaction spectrum. Within this framework, approximation error decomposes into a spectral truncation term and an encoder-realization term; sample complexity is governed by the sum of the two encoder complexities rather than their product; and a usability criterion based on spectral decay determines when the architecture can succeed. The same framework exposes a central identifiability problem: the encoders are defined only up to a linear gauge symmetry that leaves the learned coordinates arbitrary. We show that normalization is gauge fixing and that whitening pins the interaction modes up to permutation and sign, thereby explaining the uninterpretability of contrastive dimensions and providing a constructive remedy. Experiments on synthetic kernels, operator learning, and CLIP models validate the theoretical predictions: spectral decay rates match the predicted scaling, whitening recovers the true modes, and independently trained CLIP models are related by a single rotation which, after removal by whitening, exposes interpretable concept axes. The code of this paper is provided at https://github.com/RS2002/Mul-Net .
Chinese Translation
乘法双编码器网络对一对输入计算一个实值输出，作为它们各自编码的内积。该架构已在算子学习、二分匹配、对比视觉-语言模型、检索以及其他领域被独立发展，但尚无统一理论来指导基本设计决策：应表示多少交互模式、如何对编码器进行归一化，以及何时应避免使用该架构。我们通过引入低交互秩函数类来提供这样的基础，该类的内在复杂度由其交互谱度量。在这一框架内，近似误差分解为一个谱截断项和一个编码器实现项；样本复杂度由两个编码器复杂度之和而非其乘积决定；基于谱衰减的可用性准则决定了该架构何时能够成功。同一框架揭示了一个核心的可辨识性问题：编码器仅被定义到相差一个使所学坐标任意的线性规范对称性的程度。我们证明归一化就是规范固定，并且白化将交互模式固定到只差置换和符号的程度，从而解释了对比维度的不可解释性，并提供了一种建设性的补救方法。在合成核、算子学习和 CLIP 模型上的实验验证了理论预测：谱衰减率与预测的缩放规律相匹配，白化恢复了真实模式，并且独立训练的 CLIP 模型之间由单一旋转相关联，在通过白化去除该旋转后，暴露出可解释的概念轴。本文代码提供于 https://github.com/RS2002/Mul-Net 。
cs.LG / 47 / 2608.11669
Rubric Dropout: A Simple Way to Mitigate Reward Hacking in Rubric-as-Reward RL
Rubric Dropout：一种缓解以评分标准为奖励的强化学习中奖励黑客行为的简单方法
Abstract
Reinforcement learning against rubrics, lists of criteria graded by an LLM judge, has become a standard way to post-train language models on tasks with no deterministic answer. The rubric, however, is a fixed proxy for quality, never a complete description of it, and a policy trained against it long enough will learn to exploit the difference. We measure this directly. Training Qwen3-8B with Group Relative Policy Optimization (GRPO) on medical and science rubrics and grading out-of-distribution (OOD) benchmarks with both the training judge and a stronger gold judge, we find that the two scores diverge during training. The training judge's score keeps climbing while the gold judge's score peaks and then falls, by 3 points on HealthBench-Hard and by 22 points on ResearchQA. A judge with a fixed bias would shift the gold curve by a constant, not send it down while the training score rises, so the divergence is reward hacking, not judge noise. We propose Rubric Dropout, a one-line fix borrowed from neuron dropout. At every step, we randomly drop a subset of the rubric's criteria before computing the reward, so the policy never optimizes the same rubric twice. The dropped subset is shared across each rollout group, so GRPO's group-relative advantages stay comparable, and evaluation always uses the full rubric. Comparing no dropout against dropout at 30% and 50% on both benchmark pairs, dropout raises the OOD gold score at every matched checkpoint (+1 to +2 points on HealthBench-Hard, +6 to +7 points on ResearchQA), lowers the two hacking measures we track, and costs nothing in domain. Sweeping the dropout fraction shows a broad 30-50% sweet spot, while the natural alternative, reweighting criteria by how useful they are to training, performs worse than no intervention at all in our setting.
Chinese Translation
以评分标准（即由 LLM 评判器打分的标准列表）为奖励信号进行强化学习，已成为在没有确定性答案的任务上对语言模型进行后训练的标准方式。然而，评分标准只是质量的一个固定代理指标，从来不是对质量的完整描述；针对它训练足够久的策略将学会利用这种差异。我们对这一现象进行了直接测量。在使用医疗和科学评分标准并采用群体相对策略优化（GRPO）训练 Qwen3-8B 时，我们同时用训练评判器和一个更强的金标准评判器对分布外（OOD）基准进行评分，发现两个分数在训练过程中出现分化。训练评判器的分数持续上升，而金标准评判器的分数先达到峰值然后下降：在 HealthBench-Hard 上下降 3 分，在 ResearchQA 上下降 22 分。一个具有固定偏差的评判器只会使金标准分数曲线平移一个常数，而不会在训练分数上升时使其下降；因此这种分化是奖励黑客行为，而不是评判器噪声。我们提出 Rubric Dropout，一个借鉴神经元 dropout 的一行代码修复方法。在每一步，我们在计算奖励之前随机丢弃评分标准中的一个子集，因此策略从不会两次优化完全相同的评分标准。被丢弃的子集在每个 rollout 组内共享，因此 GRPO 的组相对优势保持可比，而评估始终使用完整评分标准。在两组基准上比较无 dropout、30% dropout 和 50% dropout，dropout 在每个匹配检查点都提高了 OOD 金标准分数（在 HealthBench-Hard 上 +1 到 +2 分，在 ResearchQA 上 +6 到 +7 分），降低我们跟踪的两个黑客行为指标，并且在域内没有任何损失。对 dropout 比例进行扫描，结果显示 30%–50% 是一个较宽的最佳区间；而一种自然的替代方案——根据标准对训练的有用程度重新加权各个标准——在我们的设定中表现甚至比完全不干预更差。
cs.LG / 48 / 2608.11675
FunnelCausalNet: Funnel-aware Joint Conversion-Revenue Uplift for Multi-tier Coupon Allocation
FunnelCausalNet：面向多层优惠券分配的漏斗感知联合转化-收入提升
Abstract
Coupon campaigns seek to lift both conversion and revenue, but gross merchandise value (GMV) follows a deterministic funnel from conversion to conditional order value and is zero-inflated and heavy-tailed. We propose FunnelCausalNet, an uplift estimator coupling a binary conversion head with a nonnegative conditional-value head through $μ_{\mathrm{gmv}}=μ_{\mathrm{conv}}μ_{\mathrm{val}}$. Under explicit RCT, support, rate-gap, and cross-head covariance-control assumptions, an idealized leading-order MSE comparison identifies a regime in which funnel composition can reduce pointwise variance; this is a heuristic, not a guarantee for the shared-representation neural model. The estimator is paired with marginal split-conformal CATE summaries, combined through a Bonferroni union as audit bands, and a Lagrangian budgeted allocator using RCT-anchored estimates for subsidy-aware ROI accounting. On semi-synthetic multi-tier Criteo-MT7, FunnelCausalNet's mean AUUC_GMV is within one seed standard deviation of the leading feature-interaction baseline among eleven baselines, while a controlled ablation reduces GMV effect error versus direct GMV regression by 18--48% across tested zero-inflation regimes. On de-identified industrial Hotel-Coupon RCT logs with about 4.9 million hold-out exposure records per seed, expected-outcome evaluation sweeps full LP frontiers; FunnelCausalNet has the best seed-averaged mean DeltaROI at all seven correlated anchors from 10% to 60%, which we treat as descriptive frontier consistency rather than independent significance. On sparse binary-spend public benchmarks, revenue-focused rankers can dominate uplift-curve proxies, defining an explicit regime boundary.
Chinese Translation
优惠券活动旨在同时提升转化和收入，但商品交易总额（GMV）遵循从转化到条件订单价值的确定性漏斗，并且是零膨胀和重尾的。我们提出 FunnelCausalNet，一种通过 $\mu_{\mathrm{gmv}}=\mu_{\mathrm{conv}}\mu_{\mathrm{val}}$ 将二元转化头与非负条件价值头耦合的提升估计器。在明确的 RCT、支持、率差和跨头协方差控制假设下，理想化的首阶 MSE 比较识别出漏斗组合能够降低逐点方差的情形；这对共享表示神经模型而言是一种启发式，而非保证。该估计器与边际分裂共形 CATE 汇总配对，通过 Bonferroni 联合将其组合为审计带，并与使用 RCT 锚定估计进行补贴感知 ROI 核算的拉格朗日预算分配器配对。在半合成多层 Criteo-MT7 上，FunnelCausalNet 的平均 AUUC_GMV 处于十一个基线中领先特征交互基线的一个种子标准差范围内，而受控消融在测试的零膨胀情形下将 GMV 效应误差相较于直接 GMV 回归降低了 18--48%。在去标识化的工业 Hotel-Coupon RCT 日志（每个种子约 490 万条保留曝光记录）上，期望结果评估扫描完整 LP 前沿；FunnelCausalNet 在从 10% 到 60% 的所有七个相关锚点处均具有最佳的种子平均平均 DeltaROI，我们将其视为描述性前沿一致性而非独立显著性。在稀疏二元支出公共基准上，以收入为重点的排序器可以主导提升曲线代理，从而定义了明确的情形边界。
cs.LG / 49 / 2608.11690
Drift and Dependence: Layer-wise Information-Theoretic Bounds for Replay-Based Continual Learning
漂移与依赖：基于重放的持续学习的逐层信息论界
Abstract
Continual learning must absorb new tasks without erasing old ones, and replay---mixing a small buffer of past examples into current training---is among the most effective remedies for catastrophic forgetting. Yet its generalization behavior is shaped by two coupled effects that existing analyses fold into a single hypothesis-level quantity: finite memory replaces each past distribution with an empirical proxy, and repeated reuse couples the buffer, the current data, and the final hypothesis through a shared optimization trajectory. We develop a layer-wise information-theoretic framework that separates these effects at every depth. Our main result decomposes the expected generalization gap into a replay-induced representation drift and an optimization-dependence term, the latter further resolved into stability, plasticity, interaction, and residual-coupling components. Two refinements make the framework operational. A Wasserstein relaxation of the drift term, valid under support mismatch, yields a depth-dependent drift--sensitivity trade-off whose minimizer identifies which interior layer to stabilize. An SGLD instantiation of the optimization term reduces it to a trajectory-level log-determinant budget, exposing a curvature-aware gradient-alignment statistic that serves as an online diagnostic of task-wise forgetting. Controlled and benchmark experiments confirm the predicted memory scaling, the interior funnel, and the alignment signal's link to forgetting.
Chinese Translation
持续学习必须吸收新任务而不抹除旧任务，而重放——将一小缓冲区的过往样本混合到当前训练中——是应对灾难性遗忘最有效的方法之一。然而，其泛化行为由两个相互耦合的效应塑造，而现有分析将它们合并为一个假设层面的量：有限记忆用经验代理替代每个过往分布，并且反复重用通过共享的优化轨迹将缓冲区、当前数据和最终假设耦合在一起。我们提出了一个逐层的信息论框架，在每一深度上分离这些效应。我们的主要结果将期望泛化差距分解为一项由重放引起的表示漂移和一项优化依赖项，后者进一步分解为稳定性、可塑性、交互和残余耦合分量。两项改进使该框架具有可操作性。对漂移项进行的Wasserstein松弛在支撑不匹配情形下仍成立，并产生一个随深度变化的漂移–敏感性权衡，其最小化解标识出应稳定哪一个内部层。对优化项的一种SGLD实例化将其约化为轨迹层面的对数行列式预算，暴露出一个曲率感知的梯度对齐统计量，可作为任务层面遗忘的在线诊断指标。受控实验和基准实验证实了所预测的记忆规模缩放、内部漏斗现象以及对齐信号与遗忘之间的关联。
cs.LG / 50 / 2608.11691
LEMUR: Latent Entropy-aware Multimodal Unlearning via Visual-anchored Reasoning Redirection
LEMUR：通过视觉锚定推理重定向实现潜在熵感知的多模态遗忘
Abstract
Reinforcement-learning (RL) post-training equips multimodal large reasoning models (MLRMs) with exploratory chains of thought (CoT), substantially improving visual reasoning. However, we find that this capability introduces a distinct privacy vulnerability: even when a sensitive fact is successfully unlearned from the final answer, the model may still reproduce it in its reasoning trace. This leakage is substantially more pronounced in natively RL-trained MLRMs than in their non -reasoning base models, revealing a privacy risk that existing unlearning methods are not designed to address. We show that RL-induced exploration leaves sensitive content with a distinctive token-level entropy signature that is largely absent from base models. Based on this observation, we propose LEMUR, a fully training-free, inference-time unlearning framework for natively RL-trained multimodal models. LEMUR uses entropy dynamics as a control signal to identify when sensitive reasoning begins and when sanitization should stop. During this interval, it redirects the reasoning trajectory through entropy-modulated visual-anchor latent injection, replacing committed tokens with sanitized, probability-weighted embeddings re-grounded in the input image. Across diverse MLRMs, LEMUR consistently outperforms existing unlearning met hods in suppressing both reasoning-trace and answer leakage, while better preserving non-sensitive utility and output fluency. These results demonstrate that RL-induced entropy dynamics provide a distinctive signal for privacy leakage and that exploiting this signal enables effective training-free unlearning for reasoning-capable multimodal models.
Chinese Translation
强化学习（RL）后训练使多模态大推理模型（MLRMs）具备探索性思维链（CoT），显著提升了视觉推理能力。然而，我们发现这一能力引入了一种独特的隐私漏洞：即使敏感事实已从最终答案中成功遗忘，模型仍可能在其推理轨迹中复现该事实。这种泄漏在原生经 RL 训练的 MLRMs 中比在其非推理基础模型中更为显著，揭示了一种现有遗忘方法并未设计解决的隐私风险。我们表明，RL 诱导的探索会给敏感内容留下独特的 token 级熵特征，而这种特征在基础模型中基本不存在。基于这一观察，我们提出了 LEMUR，一个完全无需训练、推理时遗忘框架，用于原生经 RL 训练的多模态模型。LEMUR 使用熵动态作为控制信号，以识别敏感推理何时开始以及净化何时应停止。在此区间内，它通过熵调制的视觉锚定潜在注入来重定向推理轨迹，将已确定 token 替换为基于输入图像重新锚定的、经净化的概率加权嵌入。在多种 MLRMs 上，LEMUR 在抑制推理轨迹泄漏和答案泄漏方面始终优于现有遗忘方法，同时更好地保留了非敏感效用和输出流畅性。这些结果表明，RL 诱导的熵动态为隐私泄漏提供了一种独特信号，并且利用该信号能够为具备推理能力的多模态模型实现有效的无需训练遗忘。
cs.LG / 51 / 2608.11698
REOPD: Reliability-Adaptive Reward Extrapolation for On-Policy Distillation
REOPD：用于同策略蒸馏的可靠性自适应奖励外推
Abstract
On-policy distillation (OPD) trains a student on its own trajectories under dense token-level supervision from a teacher. Reward-extrapolation methods such as ExOPD amplify the teacher-reference log-likelihood ratio to move beyond direct imitation, but apply a single global coefficient $λ$ to every token. This can drive the student to fit extreme peaks in the implicit reward, causing reward hacking and unstable training, and the optimal $λ$ varies across domains, requiring costly sweeps. We propose REOPD, a reliability-adaptive reward extrapolation framework for OPD. REOPD combines a token-level compatibility weight with a batch-level adaptive budget, yielding a token-wise coefficient $λ_{b,t}=1+γ_b q_t$ that preserves teacher alignment while selectively extrapolating along reliable teacher-reference directions. It requires no verifier, reward model, value model, or extra rollout beyond standard OPD. REOPD outperforms G-OPD on single-teacher mathematics and on both domains in the multi-teacher setting, while matching G-OPD on single-teacher code, demonstrating effective fine-grained reliability adaptation across domains and teacher configurations.
Chinese Translation
同策略蒸馏（OPD）在来自教师的密集 token 级监督下，对学生自身的轨迹进行训练。诸如 ExOPD 等奖励外推方法会放大教师-参考对数似然比，以超越直接模仿，但对每个 token 都应用单一的全局系数 $λ$。这可能会促使学生拟合隐式奖励中的极端峰值，导致奖励黑客和训练不稳定，并且最优 $λ$ 因领域而异，需要昂贵的调参扫描。我们提出 REOPD，一种用于 OPD 的可靠性自适应奖励外推框架。REOPD 将 token 级兼容性权重与批次级自适应预算相结合，得到逐 token 系数 $λ_{b,t}=1+γ_b q_t$，该系数在保持教师一致性的同时，沿着可靠的教师-参考方向选择性地进行外推。除了标准 OPD 外，它不需要验证器、奖励模型、价值模型或额外的 rollout。REOPD 在单教师数学任务上以及多教师设置下的两个领域上均优于 G-OPD，同时在单教师代码任务上与 G-OPD 表现相当，证明了在跨领域和教师配置中有效的细粒度可靠性自适应。
cs.LG / 52 / 2608.11701
Consolidator: Learning Persistent Routed Memory Across Context Boundaries
Consolidator：学习跨上下文边界的持久路由记忆
Abstract
Copying short-term memory (STM) into a slower store can preserve state across a context boundary, but persistence alone does not ensure that the retained state influences subsequent memory access. We test this distinction in a Phasor Memory Network (PMNet) using Consolidator, a shared slot-local operator that transforms routed STM before accumulating it into long-term memory (LTM), without replaying the source tokens. After each consolidation, the KV cache and STM are cleared. The retained LTM can still be read and is also fed into the hierarchical router, thereby conditioning which explicit-memory slots subsequent inputs access. We evaluate this mechanism on a two-segment modulo-10 mapping task in which the second segment updates the mapping at the same memory address. Following a second consolidation and reset, a held-out query must recover the updated mapping from LTM. The backbone and memory interface are frozen, leaving only 12.35K Consolidator parameters trainable (0.041\% of a 29.95M model). Across five paired runs from the same STM-pretraining checkpoint, direct LTM routing raises updated-mapping recall from $44.38\pm1.94\%$ to $87.02\pm1.76\%$ ($+42.64\pm1.10$ percentage points), while immediate STM recall remains 89.90\% in both conditions; both train separate Consolidators and retain the same LTM read paths. Learned consolidation outperforms forced identity accumulation by $21.40\pm1.91$ percentage points without routing and $68.70\pm1.76$ with routing. Thus, on this task, consolidated LTM serves as both retrievable content and an access state that shapes subsequent slot selection.
Chinese Translation
将短期记忆（STM）复制到较慢的存储中可以跨上下文边界保留状态，但仅靠持久性并不能确保所保留的状态会影响后续的记忆访问。我们在相量记忆网络（PMNet）中通过 Consolidator 检验这一区别；Consolidator 是一种共享的、槽局部的算子，它在将经路由的 STM 累积到长期记忆（LTM）之前对其进行变换，而无需重放源 token。每次整合之后，KV 缓存和 STM 都会被清空。保留下来的 LTM 仍然可以被读取，并且也会被输入到分层路由器中，从而决定后续输入访问哪些显式记忆槽。我们在一个两段式的模 10 映射任务上评估该机制；在该任务中，第二段会在同一记忆地址更新映射。在第二次整合和重置之后，一个留出的查询必须从 LTM 中恢复更新后的映射。骨干网络和记忆接口被冻结，仅剩下 12.35K 个 Consolidator 参数可训练（占 29.95M 模型的 0.041\%）。在来自同一 STM 预训练检查点的五次配对运行中，直接 LTM 路由将更新映射的召回率从 $44.38\pm1.94\%$ 提高到 $87.02\pm1.76\%$（$+42.64\pm1.10$ 个百分点），而在两种条件下即时 STM 召回率均保持为 89.90\%；两者都训练了各自的 Consolidator，并保留了相同的 LTM 读取路径。学习到的整合优于强制恒等累积，在无路由时高出 $21.40\pm1.91$ 个百分点，在有路由时高出 $68.70\pm1.76$ 个百分点。因此，在该任务上，整合后的 LTM 既充当可检索的内容，也充当塑造后续槽选择的一种访问状态。
cs.LG / 53 / 2608.11704
Robust and Efficient Noisy-Label Time-Series Classification via Dynamic Time Warping Based Granular Ball Computing
基于动态时间规整的粒球计算实现鲁棒且高效的噪声标签时间序列分类
Abstract
Dynamic Time Warping (DTW)-based Nearest-Neighbor (NN) classifiers are effective for time-series classification but are vulnerable to mislabeled training samples and require numerous DTW computations during inference. We propose DTW-based Granular Ball Computing (DTW-GBC), which organizes temporally similar training samples into granular balls and performs classification at the granule level. We further develop two granular-ball construction strategies for DTW-GBC. Experiments on four benchmark datasets with symmetric label noise show that the two DTW-GBC variants generally mitigate the performance degradation caused by label noise while requiring substantially fewer comparisons than DTW-based 1-NN during inference. These findings suggest that DTW-GBC provides a favorable balance between classification robustness and inference efficiency.
Chinese Translation
基于动态时间规整（DTW）的最近邻（NN）分类器对时间序列分类有效，但容易受到错误标记训练样本的影响，并且在推理过程中需要大量 DTW 计算。我们提出基于 DTW 的粒球计算（DTW-GBC），该方法将时序相似的训练样本组织成粒球，并在粒级上进行分类。我们进一步为 DTW-GBC 提出了两种粒球构建策略。在四个带有对称标签噪声的基准数据集上的实验表明，两种 DTW-GBC 变体通常能够减轻标签噪声引起的性能下降，同时在推理过程中所需的比较次数远少于基于 DTW 的 1-NN。这些发现表明，DTW-GBC 在分类鲁棒性和推理效率之间提供了良好的平衡。
cs.LG / 54 / 2608.11713
High-dimensional Multi-objective Bayesian Optimization with Learned Variable Interactions
基于学习到的变量交互的高维多目标贝叶斯优化
Abstract
Multi-objective Bayesian optimization (MOBO) is effective in identifying the Pareto fronts for expensive black-box problems. However, most current MOBO approaches are limited to low-dimensional decision space due to its exponential sampling complexity. This paper presents decision variable interaction analysis-based MOBO, ViaMOBO, a generic framework for expensive multi-objective problems with high-dimensional decision space. The key idea of ViaMOBO is that it utilizes a variable interaction analysis model to determine whether the decision space can be completely or partially divided, and then performs local Bayesian optimization in the divided decision subspaces. Through the variable analysis model, it can be derived whether the objectives in black-box problems are separable, partially separable, or non-separable based on the potential independent or interdependent relationships among decision variables without any strong assumptions. We compare ViaMOBO with the state-of-the-art MOBO methods on both synthetic and real-world benchmarks. The experimental results demonstrate that ViaMOBO outperforms other related MOBO baselines in approximating the Pareto front of high-dimensional expensive multi-objective problems.
Chinese Translation
多目标贝叶斯优化（MOBO）在识别昂贵黑箱问题的帕累托前沿方面是有效的。然而，由于其指数级采样复杂度，当前大多数MOBO方法仅限于低维决策空间。本文提出了基于决策变量交互分析的MOBO，即ViaMOBO，这是一个面向具有高维决策空间的昂贵多目标问题的通用框架。ViaMOBO的核心思想是：利用变量交互分析模型判断决策空间能否被完全或部分划分，然后在划分后的决策子空间中执行局部贝叶斯优化。通过变量分析模型，可以在无需任何强假设的情况下，根据决策变量之间潜在的独立或相互依赖关系，推导出黑箱问题中的目标是可分离、部分可分离还是不可分离的。我们在合成基准和真实世界基准上将ViaMOBO与最先进的MOBO方法进行了比较。实验结果表明，在逼近高维昂贵多目标问题的帕累托前沿方面，ViaMOBO优于其他相关的MOBO基线方法。
cs.LG / 55 / 2608.11716
Chain-of-Thought Shows the Path to a Tree: Realizing Branching Complexity
思维链展示了通往树的路径：实现分支复杂性
Abstract
Chain of Thought (CoT) lifts the expressive ceiling of bounded-depth Transformers, with characterizations tying the number of CoT steps to circuit complexity classes. What remains largely missing are concrete instantiations with explicit, depth-bounded constructions, and the traversal procedures such characterizations presuppose. We close this gap for branching complexity. We give CoT realizations of depth-first search (DFS) and of Dijkstra algorithm, the latter subsuming breadth-first search, by unique hard-attention decoders of at most two layers, and use them as a shared computational substrate: reusing the DFS decoder yields the Strahler number of an $n$-vertex tree in $2n-1$ steps with four layers, and reusing the Dijkstra decoder yields its width in $n-1$ steps with three. Since computing the Strahler number of a binary tree given as a term is \textsf{NC\textsuperscript{1}}-complete, and our constructions handle arbitrary $n$-ary trees without layer normalization or positional encodings, this is a non-trivial witness for the linear-step regime of the CoT hierarchy. Exploiting the classical bijection between ordered trees and Dyck paths, itself realized by our DFS construction, which emits the path as it traverses, we give independent constructions for both measures on the path representation.
Chinese Translation
思维链（CoT）提高了有界深度 Transformer 的表达上限，其特征将 CoT 步数与电路复杂度类联系起来。仍然在很大程度上缺失的是具有显式、深度有界构造的具体实例，以及这些特征所预设的遍历过程。我们针对分支复杂性弥补了这一空白。我们通过至多两层的唯一硬注意力解码器给出了深度优先搜索（DFS）和 Dijkstra 算法的 CoT 实现，其中后者包含广度优先搜索，并将它们用作共享计算基底：复用 DFS 解码器可在 $2n-1$ 步内、用四层得出 $n$ 个顶点的树的 Strahler 数；复用 Dijkstra 解码器可在 $n-1$ 步内、用三层得出其宽度。由于计算以项形式给出的二叉树的 Strahler 数是 \textsf{NC\textsuperscript{1}}-完全的，且我们的构造处理任意 $n$ 叉树，而无需层归一化或位置编码，因此这是 CoT 层级线性步数阶段的一个非平凡见证。利用有序树与 Dyck 路径之间的经典双射——该双射本身由我们的 DFS 构造实现，它在遍历时输出路径——我们针对路径表示上的这两种度量给出了独立的构造。
cs.LG / 56 / 2608.11746
Epiplexity Guided Data Selection and Generation for Out-of-Distribution Generalization
面向分布外泛化的 Epiplexity 引导数据选择与生成
Abstract
Modern systems are increasingly expected to transfer across tasks not specified during training. What data facilitates generalization in these new, unanticipated settings? One hypothesis is that data with more structural information could contain shared circuits and subprograms that could be recycled in a wider array of downstream settings. Epiplexity, a recently proposed measure of the structural information a compute-bounded learner can extract from data, provides a mechanism to reason about this relationship. In this paper, we show how to operationalize epiplexity as an online training signal for data selection and synthetic data generation. For selection, we fit scaling laws to the training loss curves of natural data domains to predict the expected epiplexity gain as a function of training tokens, and use this signal to adaptively determine the sampling weights over domains during training. For synthetic data generation, we define a generator's reward as the change in learner epiplexity over a buffer of previously generated data and use REINFORCE policy gradients to guide the generator toward an epiplexity-maximizing distribution. In both cases, higher epiplexity predicts improved downstream performance on zero-shot and fine-tuning based tasks, supporting the hypothesis that data rich in structural information yield representations that transfer across domains.
Chinese Translation
现代系统日益被期望能够迁移到训练期间未指定的任务上。什么样的数据有助于在这些新的、未预料到的场景中实现泛化？一种假设是，具有更多结构信息的数据可能包含共享电路和子程序，这些电路和子程序可在更广泛的下游场景中被重复利用。Epiplexity 是最近提出的一种对受计算约束的学习者能够从数据中提取的结构信息的度量，它为对这种关系进行推理提供了一种机制。在本文中，我们展示了如何将 Epiplexity 操作化为一种在线训练信号，用于数据选择和合成数据生成。对于数据选择，我们将缩放定律拟合到自然数据域的训练损失曲线上，以预测期望的 Epiplexity 增益作为训练 token 数量的函数，并利用该信号在训练过程中自适应地确定各域上的采样权重。对于合成数据生成，我们将生成器的奖励定义为学习器在由先前生成的数据组成的缓冲区上的 Epiplexity 变化，并使用 REINFORCE 策略梯度将生成器引向一个使 Epiplexity 最大化的分布。在两种情况下，较高的 Epiplexity 都预示着在零样本任务和基于微调的任务上更好的下游性能，这支持了如下假设：结构信息丰富的数据能够产生可跨域迁移的表征。
cs.LG / 57 / 2608.11749
MOON: Multi-Objective OrthoNormalized Updates for Multitask Learning
MOON：用于多任务学习的多目标正交归一化更新
Abstract
Multi-objective optimization (MOO) has demonstrated significant success in multi-task learning by mitigating task conflicts through gradient manipulation. However, most existing methods flatten model parameters into vectors and perform gradient manipulation under Euclidean geometry, thereby overlooking the matrix structure prevalent in modern architectures such as Transformers. In this paper, we show that gradient manipulation in Euclidean space does not generally yield the steepest descent direction under matrix geometry, potentially limiting optimization efficiency. Drawing from the theory of steepest descent for matrix-valued parameters, we propose MOON (Multi-Objective OrthoNormalized Updates), which performs gradient manipulation under spectral--nuclear norm geometry and uses the orthonormalized manipulated gradient for parameter updates. Theoretically, for smooth non-convex objectives, we establish convergence of the averaged Pareto-stationarity measure at rates of $\mathcal{O}(T^{-1/2})$ in the deterministic setting and $\mathcal{O}(T^{-1/4})$ under stochastic gradients. Empirical results across various benchmarks show that MOON consistently improves both optimization efficiency and final multi-task performance. Our code is available at https://github.com/KunlinLyu/MOON.
Chinese Translation
多目标优化（MOO）通过梯度操纵缓解任务冲突，在多任务学习中已展现出显著成功。然而，现有大多数方法将模型参数展平为向量，并在欧氏几何下进行梯度操纵，从而忽视了现代架构（如 Transformer）中普遍存在的矩阵结构。在本文中，我们表明，欧氏空间中的梯度操纵在矩阵几何下通常不会产生最速下降方向，这可能限制优化效率。借鉴矩阵值参数最速下降理论，我们提出了 MOON（多目标正交归一化更新），该方法在谱--核范数几何下进行梯度操纵，并使用正交归一化后的操纵梯度进行参数更新。理论上，对于光滑非凸目标，我们建立了平均帕累托平稳性度量的收敛性：在确定性设置下收敛速率为 $\mathcal{O}(T^{-1/2})$，在随机梯度下为 $\mathcal{O}(T^{-1/4})$。跨多个基准测试的经验结果表明，MOON 一致地改善了优化效率和最终多任务性能。我们的代码可在 https://github.com/KunlinLyu/MOON 获取。
cs.LG / 58 / 2608.11790
High-Order Liquid Evidence Encoding for Gradual GNSS Spoofing Detection in Autonomous Driving
面向自动驾驶中渐进式 GNSS 欺骗检测的高阶液态证据编码
Abstract
Accurate Global Navigation Satellite System (GNSS)-based localization is essential for safe and reliable autonomous driving. However, spoofing attacks can manipulate vehicle position estimates. Continuous and subtle attacks are particularly difficult to detect because individual GNSS observations may remain plausible while the inconsistency between GNSS-implied displacement and onboard vehicle motion gradually increases. Existing methods often rely on static vehicle-behavior features or a single residual signal and do not explicitly model this evolution. To address this problem, we propose a causal high-order liquid evidence framework for GNSS spoofing detection. The method first constructs a physics-guided GNSS--motion inconsistency residual by comparing GNSS-implied displacement with onboard-motion-derived displacement. It then forms separate evidence streams for the residual level and its first- and second-order discrete variations, with relevant contextual cues selected according to the evidence order. Each stream is processed by a separate adaptive liquid encoder, and the resulting temporal states are hierarchically coupled to predict spoofing at the window endpoint using only current and past observations. Experiments on three subsets of the real-world AV-GPS dataset show that the proposed method achieves the highest F1-scores among the evaluated temporal models on Dataset~1 and Dataset~3, reaching 0.9535 and 0.9777, respectively. On Dataset~3, it detects both labeled normal-to-attack transitions within four sampling steps. Code and datasets are publicly available at: https://github.com/pangjunbiao/GNSS_Spoofing.git.
Chinese Translation
基于全球导航卫星系统（GNSS）的精确定位对于安全可靠的自动驾驶至关重要。然而，欺骗攻击可以操纵车辆位置估计。连续且隐蔽的攻击尤其难以检测，因为单个 GNSS 观测可能仍然看似合理，而 GNSS 隐含位移与车载运动之间的不一致性逐渐增大。现有方法通常依赖静态车辆行为特征或单一残差信号，并未显式建模这种演变过程。为了解决这一问题，我们提出了一种用于 GNSS 欺骗检测的因果高阶液态证据框架。该方法首先通过比较 GNSS 隐含位移与由车载运动导出的位移，构造一个物理引导的 GNSS--运动不一致残差。然后，它分别为残差水平及其一阶和二阶离散变化形成独立的证据流，并根据证据阶数选择相关的上下文线索。每个证据流由一个单独的自适应液态编码器处理，所得到的时间状态被分层耦合，从而仅使用当前和过去的观测来预测窗口终点处的欺骗。在真实世界 AV-GPS 数据集的三个子集上的实验表明，所提出的方法在评估的时间模型中，在 Dataset~1 和 Dataset~3 上取得了最高的 F1 分数，分别达到 0.9535 和 0.9777。在 Dataset~3 上，它在四个采样步内检测到了两个已标注的正常到攻击转变。代码和数据集公开可用：https://github.com/pangjunbiao/GNSS_Spoofing.git。
cs.LG / 59 / 2608.11797
Orientation, not magnitude: the causal structure of task-vector interference in merged language models
方向，而非幅度：合并语言模型中任务向量干扰的因果结构
Abstract
Model merging by task arithmetic works until it doesn't, and the field diagnoses why with magnitudes: layerwise representation bias, deviations from cross-task linearity, parameter overlap. Tracking the exact layerwise cross-term of merged LLMs through a factorial ledger and intervening on it directly, we find magnitude insufficient - and inconsistent across model families - as a diagnostic axis. An exact decomposition of the layerwise flux shows it is dominated by amplifying transport of the existing cross-term (~65-70% in both families, gain >1 per late block), and erasing the term is undone by propagation - rebuilt to 99% of its norm at cosine 0.99 - unless applied near the output; a basin test with six starting displacements establishes the carried direction as an attractor of the forward pass. That direction is causally load-bearing: erasure along it removes expressed interference dose-dependently and saturates at exact erasure, while norm-matched wrong-direction controls fail or backfire. Instruction wrappers gate the effect: the same erasure finds 13x less relative interference to remove under a wrapper that internally amplifies the cross-term, because the wrapper drowns the interaction in a template-pinned main effect rather than shrinking it - a structure that replicates across further instruction templates but not under a length-matched control. Magnitude, by contrast, is at best a coarse correlate, and the striking +-15% "universality" of naive bfloat16 generation turns out to be quantization roughness. Task pairs whose local cross-term generation differs by at most 1.9x differ by 14x-337x in causally removable interference. All 46 predictions were preregistered and frozen before their data; falsifications, including of our own headline expectations and of behavioral recovery under a validated continuous endpoint, are reported as such.
Chinese Translation
基于任务算术的模型合并在失效之前一直有效，而该领域用幅度来诊断其失效原因：逐层表示偏差、对跨任务线性的偏离、参数重叠。通过一个因子台账追踪合并后大语言模型的精确逐层交叉项，并直接对其进行干预，我们发现幅度作为诊断轴是不充分的——并且在模型家族之间不一致。对逐层通量的精确分解显示，它主要由对现有交叉项的放大传输所主导（在两个家族中约为65–70%，每个后期块增益>1）；并且擦除该项会被传播所抵消——除非在接近输出处施加，否则它会被重建到其范数的99%、余弦相似度0.99；一项使用六个起始位移的吸引盆测试证实，所携带的方向是前向传播的一个吸引子。该方向在因果上具有承重作用：沿着它进行擦除会以剂量依赖方式移除已表达的干扰，并在精确擦除时达到饱和；而范数匹配的错误方向对照则失败或适得其反。指令包装器对该效应起门控作用：在内部放大交叉项的包装器下，同一种擦除可移除的相对干扰少13倍，因为该包装器将交互作用淹没在模板固定的主效应中，而不是将其缩小——这一结构在更多指令模板中复现，但在长度匹配的对照下不复现。相比之下，幅度充其量只是一个粗略相关因素；朴素 bfloat16 生成中引人注目的 ±15% “普遍性” 原来是量化粗糙度所致。局部交叉项生成最多相差1.9倍的任务对，在可因果移除的干扰上却相差14倍至337倍。全部46项预测均在数据出现之前预先注册并冻结；对预测的证伪——包括对我们自己主要预期的证伪，以及对经过验证的连续终点下行为恢复的证伪——均如实报告。
cs.LG / 60 / 2608.11801
JAPE: Joint Anomaly Prediction and Intrinsic Explanation in Multivariate Time Series
JAPE：多变量时间序列中的联合异常预测与内在解释
Abstract
Multivariate time-series anomaly prediction aims to identify whether and when anomalies will occur over a future horizon from historical observations. Existing methods primarily characterize anomalies as deviations in future numerical values, which may overlook subtle dependency changes induced by weak anomaly precursors and provide no native variable-level explanation together with the alert. To bridge these gaps, we propose JAPE, a Joint Anomaly Prediction and Explanation framework that lifts anomaly prediction from numerical-deviation modeling to dependency-structure modeling. JAPE is the first anomaly prediction framework to explicitly model evolving dependency structures for both point-wise alerting and native variable-level explanation. Specifically, JAPE (i) proposes a Decoupled Spatio-Temporal Representation (DSTR) backbone that decouples temporal and spatial modeling and captures lag-aware dependencies via learnable lag aggregation, thereby perceiving structural precursors before numerical deviations emerge; (ii) designs a dual-view alerting mechanism that fuses numerical forecasts with evolving dependency graphs for point-wise anomaly prediction, capturing structural evidence even under subtle numerical deviations; and (iii) presents Native Predictive Explanation (NPE), which directly reuses the predicted dependency graphs to rank variables by structural deviations without additional models or training. Extensive experiments on five real-world benchmarks across three prediction horizons demonstrate that JAPE improves average F1 and AUC-PR by 19.7% and 41.3%, respectively, while improving explainability with 26.6% gain in MRR.
Chinese Translation
多变量时间序列异常预测旨在从历史观测中识别未来时间范围内是否以及何时会发生异常。现有方法主要将异常描述为未来数值的偏离，这可能忽略由微弱异常前兆引起的细微依赖变化，并且不随警报提供原生的变量级解释。为了弥合这些差距，我们提出了 JAPE，一个联合异常预测与解释框架，将异常预测从数值偏离建模提升到依赖结构建模。JAPE 是首个显式建模演化依赖结构、同时用于逐点警报和原生变量级解释的异常预测框架。具体而言，JAPE (i) 提出了一个解耦时空表示（DSTR）骨干网络，它将时间建模和空间建模解耦，并通过可学习的滞后聚合捕获滞后感知的依赖关系，从而在数值偏离出现之前感知结构前兆；(ii) 设计了一种双视图警报机制，将数值预测与演化依赖图融合用于逐点异常预测，即使在细微数值偏离下也能捕获结构证据；以及 (iii) 提出了原生预测解释（NPE），它直接复用预测的依赖图，根据结构偏离对变量进行排序，无需额外模型或训练。在五个真实世界基准数据集上、跨越三个预测时间范围的广泛实验表明，JAPE 将平均 F1 和 AUC-PR 分别提高了 19.7% 和 41.3%，同时以 26.6% 的 MRR 增益提升了可解释性。
cs.LG / 61 / 2608.11815
Learning with Bilevel-Minimax Optimization for Efficient and Reliable Transfer Attacks
通过双层极小极大优化学习实现高效可靠的迁移攻击
Abstract
Transfer-based adversarial attacks craft adversarial examples using surrogate models to mislead black-box victim models. Beyond perturbation generation, transferability is fundamentally governed by the coupling of initialization, surrogate adaptation, and gradient dynamics. We revisit this challenge from a bilevel-minimax perspective and propose BMAT (Bilevel-Minimax Adversarial Transfer). The bilevel formulation captures the dependency between initialization and perturbation, while the inner minimax problem promotes surrogate robustness for cross-architecture generalization. Algorithmically, we develop an integrated bottom-up solver that combines a Soft Weight Modulator and an Implicit Gradient Approximator to enable ternary coupling among initialization, surrogate adaptation, and perturbation optimization. We further provide theoretical insights into the optimization dynamics of the proposed bilevel-minimax framework. Extensive experiments on classification and segmentation benchmarks show that BMAT outperforms more than 10 strong baselines across more than 30 victim models, improving both intra- and cross-architecture transfer and yielding up to a 2x reduction in mIoU. Code is available at https://github.com/callous-youth/BMAT.
Chinese Translation
基于迁移的对抗攻击通过替代模型构造对抗样本来误导黑盒受害者模型。除扰动生成外，可迁移性本质上由初始化、替代模型自适应和梯度动力学之间的耦合所决定。我们从双层极小极大视角重新审视这一挑战，并提出 BMAT（Bilevel-Minimax Adversarial Transfer，即双层极小极大对抗迁移）。该双层优化形式刻画了初始化与扰动之间的依赖关系，而内层极小极大问题提升了替代模型的鲁棒性，以实现跨架构泛化。在算法层面，我们设计了一种集成的自底向上求解器，它结合了软权重调制器（Soft Weight Modulator）和隐式梯度近似器（Implicit Gradient Approximator），以实现初始化、替代模型自适应与扰动优化三者之间的耦合。我们进一步对所提出的双层极小极大框架的优化动力学提供了理论见解。在分类和分割基准上的大量实验表明，BMAT 在 30 多个受害者模型上优于 10 多个强基线，同时提升了架构内和跨架构可迁移性，并实现了高达 2 倍的 mIoU 下降。代码可在 https://github.com/callous-youth/BMAT 获取。
cs.LG / 62 / 2608.11829
Towards Understanding On-Policy Distillation through the Lens of Test-Time Scaling
从测试时缩放的视角理解同策略蒸馏
Abstract
On-policy distillation (OPD) has emerged as a promising post-training technique for enhancing LLM reasoning. It is commonly believed to enable the student model to distill knowledge from a stronger teacher model, thereby expanding capabilities beyond the pre-OPD base model. In this study, we examine this view through the lens of test-time scaling by varying the sampling budget K and evaluating performance with pass@K and avg@K. Specifically, across several OPD variants, we observe that OPD-trained models maintain superior avg@K performance across sampling budgets, while the advantage in pass@K gradually shifts to the pre-OPD base models as K increases. These results suggest that OPD primarily improves sampling efficiency rather than consistently expanding the student's reasoning capability boundary. The pass@K dynamics throughout OPD training further reveal a progressive shift toward stronger small-K performance at the expense of the large-K capability boundary. Furthermore, a problem-level solvability analysis using pass@1024 as the criterion reveals an asymmetry: OPD causes more previously solvable problems to become unsolvable than previously unsolvable problems to become solvable. Together, these findings suggest that, from the perspective of capability expansion, OPD behaves more like an "illusory distillation": its apparent gains arise primarily from improved sampling efficiency rather than from acquiring genuinely new reasoning capabilities from the teacher.
Chinese Translation
同策略蒸馏（OPD）已成为增强大语言模型推理能力的一种有前景的后训练技术。通常认为，它使学生模型能够从更强的教师模型中蒸馏知识，从而将能力扩展到未经 OPD 的基座模型之外。在本研究中，我们通过改变采样预算 K 并用 pass@K 和 avg@K 评估性能，从测试时缩放的视角审视这一观点。具体而言，在多个 OPD 变体中，我们观察到经过 OPD 训练的模型在不同采样预算下保持更优的 avg@K 性能，而随着 K 增大，pass@K 上的优势逐渐转向未经 OPD 的基座模型。这些结果表明，OPD 主要提高采样效率，而非持续扩展学生模型的推理能力边界。在整个 OPD 训练过程中，pass@K 的动态进一步揭示出一种逐渐向更强的小 K 性能转变、并以牺牲大 K 能力边界为代价的趋势。此外，以 pass@1024 为标准的题目级可解性分析揭示了一种不对称性：OPD 导致更多原本可解的问题变得不可解，而将原本不可解的问题变得可解的数量更少。综合来看，这些发现表明，从能力扩展的角度来看，OPD 更像是一种“幻觉式蒸馏”：其表面上的收益主要来自采样效率的提高，而不是从教师模型那里获得了真正新的推理能力。
cs.LG / 63 / 2608.11831
Kernel Methods for Learning Operators with Multiple Inputs and Outputs
用于学习多输入多输出算子的核方法
Abstract
Learning mappings between infinite-dimensional objects is a central challenge in scientific machine learning. We introduce a general kernel-based encoder-decoder framework for operator learning that separates observation, representation, learning, and reconstruction. We develop this framework for multi-input, multi-output operator learning, where operators map between products of potentially distinct function spaces. Our approximation theory shows that, although the number of inputs and outputs can increase, the convergence rate is governed by the most challenging constituent approximation problem rather than the overall problem dimension. The framework leads to practical kernel methods with closed-form training and inference, combining mathematical tractability with computational efficiency. We further specialize the approach to multiple operator learning by introducing KernelMO, a family of kernel methods with complementary operator-valued and product-space formulations. Across five families of parametric partial differential equations, the proposed methods achieve competitive or state-of-the-art predictive accuracy while reducing training and inference costs relative to neural operator architectures and deep learning based models, offering an efficient and lightweight alternative.
Chinese Translation
学习无限维对象之间的映射是科学机器学习中的一个核心挑战。我们引入一个通用的基于核的编码器-解码器框架用于算子学习，该框架将观测、表示、学习和重建分离开来。我们针对多输入、多输出算子学习开发该框架，其中算子是在可能不同的函数空间的乘积之间进行映射。我们的逼近理论表明，尽管输入和输出的数量可能增加，收敛速率由最具挑战性的组成逼近问题而非整体问题维度决定。该框架产生了具有闭式训练和推断的实用核方法，将数学可处理性与计算效率结合起来。我们进一步通过引入 KernelMO 将该方法专门用于多算子学习，KernelMO 是一类具有互补的算子值形式和乘积空间形式的核方法。在五类参数化偏微分方程上，所提方法取得了具有竞争力或最先进的预测精度，同时相较于神经算子架构和基于深度学习的模型降低了训练和推断成本，提供了一种高效且轻量化的替代方案。
cs.LG / 64 / 2608.11839
Air Quality Station Simulation via LSTM and Attention-Based Modelling
基于LSTM与注意力机制建模的空气质量站点模拟
Abstract
Poor air quality in urban areas is driven by a complex chain of processes and presents a significant public health concern. To better understand and control the mechanisms that determine air quality, cities deploy networks of measurement stations, and launch initiatives for collecting denser data about the concentration of pollutants in the atmosphere. Extracting insights from the stations relies on their reliable and uninterrupted operation. However, hardware is susceptible to faults and black- outs that may result in data unavailability, which affects the overall quality of analyses. In this paper, we present a deep-learning model, called SATADL, which can infer complex relations and output multiple-hour-ahead air-quality forecasts. The goal of the model is to simulate the mea- surements of an unresponsive station until its operation is restored. The architecture of the model, which allows it to extract information from different aspects of the data, is described in detail and a careful examination of all of its components is provided. We demonstrate the performance of SATADL on four sets of air quality stations from around the world, by using it to simulate the concentration of PM10 for periods of hypothetical failures of one of the measurement stations, lasting for as long as 48 hours. A selection of baseline and published deep learning models were trained and used as a benchmark. The results show that SATADL per- forms better across different prediction windows, for both coefficient of determination and root mean squared error, demonstrating its suitability as a virtual proxy station.
Chinese Translation
城市地区的空气质量差是由复杂的过程链所驱动的，并构成一个重大的公共卫生问题。为了更好地理解和控制决定空气质量的机制，城市部署了测量站网络，并启动多项举措来收集更密集的大气污染物浓度数据。从这些站点中提取见解依赖于其可靠且不间断的运行。然而，硬件容易发生故障和断电，可能导致数据不可用，从而影响分析的总体质量。在本文中，我们提出了一种名为SATADL的深度学习模型，该模型能够推断复杂关系并输出提前多个小时的空气质量预测。该模型的目标是模拟一个无响应站点的测量值，直到其运行恢复。我们详细描述了该模型的架构，该架构使其能够从数据的不同方面提取信息，并对其所有组成部分进行了仔细考察。我们使用SATADL模拟其中一个测量站在假设故障期间（持续时间长达48小时）的PM10浓度，从而在来自世界各地的四组空气质量站点上展示了其性能。我们训练了一组基线模型和已发表的深度学习模型，并将其用作基准。结果表明，SATADL在不同预测窗口下均表现更优，在决定系数和均方根误差方面均如此，证明了其作为虚拟代理站点的适用性。
cs.LG / 65 / 2608.11859
Small-Scale Experiments: Are We There Yet?
小规模实验：我们做到了吗？
Abstract
Scaling laws promised cost-effective experiments; six years later, they have yet to fully deliver. Instead, researchers have found them unreliable at small scales (starting at 4M parameters) and concluded that sizable models cannot be avoided. We show this is not the case: the confounding factor is hyperparameters. Small models are highly sensitive, but hyperparameter sensitivity fades with scale. This small-scale sensitivity makes scaling laws easy to miss because they only emerge on the fully tuned frontier, and reaching that frontier requires an extensive search far beyond what most ever run. By ablating the basic scaling law recipe, we show well-tuned hyperparameters matter more than any other ingredient. Further, we reveal why those hyperparameters become easier to find: as scale increases, the hyperparameter loss surface becomes lower dimensional. Nevertheless while scaling laws exist in small models, extrapolation hits statistical limitations. A holistic approach is required. Synthesizing our insights with the recent literature, we develop a new methodology for model-centric research and demonstrate it on a question that once took the field years to settle: where to place normalization layers in the transformer architecture. From small-scale experiments, we recover the large scale result: pre-normalization works better as models grow in size. With the right tools and a better understanding, small-scale experiments can deliver on scaling laws' long-awaited promise.
Chinese Translation
缩放定律曾承诺实现具有成本效益的实验；六年后，它们仍未完全兑现。相反，研究人员发现它们在规模较小时（从4M参数开始）不可靠，并得出结论：无法避免使用较大规模的模型。我们证明情况并非如此：干扰因素是超参数。小模型高度敏感，但超参数敏感度随规模增大而减弱。这种小规模敏感度使得缩放定律容易被忽略，因为它们只在完全调优的前沿显现，而达到该前沿所需的搜索范围远超大多数人通常执行的规模。通过消融基本的缩放定律配方，我们表明，经过良好调优的超参数比任何其他成分都更重要。此外，我们揭示了为什么这些超参数变得更容易找到：随着规模增大，超参数损失曲面变得更低维。然而，尽管缩放定律在小模型中存在，外推却遇到统计局限。需要一种整体性方法。综合我们的见解与近期文献，我们开发了一种以模型为中心的研究新方法，并在一个曾让该领域花费多年才解决的问题上进行了演示：在Transformer架构中归一化层应放置在哪里。通过小规模实验，我们复现了大规模结果：随着模型规模增大，前置归一化效果更好。借助正确的工具和更深入的理解，小规模实验能够兑现缩放定律期待已久的承诺。
cs.LG / 66 / 2608.11868
Forward and Inverse Virtual Metrology for Phototransistor Gain: A Hierarchical, Uncertainty-Aware Approach for Small Production Datasets
光电晶体管增益的正向与逆向虚拟量测：一种面向小规模生产数据集的层次化、不确定性感知方法
Abstract
The customization, optimization and stabilization of the process flow of a silicon bipolar phototransistor commits months of cleanroom time before a finished device can be measured, so a model that predicts device gain from process parameters before a run has value out of proportion to its accuracy. We study this problem on a real fabrication history, thirteen to fourteen process runs of a single device: a small-sample, hierarchically structured setting unlike the large-corpus regime of conventional virtual metrology. Decomposing the variance of device gain, we find that roughly half of it lies between process runs rather than within them, so recipe-only prediction is bounded by construction. Building on these findings we provide a forward gain predictor with a relative, uncertainty-aware signal, an inverse search that returns recipes for a target gain, and, as the foundation for all of it, a multi-level data-quality assessment tailored to the nested physical entities of fabrication (batch, wafer, die) with an explicit cross-level linkage score. The normalized dataset and analysis code are released for full reproducibility.
Chinese Translation
硅双极光电晶体管工艺流程的定制、优化与稳定化在成品器件能够被测量之前需要投入数月的洁净室时间，因此，一个能在运行前根据工艺参数预测器件增益的模型，其价值与其精度不成比例地高。我们在一段真实的制造历史，即单一器件的十三至十四次工艺运行上研究该问题：这是一种小样本、层次化结构的环境，不同于传统虚拟量测的大规模语料范式。通过分解器件增益的方差，我们发现其中大约一半位于工艺运行之间而非运行之内，因此仅基于配方的预测在构造上就受到固有限制。基于这些发现，我们提供：一个具有相对且不确定性感知信号的正向增益预测器；一个为给定目标增益返回配方的逆向搜索；以及作为上述一切基础、针对制造中嵌套物理实体（批次、晶圆、裸片）量身定制并带有显式跨层级关联评分的多层级数据质量评估。标准化数据集和分析代码已发布，以确保完全可复现性。
cs.LG / 67 / 2608.11873
DCM Bandits: Multiplayer Information Asymmetric Cascading Bandits for Multiple Clicks
DCM 老虎机：面向多次点击的多智能体信息非对称级联老虎机
Abstract
In this work, we extend the Dependent Click Model (DCM) Bandits to a multiplayer information-asymmetric setting, where multiple agents interact with a shared ranked list and may observe multiple clicks per session, introducing new challenges for selection strategies. We study asymmetry in (1) actions and (2) rewards, providing sublinear regret guarantees for three settings where at least one asymmetry is present. Establishing matching information-theoretic lower bounds for these settings is left as an open problem. We further show that for small termination probabilities, the termination ranking need not be known, improving on prior single-agent results. Experiments confirm that our algorithms perform well across asymmetric environments and highlight the critical role of feedback structure, specifically the distinction between full versus first-click feedback, in coordinating exploration and minimizing regret.
Chinese Translation
在本文中，我们将依赖点击模型（DCM）老虎机扩展到多智能体信息非对称场景，其中多个智能体与共享的排序列表交互，并且每个会话可能观察到多次点击，这为选择策略带来了新的挑战。我们研究了（1）动作和（2）奖励中的非对称性，并针对至少存在一种非对称性的三种场景提供了次线性遗憾保证。为这些场景建立匹配的信息论下界仍是一个开放问题。我们进一步证明，对于较小的终止概率，终止排序无需已知，这改进了先前单智能体的结果。实验证实，我们的算法在各种非对称环境中表现良好，并凸显了反馈结构的关键作用，尤其是完整反馈与首次点击反馈之间的区别在协调探索和最小化遗憾方面的作用。
cs.LG / 68 / 2608.11909
Disentangling the Expressivity of RoPE
厘清 RoPE 的表达能力
Abstract
Two accounts recur in explanations of the success of rotary position embeddings (RoPE). Expressivity studies associate periodic position information with modular predicates, whereas mechanistic and long-context studies emphasize positional anchors and local offsets. We formalize both accounts for fully uniform, finite-precision soft-attention transformers. We find that, if every rotary component is periodic, RoPE transformers recognize exactly the languages definable in past temporal logic with modular predicates. Conventional RoPE is different: The rotations it computes never repeat. This yields a precision-dependent bounded simulation of fixed-offset look-back operators, rather than an all-length modular characterization. Controlled experiments match this separation: Constructed periodic schedules length-generalize on modular languages, while conventional RoPE behaves more like a bounded locality bias and can impair tasks requiring position-invariant access to distant context. Altogether, our findings shed light on RoPE transformers, bringing theoretical expressivity characterizations closer to models used in practice.
Chinese Translation
在解释旋转位置嵌入（RoPE）取得成功的原因时，有两种论述反复出现。表达能力研究将周期性位置信息与模谓词联系起来，而机制研究与长上下文研究则强调位置锚点和局部偏移。我们针对完全均匀的、有限精度的 soft-attention Transformer，将这两种论述形式化。我们发现，如果每个旋转分量都是周期性的，那么 RoPE Transformer 恰好能识别可在带模谓词的过去时态逻辑中定义的语言。传统 RoPE 则不同：它计算出的旋转永远不会重复。这带来的是对固定偏移回看算子的、依赖于精度的有界模拟，而非一种适用于所有长度的模刻画。受控实验与这一区分相一致：构造的周期性调度在模语言上能实现长度泛化，而传统 RoPE 的表现更像一种有界局部性偏置，并且可能损害需要以位置不变方式访问远端上下文的任务。总而言之，我们的研究结果阐明了 RoPE Transformer 的特性，使理论表达能力刻画更接近实践中使用的模型。
cs.LG / 69 / 2608.11917
A Factor Graph Approach to Scalable Multi-Output Gaussian Process Regression
一种面向可扩展多输出高斯过程回归的因子图方法
Abstract
Multi-output Gaussian process regression scales cubically in the number of observations times outputs, and dense kernel-matrix methods need bespoke handling whenever different outputs are observed at different inputs. We express multi-output Gaussian process regression as a Forney-style factor graph in which a nearest-neighbor chain orders a fixed candidate set of $C$ inputs into a one-dimensional sequence. Along this chain, latent Matérn processes evolve through linear-Gaussian transition factors, while the linear model of coregionalization mixes $L$ latent processes into $D$ outputs through a deterministic mixing factor and per-output scalar observation factors. Posterior computation reduces to exact Gaussian message passing on the chain at cost $\mathcal{O}(C(DL^2 + L^3))$ after chain construction, and missing observations omit their local factor without any covariance-matrix restructuring. The formulation therefore scales in the number of data samples and in the rate of missing observations, while remaining best suited to candidate sets in low input dimension.We compare the factor-graph formulation against an exact kernel-matrix baseline, a sparse-variational inducing-point baseline, and a nearest-neighbor baseline on a synthetic input-dimension sweep and on electricity time series forecasting. At low input dimension the factor-graph posterior tracks the exact kernel-matrix posterior closely, and the gap grows gradually as input dimension increases while staying competitive with both approximate baselines. On the electricity time series our factor-graph formulation matches all three baselines in forecast accuracy while scaling linearly in the number of data points, where the exact kernel-matrix method becomes infeasible and the inducing-point baseline remains substantially slower.
Chinese Translation
多输出高斯过程回归在观测数乘以输出数上的计算规模呈三次方增长，并且当不同输出在不同输入处被观测时，稠密核矩阵方法需要定制化处理。我们将多输出高斯过程回归表示为一个 Forney 式因子图，其中最近邻链将固定的由 $C$ 个输入组成的候选集排序为一维序列。沿着这条链，潜在 Matérn 过程通过线性高斯转移因子演化，而协同区域化线性模型通过确定性混合因子和每个输出的标量观测因子将 $L$ 个潜在过程混合为 $D$ 个输出。链构建后，后验计算归结为链上的精确高斯消息传递，成本为 $\mathcal{O}(C(DL^2 + L^3))$，并且缺失观测会省略其局部因子而无需任何协方差矩阵重构。因此，该公式在数据样本数量和缺失观测率方面具有可扩展性，同时仍然最适合低输入维度的候选集。我们在合成输入维度扫描和电力时间序列预测上，将因子图公式与精确核矩阵基线、稀疏变分诱导点基线和最近邻基线进行了比较。在低输入维度下，因子图后验紧密跟随精确核矩阵后验，随着输入维度增加，差距逐渐增大，但仍与两个近似基线保持竞争力。在电力时间序列上，我们的因子图公式在预测精度上与所有三个基线相当，同时在数据点数量上线性扩展，而精确核矩阵方法变得不可行，诱导点基线仍然明显更慢。
cs.LG / 70 / 2608.11937
Distillation of Foundation Models for Time-dependent PDEs
面向时间相关偏微分方程的基础模型蒸馏
Abstract
Foundation models for time-dependent partial differential equations (PDEs) are trained on large and diverse collections of physical systems and can generalize effectively to new downstream tasks. After fine-tuning on only a few trajectories from a target domain, they can achieve strong accuracy in low-data regimes. However, these models are typically large and computationally intensive, limiting their usefulness as fast surrogates for numerical solvers. We propose Teacher Rollout Extension (TREX), a knowledge distillation framework that transfers the predictive capability of a pretrained foundation model into a compact and efficient student. Starting from a fine-tuned teacher, TREX augments limited downstream data by generating long synthetic trajectories through teacher rollouts, optionally with periodic noise injection. This procedure samples from the teacher-induced rollout distribution without requiring explicit knowledge of the initial-condition distribution, while exposing the student to long-horizon states and local recovery behavior around states encountered during autoregressive prediction. The student can further incorporate task-specific inductive biases, such as equivariance, that the teacher does not necessarily enforce. We evaluate TREX on multiple PDE benchmarks. The resulting students can match or surpass the teacher's accuracy while reducing the number of parameters by several orders of magnitude and achieving more than an order-of-magnitude speedup in inference.
Chinese Translation
面向时间相关偏微分方程（PDE）的基础模型在大规模且多样化的物理系统集合上训练，并能够有效泛化到新的下游任务。仅在目标域的少量轨迹上进行微调后，它们能够在低数据情形下达到很强的精度。然而，这些模型通常规模庞大且计算密集，限制了它们作为数值求解器快速替代模型的实用性。我们提出了教师推演扩展（TREX），一种知识蒸馏框架，将预训练基础模型的预测能力迁移到一个紧凑且高效的学生模型中。从微调后的教师模型出发，TREX 通过教师推演生成长合成轨迹，并可选地加入周期性噪声注入，从而增广有限的下游数据。该过程从教师诱导的推演分布中采样，而无需显式了解初始条件分布，同时使学生模型接触到长程状态以及在自回归预测过程中遇到的状态周围的局部恢复行为。学生模型还可以进一步融入任务特定的归纳偏置，例如等变性，而教师模型未必强制施加这些偏置。我们在多个 PDE 基准上评估了 TREX。所得到的学生模型能够达到或超过教师模型的精度，同时将参数数量减少数个数量级，并在推理中实现超过一个数量级的加速。
cs.LG / 71 / 2608.11951
TailBooster: A Dual-Layer Generative Framework for Extreme Value Augmentation with Operational Validity Enforcement
TailBooster：一种用于极值增强且具备运行有效性强制机制的双层生成框架
Abstract
Extreme events in air transport, such as severe arrival delays and abnormal air times, cause cascading network disruptions with substantial operational, economic, and safety costs. Such events are rare in historical records, leaving insufficient training signal for machine learning models. Synthetic data augmentation offers a principled solution, but conventional generative models under-represent distributional tails and give no guarantee against operationally infeasible instances, such as a short air time paired with a long flight distance. No existing approach addresses both limitations for mixed-type tabular records. We propose TailBooster, a dual-layer generative framework combining generative modelling with two anomaly detection layers. A statistical layer extracts extremes via the interquartile range, supplying tail-concentrated training signal to dedicated generative models, here a Tabular Variational Autoencoder. A deep learning layer then applies autoencoder-based cleaning, discarding synthetic records that violate the operational envelope learned from historical data. The framework was evaluated on US flight records across five dimensions: diversity, statistical similarity, fidelity, operational validity, and utility, the latter two being the primary improvement targets. Data-driven cleaning markedly improved operational validity, while targeted augmentation enhanced utility for extreme-event prediction. Across six regression algorithms, training on the framework's records reduced Mean Absolute Error by 47-49% on extreme air time and 29-57% on extreme arrival delay prediction relative to conventional synthetic data, with comparable gains when real records were enriched with synthetic extremes. Being fully data-driven and model-agnostic, TailBooster extends to domains where extreme-event prediction is critical and domain-specific rules are unavailable.
Chinese Translation
航空运输中的极端事件，如严重到达延误和异常飞行时间，会造成级联式网络中断，并带来巨大的运行、经济和安全成本。此类事件在历史记录中很少见，导致机器学习模型的训练信号不足。合成数据增强提供了一种有原则的解决方案，但传统生成模型对分布尾部的表示不足，并且无法保证不会生成运行上不可行的实例，例如将较短的飞行时间与较长的飞行距离配对。现有方法均无法针对混合类型表格记录同时解决这两个局限性。我们提出TailBooster，这是一个双层生成框架，将生成建模与两个异常检测层相结合。统计层通过四分位距提取极值，为专用生成模型（此处为表格变分自编码器）提供集中于尾部的训练信号。然后，深度学习层应用基于自编码器的清洗，丢弃违反从历史数据中学习到的运行包线的合成记录。该框架在美国航班记录上从五个维度进行了评估：多样性、统计相似性、保真度、运行有效性和实用性，其中后两者是主要改进目标。数据驱动的清洗显著提高了运行有效性，而针对性增强则提升了极端事件预测的实用性。在六种回归算法中，与传统合成数据相比，在该框架生成的记录上训练将极端飞行时间预测的平均绝对误差降低了47%–49%，将极端到达延误预测的平均绝对误差降低了29%–57%；当真实记录用合成极值进行扩充时，也获得了相当的提升。TailBooster完全由数据驱动且与模型无关，可扩展到极端事件预测至关重要且缺乏领域特定规则的领域。
cs.LG / 72 / 2608.11970
TESLA: Taylor Expansion of Sinusoidal Learnable Activations
TESLA：正弦可学习激活的泰勒展开
Abstract
The parity problem--deciding whether the number of ones in a binary vector is odd or even--remains challenging for standard neural networks due to linear inseparability and the need for global interactions. We propose TESLA, an activation defined as a learnable combination of sine and cosine terms, enabling explicit control over polynomial degree and selective amplification of high-order components. Theoretically, we show that constraining TESLA's coefficients yields Lipschitz/Rademacher complexity bounds and shapes the training dynamics to emphasize higher-frequency structure. Empirically, on parity with input length n = 32, TESLA attains strong generalization with 100K training samples (approximately 0.002% of the 2^32 input space) and remains robust under heavy corruption, retaining high accuracy with up to 30% label noise. We also compare against periodic and frequency-based baselines (SIREN, SNAKE, and Fourier feature embeddings) on parity and Forrelation. Beyond synthetic structure, TESLA delivers comparable performance on ImageNet-100, indicating that activation-level degree control transfers to more general vision workloads. Code: https://github.com/KAU-QuantumAILab/TESLA
Chinese Translation
奇偶性问题——判断一个二进制向量中 1 的个数是奇数还是偶数——由于线性不可分以及需要全局交互，对标准神经网络仍然具有挑战性。我们提出 TESLA，一种定义为正弦项和余弦项的可学习组合的激活函数，能够显式控制多项式阶数并选择性放大高阶分量。理论上，我们证明约束 TESLA 的系数可以得到 Lipschitz/Rademacher 复杂度界，并塑造训练动态以强调高频结构。在经验上，在输入长度 n = 32 的奇偶性问题上，TESLA 在 10 万个训练样本（约为 2^32 输入空间的 0.002%）下获得了强泛化能力，并且在严重损坏下保持稳健，在高达 30% 的标签噪声下仍保持高准确率。我们还在奇偶性和 Forrelation 问题上与周期性和基于频率的基线（SIREN、SNAKE 和傅里叶特征嵌入）进行了比较。除了合成结构之外，TESLA 在 ImageNet-100 上提供了可比的性能，表明激活级别的阶数控制能够迁移到更一般的视觉工作负载。代码：https://github.com/KAU-QuantumAILab/TESLA
cs.LG / 73 / 2608.12001
Remote Sensing and Machine Learning-Based Analysis of Land Use and Vegetation Change in Dhaka District, Bangladesh
基于遥感与机器学习的孟加拉国达卡区土地利用与植被变化分析
Abstract
Rapid urbanization in Dhaka District, Bangladesh has triggered substantial alterations in land use and environmental conditions, necessitating systematic monitoring for informed urban planning and ecological sustainability. This study employs remote sensing data and machine learning techniques to analyze spatiotemporal changes in land cover and vegetation dynamics between 2019 and 2024. High-resolution satellite imagery from Sentinel-2 MSI and Landsat 8 was utilized to classify land cover types and compute spectral indices including the Normalized Difference Vegetation Index (NDVI), Normalized Difference Built-up Index (NDBI), and Normalized Difference Water Index (NDWI). A supervised machine learning approach incorporating Decision Tree, K-Nearest Neighbors (KNN), and Random Forest classifiers was applied using labeled geospatial training points within Google Earth Engine. Accuracy assessments were conducted using confusion matrices and kappa statistics. Results indicate a 59.5% increase in urban built-up areas and a significant decline in vegetation (-8.46%) and water bodies (-7.77%) over the five-year period. Land conversion from vegetated and aquatic areas to urban infrastructure was identified as a dominant trend. Among the models, Random Forest demonstrated the highest classification accuracy. These findings underscore the growing environmental pressures driven by unregulated urban expansion in Dhaka. The study highlights the potential of remote sensing and machine learning tools in providing timely, actionable data to support sustainable urban development, land-use regulation, and ecosystem conservation policies.
Chinese Translation
孟加拉国达卡区的快速城市化已引发土地利用和环境状况的重大变化，因此需要进行系统监测，以支持明智的城市规划和生态可持续性。本研究利用遥感数据和机器学习技术，分析2019年至2024年间土地覆盖和植被动态的时空变化。利用来自Sentinel-2 MSI和Landsat 8的高分辨率卫星影像对土地覆盖类型进行分类，并计算光谱指数，包括归一化差值植被指数（NDVI）、归一化差值建筑指数（NDBI）和归一化差值水体指数（NDWI）。在Google Earth Engine中，利用带标签的地理空间训练点，应用了一种包含决策树、K近邻（KNN）和随机森林分类器的监督机器学习方法。使用混淆矩阵和kappa统计量进行精度评估。结果表明，在这五年期间，城市建成区增加了59.5%，植被（-8.46%）和水体（-7.77%）显著减少。从植被和水域向城市基础设施的土地转变被确定为主要趋势。在各模型中，随机森林表现出最高的分类精度。这些发现凸显了达卡不受监管的城市扩张所带来的日益增大的环境压力。本研究强调了遥感与机器学习工具在提供及时、可操作数据以支持可持续城市发展、土地利用监管和生态系统保护政策方面的潜力。
cs.LG / 74 / 2608.12007
Dual-Model Sentiment Analysis of Consumer Reviews in the Retail Coffee Sector Using Machine Learning and Deep Learning Approaches
使用机器学习与深度学习方法的零售咖啡行业消费者评论双模型情感分析
Abstract
Consumer reviews play an important role in shaping brand perception and business strategies, particularly in service-driven industries such as retail coffee. This study presents a comparative sentiment analysis framework for Starbucks customer reviews using classical machine learning and deep learning approaches. The dataset, collected from ConsumerAffairs, contains more than 700 reviews and was analyzed through preprocessing and exploratory data analysis to identify temporal and geographic patterns. Sentiment labels were generated by binarizing star ratings, with ratings of 4 and 5 classified as positive and ratings of 1 to 3 as negative. The resulting dataset was substantially imbalanced toward negative sentiment. Five machine learning classifiers, including Logistic Regression, Support Vector Machine (SVM), Decision Tree, Random Forest, and Naive Bayes, were evaluated alongside five deep learning models: LSTM, RNN, Bidirectional LSTM, GRU, and CNN. Model performance was assessed using accuracy, precision, recall, and F1-score. SVM achieved the highest accuracy among the machine learning models at 91.0 percent, while Bidirectional LSTM showed the strongest performance among the deep learning models and demonstrated good generalization on unseen data. The findings also show that class imbalance negatively affected positive sentiment recall across several models. Overall, this study provides a comparative evaluation of machine learning and deep learning approaches for real-world consumer sentiment analysis and highlights the importance of appropriate model selection and preprocessing for customer experience analytics in the retail coffee sector.
Chinese Translation
消费者评论在塑造品牌认知和商业策略方面发挥着重要作用，尤其是在零售咖啡等服务驱动型行业中。本研究提出了一个针对星巴克顾客评论的比较情感分析框架，使用了经典机器学习和深度学习方法。数据集收集自 ConsumerAffairs，包含 700 多条评论，并通过预处理和探索性数据分析加以分析，以识别时间和地理模式。情感标签通过将星级评分二值化生成，4 星和 5 星被归类为正面，1 星至 3 星被归类为负面。所得数据集严重偏向负面情感。对五种机器学习分类器（包括逻辑回归、支持向量机（SVM）、决策树、随机森林和朴素贝叶斯）以及五种深度学习模型（LSTM、RNN、双向 LSTM、GRU 和 CNN）进行了评估。模型性能使用准确率、精确率、召回率和 F1 分数进行评估。SVM 在机器学习模型中取得了最高准确率，为 91.0%，而双向 LSTM 在深度学习模型中表现最强，并在未见数据上表现出良好的泛化能力。研究结果还表明，类别不平衡对多个模型的正面情感召回率产生了负面影响。总体而言，本研究对用于真实世界消费者情感分析的机器学习和深度学习方法进行了比较评估，并强调了在零售咖啡行业的顾客体验分析中适当选择模型和预处理的重要性。
cs.LG / 75 / 2608.12010
Reducing Symmetry Increase in Equivariant Neural Networks
减少等变神经网络中的对称性增加
Abstract
Equivariant Neural Networks (ENNs) have empowered numerous applications in scientific fields. Despite their remarkable capacity for representing geometric structures, ENNs suffer from degraded expressivity when processing symmetric inputs: the output representations are invariant to transformations that extend beyond the input's symmetries. The mathematical essence of this phenomenon is that a symmetric input, after being processed by an equivariant map, experiences an increase in symmetry. While prior research has documented symmetry increase in specific cases, a rigorous understanding of its underlying causes and general reduction strategies remains lacking. In this paper, we provide a detailed and in-depth characterization of symmetry increase together with a principled framework for its reduction: (i) For any given feature space and input symmetry group, we prove that the increased symmetry admits an infimum determined by the structure of the feature space; (ii) Building on this foundation, we develop a computable algorithm to derive this infimum, and propose practical guidelines for feature design to prevent harmful symmetry increases. (iii) Under standard regularity assumptions, we demonstrate that for most equivariant maps, our guidelines effectively reduce symmetry increase. To complement our theoretical findings, we provide visualizations and experiments on both synthetic datasets and the real-world QM9 dataset. The results validate our theoretical predictions.
Chinese Translation
等变神经网络（ENNs）已在科学领域的众多应用中发挥了重要作用。尽管它们在表示几何结构方面具有卓越能力，但在处理对称输入时，ENNs 的表达能力会下降：输出表示会对超出输入对称性的变换保持不变。这一现象的数学本质是，对称输入在经过等变映射处理后，其对称性会增加。尽管已有研究记录了特定情况下的对称性增加，但对其根本原因的严格理解以及通用的减少策略仍然缺乏。在本文中，我们对对称性增加进行了详细而深入的表征，并提出了一个用于减少对称性增加的原则性框架：(i) 对于任意给定的特征空间和输入对称群，我们证明了增加的对称性存在一个由特征空间结构决定的下确界；(ii) 在此基础上，我们开发了一种可计算算法来推导该下确界，并提出了用于特征设计的实用指南，以防止有害的对称性增加。(iii) 在标准正则性假设下，我们证明对于大多数等变映射，我们的指南能够有效减少对称性增加。为了补充我们的理论发现，我们在合成数据集和真实世界 QM9 数据集上提供了可视化和实验。结果验证了我们的理论预测。
cs.LG / 76 / 2608.12026
SoftWater: Class-Aware Rate Allocation for Softmax Quantization
SoftWater：面向 Softmax 量化的类别感知速率分配
Abstract
Post-training quantization pipelines routinely leave the softmax output layer in high precision. Yet in small LLMs with modern vocabularies, the head holds 15--30\% of all parameters, so a nominal ``2-bit'' model with an fp16 head can store several times as many bits per weight. We pose softmax-layer quantization as a rate-distortion problem under the KL divergence between the original and quantized output distributions. A second-order analysis reveals a class-aware geometry: quantization error is weighted jointly by feature covariance and class-specific softmax curvature. A separability approximation replaces the $Kn\times Kn$ Cholesky with one $n\times n$ factorization rescaled per class, making the lattice encodable by successive interference cancellation, with both statistics from a single forward pass. The resulting method, SoftWater, gives fine grids to frequent, low-variance classes and coarse grids to rare ones, a large gap under Zipfian token distributions. Across five models from 1B to 32B, SoftWater outperforms the released WaterSIC quantizer (near-optimal under linear-layer WMSE but not output KL) at matched head rates on 59 of 60 test points, using none of that pipeline's refinements and cutting head-induced KL by $6.5\times$--$8.3\times$ at 2 bits. On Llama-3.2-1B-Instruct with quantized bodies, a 2-bit head removes 45--60\% of stored bytes for a $2.9$--$3.7\%$ perplexity increase. Because the class-side statistic comes from calibration data, matching calibration to the deployment domain gives the lowest KL on that domain throughout. On a tied model, a 4-bit head is near-lossless and a 2-bit head costs under 4\% perplexity, making head quantization of such models practical.
Chinese Translation
训练后量化流程通常将 softmax 输出层保留在高精度。然而，在具有现代词表的小型 LLM 中，输出头占全部参数的 15--30%，因此一个带有 fp16 输出头的名义“2-bit”模型每权重存储的比特数可能是数倍。我们将 softmax 层量化建模为原始输出分布与量化输出分布之间 KL 散度下的率失真问题。二阶分析揭示了一种类别感知的几何结构：量化误差由特征协方差和类别特定的 softmax 曲率共同加权。一种可分离近似将 $Kn\times Kn$ 的 Cholesky 分解替换为每个类别重新缩放的一次 $n\times n$ 分解，从而使格可通过逐次干扰消除进行编码，且两种统计量均来自一次前向传播。由此得到的方法 SoftWater 为频繁且低方差的类别分配细网格，为稀有类别分配粗网格，这在 Zipf 型 token 分布下形成巨大差距。在从 1B 到 32B 的五个模型上，在匹配的输出头速率下，SoftWater 在 60 个测试点中的 59 个上优于已发布的 WaterSIC 量化器（该量化器在线性层 WMSE 下接近最优，但在输出 KL 下并非如此），且未使用该流程中的任何改进措施，并在 2 比特下将输出头引起的 KL 降低了 $6.5\times$--$8.3\times$。在主体已量化的 Llama-3.2-1B-Instruct 上，2 比特输出头可减少 45--60% 的存储字节，而困惑度增加 $2.9$--$3.7\%$。由于类别侧统计量来自校准数据，使校准数据与部署域匹配可在该域上始终获得最低 KL。在权重绑定模型上，4 比特输出头接近无损，2 比特输出头带来的困惑度增加不到 4%，这使得此类模型的输出头量化切实可行。
cs.LG / 77 / 2608.12027
Uncertainty-Aware Probabilistic Constrained Clustering from Entangled Pairwise Supervision
基于纠缠成对监督的不确定性感知概率约束聚类
Abstract
Pairwise constrained clustering typically relies on hard must-link/cannot-link labels, whereas realistic pairwise supervision may be real-valued and entangle intrinsic ambiguity, expert judgment, and stochastic corruption. Existing deep constrained clustering (DCC) methods mainly target hard, expert-agnostic constraints, treating soft labels mostly numerically rather than semantically. We formalize this setting as uncertainty-aware probabilistic constrained clustering (UPCC), defining a canonical aleatoric target through a heterogeneous observation process and analyzing its conditional identifiability. We introduce ProbPair, an angular pairwise objective for probabilistic relations, and build ECI-PP, an estimator--corrector--integrator framework that refines imperfect supervision via belief estimation, correction, and reliability-aware integration. Across challenging probabilistic supervision settings, experiments on diverse benchmarks show that ECI-PP outperforms state-of-the-art DCC methods and remains robust with a shared default configuration.
Chinese Translation
成对约束聚类通常依赖硬性的必须链接/不可链接标签，而现实中的成对监督可能是实值的，并交织着固有歧义、专家判断和随机损坏。现有深度约束聚类（DCC）方法主要针对硬性的、与专家无关的约束，其对软标签的处理大多是数值性的而非语义性的。我们将该设定形式化为不确定性感知概率约束聚类（UPCC），通过异构观测过程定义了一个规范化的偶然性目标，并分析了其条件可辨识性。我们引入了 ProbPair，一种用于概率关系的角度成对目标函数，并构建了 ECI-PP，一个估计器—校正器—集成器框架，该框架通过信念估计、校正和可靠性感知集成来细化不完美的监督。在具有挑战性的概率监督设定下，多个不同基准上的实验表明，ECI-PP 优于最先进的 DCC 方法，并在共享默认配置下仍保持鲁棒。
cs.LG / 78 / 2608.12037
Clustered Randomized Smoothing for Stochastic Prediction Functions
Abstract
Modern stochastic predictors can model rich, multi-modal outcome distributions. However, this expressive power comes with challenges in ensuring robust predictions $-$ a critical requirement in safety-critical domains. Randomized smoothing is a leading technique for improving robustness, particularly against adversarial perturbations. Yet, in stochastic multi-modal regression settings, randomized smoothing often fails due to mode collapse, yielding averaged predictions that do not reflect the underlying distribution. To address this limitation, we propose clustered $α$-smoothing, a framework that (1) partitions noisy samples using an arbitrary clustering algorithm, (2) applies $α$-smoothing locally within each cluster, and (3) combines the resulting predictions into a mixture distribution. By interpreting the smoothing distribution as a mixture of $α$-smoothers, we derive a lower bound on the probability that the smoothed prediction lies within a union of compact regions corresponding to distinct modes. We empirically evaluate our framework on two benchmarks, demonstrating substantial improvements over state-of-the-art methods. In stochastic trajectory prediction on a driving simulator dataset, our approach achieves, on average, a $27\%$ lower Wasserstein distance to the ground-truth distribution compared to $α$-smoothing. In quadrotor control, where modes correspond to distinct feasible paths to a target, our method reduces the collision rate by $81\%$ relative to the state-of-the-art randomized smoothing.
cs.LG / 79 / 2608.12057
Towards Truly Unsupervised Evaluation of Feature Selection
迈向真正无监督的特征选择评估
Abstract
Feature selection is one of the most important and fundamental tasks in data mining, tackled by a family of methods with an established set of evaluation techniques to measure the quality of a specific method. Most of the methods commonly used for the unsupervised evaluation of feature selection algorithms suffer from critical design flaws which question their unsupervised nature. In this paper, we provide a critical discussion on the established allegedly unsupervised evaluation techniques, and shed light on the reasons why they are not truly unsupervised but, at best, supervised evaluation under an unsupervised downstream task. We also propose a novel, truly unsupervised evaluation framework to measure the quality of the feature selection algorithms without any form of information about the labels. The proposed framework utilizes unsupervised Principal Component Analysis, and optimal transport to measure the quality of the feature selection methods in a truly unsupervised manner.
Chinese Translation
特征选择是数据挖掘中最重要和基础的任务之一，由一系列方法处理，并配有一套既定的评估技术来衡量特定方法的质量。大多数常用于无监督评估特征选择算法的方法存在严重的设计缺陷，这些缺陷令人质疑其无监督性质。在本文中，我们对既有的、据称是无监督的评估技术进行了批判性讨论，并阐明了它们为何并非真正无监督，而充其量只是在无监督下游任务下的有监督评估。我们还提出了一种新颖的、真正无监督的评估框架，用于在没有任何标签信息的情况下衡量特征选择算法的质量。所提出的框架利用无监督主成分分析和最优传输，以真正无监督的方式衡量特征选择方法的质量。
cs.LG / 80 / 2608.12084
NAE: Normalizing AutoEncoder
NAE：归一化自编码器
Abstract
We consider the setting of Normalizing flows with approximate inverses, an established paradigm spanning both full-dimensional ($d=D$) and bottleneck ($d<D$) settings, and group these models under the term flow autoencoders. We present a theoretical investigation into their training dynamics and prove that the proposed loss used by existing approaches is suboptimal; specifically, both encoder and decoder surrogates must be optimized in alignment with reconstruction loss. Guided by these insights, we propose Normalizing Autoencoder (NAE), which employs a novel conditional loss that aligns the surrogate loss gradient with that of reconstruction loss, directly improving upon the current standard. Extensive experiments across molecule generation, tabular data, and image benchmarks demonstrate that NAE achieves state of the art performance. Our work highlights the importance of loss alignment in flow autoencoders and establishes NAE as a powerful generative framework.
Chinese Translation
我们考虑具有近似逆的归一化流设定，这是一种涵盖全维（$d=D$）和瓶颈（$d<D$）设定的既有范式，并将这些模型统称为流自编码器。我们对其训练动态进行了理论研究，并证明现有方法所使用的损失是次优的；具体而言，编码器和解码器代理都必须与重构损失对齐地进行优化。在这些见解的指导下，我们提出了归一化自编码器（NAE），它采用一种新的条件损失，使代理损失梯度与重构损失梯度对齐，直接改进了当前标准。在分子生成、表格数据和图像基准上的大量实验表明，NAE 实现了最先进的性能。我们的工作突显了流自编码器中损失对齐的重要性，并将 NAE 确立为一个强大的生成框架。
cs.LG / 81 / 2608.12090
Task- and dataset-specific information in protein language models
蛋白质语言模型中的任务和数据集特定信息
Abstract
Protein language models (PLMs) have transferred the latest advances from natural language processing to computational biology. These models, trained on large corpora of protein sequence data, are widely used to translate amino acid sequences into latent-space embeddings, ready for use in diverse downstream tasks (DTs). By a common consensus, embeddings from the model's last layer are used, and the model's internal behavior remains poorly understood. We analyzed 13 PLMs across 15 DTs from 11 datasets to investigate the informativeness of embeddings created in intermediate PLM layers. We trained probe models on embeddings from each layer, compared their performance, and computed characteristics of the latent spaces they span to estimate the information they contain, and found that the last layers of PLMs rarely contained embeddings that led to the best results on downstream tasks. Furthermore, we identified a connection between DTs and the distribution across PLMs' layers of the relevant information to predict that task. For example, similarity between the pre-training objective and the objective of predicting properties of individual residues leads to a steady increase in understanding of such tasks across the layers of PLMs. On the other hand, for whole-protein tasks, we observe that the dataset, rather than the task itself, defines PLMs' ability to perform well on a DT. Embeddings from shallow layers of PLMs perform better for datasets that contain deep mutational scan (DMS) data, while datasets containing diverse natural proteins find most useful embeddings in the models' deeper layers. Additionally, we discover that the performance of PLMs drops significantly when tasks are introduced for artificial proteins.
Chinese Translation
蛋白质语言模型（PLMs）已将自然语言处理的最新进展迁移到计算生物学。这些模型在蛋白质序列数据的大型语料库上训练，被广泛用于将氨基酸序列转换为潜在空间嵌入，以便用于各种下游任务（DTs）。根据普遍共识，使用的是模型最后一层的嵌入，而模型的内部行为仍未被充分理解。我们分析了来自11个数据集的15个下游任务中的13个蛋白质语言模型，以研究在蛋白质语言模型中间层生成的嵌入的信息量。我们在每一层的嵌入上训练探针模型，比较其性能，并计算它们所张成的潜在空间的特征以估计其所含信息，结果发现蛋白质语言模型的最后几层很少包含能在下游任务上带来最佳结果的嵌入。此外，我们发现下游任务与预测该任务所需相关信息在蛋白质语言模型各层中的分布之间存在联系。例如，预训练目标与预测单个残基性质的目标之间的相似性，导致随着蛋白质语言模型层数的增加，对此类任务的理解稳步提升。另一方面，对于全蛋白任务，我们观察到，决定蛋白质语言模型在下游任务上表现良好能力的，是数据集而非任务本身。对于包含深度突变扫描（DMS）数据的数据集，蛋白质语言模型浅层嵌入表现更好；而包含多样天然蛋白质的数据集则在模型更深层中找到最有用的嵌入。此外，我们发现，当针对人工蛋白质引入任务时，蛋白质语言模型的性能会显著下降。
cs.LG / 82 / 2608.12100
Confidence Calibration of Deep Learning Systems
深度学习系统的置信度校准
Abstract
In high-stakes applications, reliable confidence estimates are as important as the predictions themselves. Confidence calibration ensures that predicted probabilities reflect the likelihood of correctness, making it essential for safe deployment of deep learning models. However, existing methods typically assume access to clean validation data, which is often unrealistic due to label noise and domain shifts. This thesis develops methods for improving calibration under these conditions. First, we address calibration under label noise. Standard methods can produce misleading confidence estimates when labels are unreliable. We propose a framework that uses an estimated noise model to reconstruct noise-free confidence estimates by modeling the relationship between noisy and clean label distributions. We extend this approach to Conformal Prediction (CP), which provides set-valued predictions with guaranteed coverage. Our noise-aware CP method estimates clean conformity scores despite label noise, enabling reliable uncertainty quantification. Next, we study calibration in unsupervised domain adaptation, where a model trained on a labeled source domain is adapted to an unlabeled target domain. Since labeled target data are unavailable, we estimate target-domain accuracy from source performance and domain discrepancies, enabling calibration without target labels. We also consider privacy-preserving settings in which user labels and model outputs must remain protected. We propose a locally differentially private conformal prediction framework that provides valid uncertainty quantification while maintaining privacy guarantees and balancing privacy, computational feasibility, and prediction reliability. Our results bridge calibration theory and practical deployment in safety-critical applications, contributing to reliable, privacy-preserving, and noise-resilient neural network predictions.
Chinese Translation
在高风险应用中，可靠的置信度估计与预测本身同样重要。置信度校准确保预测概率反映正确性的可能性，使其对于深度学习模型的安全部署至关重要。然而，现有方法通常假设能够获得干净的验证数据，由于标签噪声和领域偏移，这往往是不现实的。本论文开发了在这些条件下改进校准的方法。首先，我们处理标签噪声下的校准问题。当标签不可靠时，标准方法可能产生误导性的置信度估计。我们提出了一个框架，利用估计的噪声模型，通过建模带噪标签分布与干净标签分布之间的关系来重建无噪声的置信度估计。我们将该方法扩展到共形预测（CP），它提供具有覆盖保证的集合值预测。我们的噪声感知 CP 方法在存在标签噪声的情况下估计干净的一致性分数，从而实现可靠的不确定性量化。接下来，我们研究无监督领域自适应中的校准问题，其中在带标签源域上训练的模型被适配到无标签目标域。由于无法获得带标签的目标数据，我们根据源域性能和领域差异估计目标域准确率，从而在无需目标标签的情况下实现校准。我们还考虑了用户标签和模型输出必须受到保护的隐私保护场景。我们提出了一个本地差分隐私共形预测框架，该框架在保持隐私保证的同时提供有效的不确定性量化，并平衡隐私、计算可行性和预测可靠性。我们的结果架起了校准理论与安全关键应用中实际部署之间的桥梁，为实现可靠、隐私保护且抗噪的神经网络预测做出了贡献。
cs.LG / 83 / 2608.12108
Beyond Parameter Space: NTK-Guided Personalized Aggregation for Robust Federated Learning
超越参数空间：NTK引导的个性化聚合用于鲁棒联邦学习
Abstract
Federated learning (FL) enables collaborative model training across distributed clients while keeping data local. A central challenge is determining which client updates are beneficial for aggregation with respect to each client's target domain. Existing methods typically address this problem in parameter space by comparing model parameters or gradients. However, parameter-space similarity can be a poor proxy for predictive behavior, especially under heterogeneous, non-IID data. Consequently, updates that are misaligned with a client's target domain, including those caused by heterogeneous data or malfunctioning clients, may degrade local model performance. We propose Local Inference Guided Aggregation for Heterogeneous Training Environments to Yield Enhancement Through Agreement and Regularization (LIGHTYEAR), a federated learning framework that performs update selection in function space. LIGHTYEAR uses an NTK-based agreement score to characterize predictive behavior and determine a personalized aggregation set for each client. By relating model parameters to local predictive responses, the Neural Tangent Kernel (NTK) provides a more expressive criterion for update selection than parameter-space similarity alone. Because function-space information is not available before aggregation in conventional centralized FL, LIGHTYEAR uses a peer-to-peer (P2P) topology in which clients exchange updates directly and evaluate incoming models on private validation data. Each client selects only updates that are beneficial for its own target domain and aggregates them using a regularized rule that improves stability under heterogeneity. Across five datasets and nine baseline methods, LIGHTYEAR consistently outperforms centralized FL baselines and existing P2P approaches.
Chinese Translation
联邦学习（FL）能够在数据保留在本地的同时，跨分布式客户端进行协作模型训练。一个核心挑战在于，针对每个客户端的目标域，确定哪些客户端更新对聚合是有益的。现有方法通常通过比较模型参数或梯度，在参数空间中解决这一问题。然而，参数空间中的相似性可能无法很好地代表预测行为，尤其是在异构、非独立同分布数据下。因此，与客户端目标域不一致的更新——包括由异构数据或故障客户端引起的更新——可能会降低本地模型性能。我们提出 LIGHTYEAR（面向异构训练环境的局部推理引导聚合，通过一致性与正则化实现增强），这是一个在函数空间中执行更新选择的联邦学习框架。LIGHTYEAR 使用基于 NTK 的一致性分数来刻画预测行为，并为每个客户端确定个性化聚合集。通过将模型参数与局部预测响应相关联，神经正切核（NTK）为更新选择提供了比仅使用参数空间相似性更具表达力的准则。由于在传统集中式联邦学习中，函数空间信息在聚合前无法获得，LIGHTYEAR 采用点对点（P2P）拓扑，客户端直接交换更新并在私有验证数据上评估传入模型。每个客户端仅选择对自身目标域有益的更新，并使用一种在异构条件下提高稳定性的正则化规则对其进行聚合。在五个数据集和九种基线方法上，LIGHTYEAR 持续优于集中式联邦学习基线和现有的点对点方法。
cs.LG / 84 / 2608.12117
Attractor Image-Based Deep Learning of Arterial Pulse Waves for Age Classification
基于吸引子图像的动脉脉搏波深度学习用于年龄分类
Abstract
Arterial pulse waveform morphology evolves with age, reflecting structural and functional changes in the cardiovascular system. Thus, vascular age is a valuable surrogate marker of cardiovascular health, and premature vascular ageing can indicate increased disease risk. Pulse wave analysis could support risk stratification in otherwise asymptomatic adults. We transformed pulse wave time-series data from photoplethysmography (PPG) and arterial tonometry into images, using the Symmetric Projection Attractor Reconstruction (SPAR) method. These SPAR images were used to train a convolutional neural network to classify healthy subjects into two closely spaced age groups (35-40 and 50-55 years). The model demonstrated consistent classification performance across internal and external test sets, achieving F1 scores above 70% for both PPG and tonometry signals. These results suggest that SPAR-derived pulse wave images contain discriminative morphological features even among healthy adults close in age. This proof-of-concept lays the groundwork for future research into the use of SPAR for early risk detection using smart wearables.
Chinese Translation
动脉脉搏波形形态随年龄演变，反映心血管系统结构和功能的变化。因此，血管年龄是心血管健康的一个有价值的替代标志物，过早的血管老化可提示疾病风险增加。脉搏波分析可支持在其他方面无症状的成人中进行风险分层。我们使用对称投影吸引子重构（SPAR）方法，将来自光电容积描记（PPG）和动脉张力测量的脉搏波时间序列数据转换为图像。这些SPAR图像被用于训练卷积神经网络，以将健康受试者分为两个年龄间隔较近的组（35–40岁和50–55岁）。该模型在内部和外部测试集上表现出一致的分类性能，对PPG和张力测量信号均取得了高于70%的F1分数。这些结果表明，即使是在年龄相近的健康成年人中，SPAR衍生的脉搏波图像也包含具有判别力的形态特征。这一概念验证为未来研究使用SPAR通过智能可穿戴设备进行早期风险检测奠定了基础。
cs.LG / 85 / 2608.12134
Adversarial Resilience of Poisson-Process Submodular Maximization over Matroids: From Robust Offline Optimization to Full-Bandit Learning
拟阵上泊松过程子模块最大化的对抗韧性：从鲁棒离线优化到全强盗学习
Abstract
We study nonnegative submodular maximization subject to a general matroid when the offline algorithm is given an arbitrary controlled value oracle. Our main result is an adversarial resilience theorem for the Spiteful Greedy Swap Poisson Process (SGS-Poisson): without modifying its Poisson intensity, single-element exchange rule, or spiteful drop step, the algorithm retains limiting approximation factors $1/e$ for non-monotone objectives and $1-1/e$ for monotone objectives. More precisely, under every controlled oracle $\widehat f$ satisfying $|\widehat f(S)-f(S)|\le ξ$ for every set $S$, our implementation returns a feasible set with expected value at least $(1/e-\varepsilon)\OPT-O(kξ)$ and $(1-1/e-\varepsilon)\OPT-O(kξ)$, respectively, using $\widetilde O(nk^2\varepsilon^{-2})$ oracle calls. As a consequence, the offline-to-online reduction yields full-bandit CMAB algorithms for general matroid-constrained submodular rewards with exact limiting approximation-regret factors $1/e$ and $1-1/e$ and $\widetilde O(n^{1/5}k^{4/5}T^{4/5})$ regret.
Chinese Translation
我们研究受一般拟阵约束的非负子模块最大化，其中离线算法被给定任意受控值预言机。我们的主要结果是针对恶意贪婪交换泊松过程（SGS-Poisson）的对抗韧性定理：在不修改其泊松强度、单元素交换规则或恶意丢弃步骤的情况下，该算法对非单调目标保持极限近似因子 $1/e$，对单调目标保持 $1-1/e$。更准确地说，在每一个满足对每个集合 $S$ 都有 $|\widehat f(S)-f(S)|\le ξ$ 的受控预言机 $\widehat f$ 下，我们的实现返回一个可行集，其期望值分别至少为 $(1/e-\varepsilon)\OPT-O(kξ)$ 和 $(1-1/e-\varepsilon)\OPT-O(kξ)$，并且使用 $\widetilde O(nk^2\varepsilon^{-2})$ 次预言机调用。因此，离线到在线的归约产生了针对一般拟阵约束子模块奖励的全强盗 CMAB 算法，其具有精确的极限近似-遗憾因子 $1/e$ 和 $1-1/e$，以及 $\widetilde O(n^{1/5}k^{4/5}T^{4/5})$ 的遗憾。
cs.LG / 86 / 2608.12194
HYDRA: Hyperbolic Dynamic Representation Architecture for Kolmogorov-Arnold Networks
HYDRA：用于 Kolmogorov-Arnold 网络的双曲动态表示架构
Abstract
Kolmogorov-Arnold Networks (KANs) enhance nonlinear function approximation by replacing scalar weights with learnable univariate functions. However, assigning an independent function to every connection results in substantial parameter redundancy, limiting their scalability and efficiency. To reduce this redundancy, we introduce \textbf{HY}perbolic \textbf{D}ynamic \textbf{R}epresentation \textbf{A}rchitecture (HYDRA), a parameter-efficient hyperbolic extension of KAN that combines spline-based functional learning with representations in the Poincaré ball. HYDRA maps vector-valued inputs into a bounded hyperbolic latent space, performs KAN-style updates in tangent space, and employs a low-rank prototype block to share functional transformations across hidden dimensions. The resulting hyperbolic representations provide a structured radial coordinate for interpretation, while radius control improves training stability by preventing boundary saturation. Extensive experiments across eight benchmark datasets demonstrate that HYDRA consistently achieves competitive or superior predictive performance while improving parameter efficiency and representation interpretability.
Chinese Translation
Kolmogorov-Arnold 网络（KAN）通过用可学习的单变量函数替代标量权重来增强非线性函数逼近能力。然而，为每个连接分配一个独立函数会导致大量参数冗余，从而限制了其可扩展性和效率。为了减少这种冗余，我们引入了双曲动态表示架构（HYDRA），这是 KAN 的一种参数高效的双曲扩展，它结合了基于样条的函数学习与庞加莱球中的表示。HYDRA 将向量值输入映射到有界的双曲潜在空间中，在切空间中执行 KAN 风格的更新，并采用低秩原型块在隐藏维度之间共享函数变换。所得的双曲表示为解释提供了结构化的径向坐标，而半径控制通过防止边界饱和提高了训练稳定性。在八个基准数据集上的大量实验表明，HYDRA 在持续实现具有竞争力或更优的预测性能的同时，提高了参数效率和表示可解释性。
cs.LG / 87 / 2608.12219
ScreenShot: A Foundation Model for Few-Shot Combination Drug Screening
ScreenShot：用于少样本联合用药筛选的基础模型
Abstract
Treating patients with combinations of drugs reduces the risk of resistance to any individual drug. Finding effective combinations is difficult because the large search space makes combinatorial screens prohibitively expensive, time consuming, and often technically infeasible. Predictive models can fill this gap, yet existing methods typically require molecular profiling of each sample and per-cohort training, limiting their applicability when time and tissue are scarce. To address this challenge, we introduce ScreenShot, a hierarchical transformer pretrained on 40 drug screening datasets covering 3,700 drugs and 6,000 biological samples, whose architecture mirrors the nested structure of screening data. Given a few-shot context of observations from a new patient, ScreenShot predicts the response of the sample to combination therapies through in-context learning, operating directly on functional measurements with no fine-tuning and no molecular profiling. On four held-out datasets, ScreenShot outperforms all baselines in both prediction accuracy and identification of selectively effective treatments. ScreenShot's internal representations are directly useful for experimental design: we use them to drive a weighted k-means++ active learning strategy that selects which experiments to run, achieving the same hit detection as uniform screening with a third of the budget. Source code and interactive dashboard: https://github.com/tansey-lab/screenshot.
Chinese Translation
用药物联合治疗患者可降低对任何单一药物产生耐药性的风险。寻找有效的联合用药方案十分困难，因为庞大的搜索空间使得组合筛选极其昂贵、耗时，且通常在技术上不可行。预测模型可以填补这一空白，但现有方法通常需要对每个样本进行分子谱分析，并按队列进行训练，这限制了它们在时间和组织稀缺时的适用性。为应对这一挑战，我们提出了ScreenShot，这是一个在40个药物筛选数据集上预训练的分层Transformer，涵盖3,700种药物和6,000个生物样本，其架构反映了筛选数据的嵌套结构。给定来自新患者的少量观察上下文，ScreenShot通过上下文学习预测样本对联合疗法的反应，直接对功能测量数据进行操作，无需微调，也无需分子谱分析。在四个留出数据集上，ScreenShot在预测准确性和识别选择性有效治疗方面均优于所有基线方法。ScreenShot的内部表示可直接用于实验设计：我们利用它们来驱动一种加权k-means++主动学习策略，该策略选择要进行的实验，以三分之一的预算实现了与均匀筛选相同的命中检测效果。源代码和交互式仪表板：https://github.com/tansey-lab/screenshot。
cs.LG / 88 / 2608.12231
An Efficient Near-Optimal Algorithm for Adversarial $m$-Set Bandits
对抗性 $m$-集合老虎机的一种高效近最优算法
Abstract
We study adversarial combinatorial bandits with $m$-set actions, where at each round the learner selects $m$ out of $d$ items and observes only the aggregate loss of the selected items. The resulting action set contains $K=\binom{d}{m}$ elements and can therefore be exponentially large. Nevertheless, the loss of every action is determined by the same $d$-dimensional vector of item losses. We propose a computationally efficient algorithm that exploits this structure without explicitly enumerating the action set. Against adaptive non-anticipating adversaries, it guarantees, with probability at least $1-δ$, regret against the best fixed action of \[ R_T = O\left(\sqrt{dT\log(K/δ)}\right). \] This matches the high-probability regret bound of the finite-action EXP3-KW algorithm of Zimmert and Lattimore, whose direct implementation may require exponential space. Our algorithm instead represents each sampling distribution with $d$ parameters and runs in polynomial time without enumerating the action set. Thus, it resolves the open problem posed by Maiti et al.
Chinese Translation
我们研究具有 $m$-集合动作的对抗性组合老虎机，其中在每一轮中，学习者从 $d$ 个物品中选择 $m$ 个，并且仅观察到所选物品的总损失。由此产生的动作集包含 $K=\binom{d}{m}$ 个元素，因此可能是指数级大的。然而，每个动作的损失都由同一个 $d$ 维物品损失向量决定。我们提出一种计算高效的算法，利用这种结构而无需显式枚举动作集。针对自适应非预期对手，它以至少 $1-δ$ 的概率保证，相对于最佳固定动作的遗憾为 \[ R_T = O\left(\sqrt{dT\log(K/δ)}\right). \] 这与 Zimmert 和 Lattimore 的有限动作 EXP3-KW 算法的高概率遗憾界相匹配，而该算法的直接实现可能需要指数级空间。我们的算法改为用 $d$ 个参数表示每个采样分布，并在多项式时间内运行，无需枚举动作集。因此，它解决了 Maiti 等人提出的开放问题。
cs.LG / 89 / 2608.12259
Calibration Bets on the Past: Post-Training Quantization for Financial Time-Series Forecasting
以过去为赌注的校准：面向金融时间序列预测的训练后量化
Abstract
Financial forecasting models are typically developed in full precision, yet production deployment often requires low-precision inference to reduce memory and computational cost. Post-training quantization (PTQ) enables such deployment without retraining. However, reliable activation quantization requires calibration: activation ranges are estimated from historical data before deployment and then remain fixed during future inference. The importance of this deployment choice for financial forecasting remains poorly understood. We present a systematic study of activation calibration for PTQ in cross-sectional volatility forecasting on the S&P 500. Our evaluation covers seven representative neural architectures, eight walk-forward test years (2018-2025), and 560 trained models. We find that activation calibration has little effect at 8 bits but becomes the primary determinant of predictive performance at 4 bits. Under default absolute-maximum (abs-max) calibration, static 4-bit quantization of both weights and activations removes 11-62% of the full-precision mean information coefficient in affected architectures. Replacing abs-max with percentile calibration recovers 53-94% of this degradation in the four most affected architectures. The preferred activation range also varies across market periods. Narrow ranges improve resolution under typical market conditions but lose part of their advantage when test-period market dispersion exceeds the calibration history. These findings show that activation calibration is a first-class deployment decision for reliable 4-bit PTQ in financial forecasting. When substantial degradation remains, 8-bit activations or weight-only 4-bit quantization provide more robust deployment choices.
Chinese Translation
金融预测模型通常以全精度开发，但生产部署通常需要低精度推理以减少内存和计算成本。训练后量化（PTQ）无需重新训练即可实现这种部署。然而，可靠的激活量化需要校准：激活范围在部署前根据历史数据估计，然后在未来推理期间保持固定。这一部署选择对金融预测的重要性仍鲜为人知。我们对标普500横截面波动率预测中PTQ的激活校准进行了系统研究。我们的评估涵盖七种代表性神经架构、八个前推测试年份（2018-2025）以及560个训练模型。我们发现，激活校准在8位时影响很小，但在4位时成为预测性能的主要决定因素。在默认的绝对最大值（abs-max）校准下，对权重和激活都进行静态4位量化会使受影响架构的全精度平均信息系数损失11-62%。将abs-max替换为百分位校准可在四个受影响最大的架构中挽回53-94%的性能损失。偏好的激活范围也随市场时期而变化。窄范围在典型市场条件下可提高分辨率，但当测试期市场离散度超过校准历史时，会失去部分优势。这些发现表明，激活校准是实现可靠4位PTQ的一项头等重要的部署决策。当仍存在显著性能下降时，8位激活或仅权重的4位量化提供了更稳健的部署选择。
cs.LG / 90 / 2608.12271
Earth observation embeddings are effective sub-grid descriptors for probabilistic weather downscaling
地球观测嵌入是概率性天气降尺度的有效次网格描述符
Abstract
Global weather reanalyses and forecasts resolve the evolving atmospheric state on coarse grids, but site-specific applications require predictions at arbitrary locations where near-surface conditions also depend on unresolved terrain and land-surface properties. Existing probabilistic downscalers address this gap using hand-crafted topographic descriptors. We ask instead whether Earth observation foundation models can provide transferable sub-grid surface representations for probabilistic weather downscaling. We augment a convolutional conditional neural process that downscales coarse ERA5 reanalysis fields at ~25 km resolution with a learned local surface descriptor, obtained by compressing a patch of TESSERA embeddings at 10 m resolution. Although these embeddings summarise surface conditions over annual timescales, they improve downscaling of instantaneous 2 m temperature and 10 m wind speed by encoding persistent surface properties that capture a location's departure from the coarse-grid atmospheric state. Across five climatically diverse regions, the embedding improves point and probabilistic skill at stations held out in both space and time, overall improving CRPS skill by 11.5% for 2 m temperature and 6.2% for 10 m wind speed. We further analyse how its contribution differs by variable, finding that topography explains more of temperature's sub-grid structure, while TESSERA provides additional surface information for wind speed. These improvements persist when the coarse input is changed from ERA5 to forecasts from the Aurora AI forecasting model, and when predicting at newly deployed stations with no regional history. To our knowledge, this is the first evidence that long-timescale Earth-observation embeddings can support short-timescale weather downscaling where sub-grid departures are systematically structured by persistent surface properties.
Chinese Translation
全球天气再分析和预报在粗网格上解析不断演变的大气状态，但特定地点的应用需要在任意位置进行预测，而那里的近地面条件还取决于未解析的地形和陆面特性。现有的概率降尺度器使用手工制作的地形描述符来解决这一差距。我们转而探究地球观测基础模型是否能为概率性天气降尺度提供可迁移的次网格地表表征。我们通过压缩10米分辨率的TESSERA嵌入图块得到一个学习的局部地表描述符，用它来增强一个卷积条件神经过程，该过程对约25千米分辨率的粗ERA5再分析场进行降尺度。尽管这些嵌入总结的是年际时间尺度上的地表条件，但它们通过编码持久的地表特性来改进瞬时2米气温和10米风速的降尺度，这些特性捕捉了某个位置与粗网格大气状态的偏离。在五个气候多样的区域中，该嵌入提高了在空间和时间上均被留出的站点的点值技能和概率技能，总体上将2米气温的CRPS技巧提高了11.5%，将10米风速的CRPS技巧提高了6.2%。我们进一步分析了其贡献如何因变量而异，发现地形更多地解释了气温的次网格结构，而TESSERA为风速提供了额外的地表信息。当粗输入从ERA5改为Aurora AI预报模型的预报时，以及在没有区域历史的新部署站点进行预测时，这些改进仍然存在。据我们所知，这是首个证据，表明长时间尺度的地球观测嵌入能够支持那些次网格偏离由持久地表特性系统性结构的短时间尺度天气降尺度。
cs.LG / 91 / 2608.12302
A Framework for Designing Reward Functions: From Objectives to Features to Human-Aligned Reward Functions
设计奖励函数的框架：从目标到特征再到与人类对齐的奖励函数
Abstract
We present a formal process to enable non-experts to instantiate and iterate on human-aligned reward functions, i.e. reward functions that adhere to a given preference ordering over trajectories. Given a task described in natural language, our process produces a linear reward function in three steps: distill the task's objectives into a set of fundamental objectives and derive measurable outcome variables that capture those fundamental objectives, select a causally representative subset of outcome variables as the reward terms, and fit weights to those reward terms via preference elicitation. Our contributions describe the first step and formalize the latter two steps. The first is a guided workflow for deriving outcome variables. The second is a reduction of reward term selection to minimum-cost partial cover on a causal DAG, solved in polynomial time via max-flow. The third is a geometric framing of weight fitting as a convex feasibility problem iteratively narrowed by preference queries, solved by existing separation oracle methods. To the best of our knowledge, this is the first reward-design method that maintains a deterministically conflict-free feasible weight region, narrowed to a desired tolerance via a separation oracle with O(n log κ) preference queries.
Chinese Translation
我们提出了一个形式化流程，使非专家能够实例化并迭代与人类对齐的奖励函数，即遵循给定轨迹偏好排序的奖励函数。给定一个用自然语言描述的任务，我们的流程通过三个步骤生成线性奖励函数：将任务目标提炼为一组基本目标，并推导出能够刻画这些基本目标的可测量结果变量；选择一个具有因果代表性的结果变量子集作为奖励项；并通过偏好获取为这些奖励项拟合权重。我们的贡献描述了第一步，并将后两步形式化。第一步是用于推导结果变量的引导式工作流。第二步是将奖励项选择归约为因果有向无环图上的最小代价部分覆盖问题，并通过最大流在多项式时间内求解。第三步是将权重拟合几何化地表述为一个凸可行性问题，并通过偏好查询迭代缩小可行域，由现有的分离预言机方法求解。据我们所知，这是第一种奖励设计方法，能够维持一个确定性地无冲突的可行权重区域，并通过一个具有 O(n log κ) 次偏好查询的分离预言机将其缩小到所需容差。
cs.LG / 92 / 2608.12306
Redistribution-based Cost Inference Improves Sparse Safe Offline RL
基于重分配的成本推断改善稀疏安全离线强化学习
Abstract
Safe offline RL typically assumes access to dense per-step cost annotations, but in practice supervisors provide only trajectory-level stop-feedback: a binary signal at the first unsafe transition, with no per-step attribution. We frame this as a temporal credit assignment problem and propose the Redistribution-based Cost Inference (RCI) framework, which converts sparse stop-feedback into dense per-step costs via return decomposition, then trains a constrained offline policy on the augmented dataset. We show that return-equivalent redistribution preserves the feasible policy set and the optimal Lagrangian in a CMDP, establishing that the transformation is lossless in theory while yielding better-conditioned cost critic learning in practice. Experiments on highway driving and robotic manipulation demonstrate substantially lower violation rates than sparse and classifier-based baselines, with robustness to heterogeneous dataset compositions and label noise.
Chinese Translation
安全离线强化学习通常假设可以获得密集的逐步成本标注，但在实践中监督者只提供轨迹级别的停止反馈：在首次不安全转移处给出的二元信号，且没有逐步归因。我们将此问题形式化为时间信用分配问题，并提出基于重分配的成本推断（RCI）框架，该框架通过回报分解将稀疏停止反馈转换为密集逐步成本，然后在增强数据集上训练约束离线策略。我们证明，回报等价的重分配在 CMDP 中保持了可行策略集和最优拉格朗日函数，从而表明该变换在理论上是无损的，同时在实践中使成本评论家的学习具有更好的数值条件。在高速公路驾驶和机器人操作上的实验表明，与稀疏和基于分类器的基线相比，其违规率显著更低，并且对异构数据集组成和标签噪声具有鲁棒性。
cs.LG / 93 / 2608.12307
AI4AI at Test-Time: Strong-to-Weak Capability Transfer via Harnesses
测试时 AI4AI：通过 Harness 实现强到弱能力迁移
Abstract
Recent work on distillation transfers the capabilities of large models to smaller ones often by updating the latter's parameters, through teacher forcing, on-policy distillation, and related training-time methods. In this paper, we ask whether such transfer can instead occur at test time. We study strong-to-weak scaffolding: whether a stronger builder model can construct inference-time harnesses that help a weaker target model solve tasks more reliably without any parameter updates. Using four representative Theory-of-Mind benchmarks, each builder model uses 5% of the data as a validation set to iteratively refine its harness over multiple rounds, after which the finalized harness is evaluated on the full test set. Empirically, this form of test-time capability transfer is highly effective, nearly doubling average target-model performance from 0.49 to 0.91. Our analysis shows that the gains come primarily from offloading unstable model reasoning into deterministic code, benchmark-specific routing, and strict answer-format enforcement, rather than from encouraging the target model to reason more extensively or sample more broadly. We further find that builder-model reasoning effort improves harness quality monotonically, platform effects are modest relative to the builder model's own capability, and weaker target models receive the largest gains. These results suggest that inference-time harness design is an important complement to conventional training-time distillation, enabling strong models to transfer cognitive structure to weaker models without retraining.
Chinese Translation
最近关于蒸馏的研究通常通过更新较小模型的参数，借助教师强制（teacher forcing）、同策略蒸馏（on-policy distillation）以及相关的训练时方法，将大模型的能力迁移到小模型。在本文中，我们探讨这种迁移是否可以在测试时进行。我们研究强到弱脚手架（strong-to-weak scaffolding）：一个更强的构建者模型（builder model）是否能够构造推理时 harness，在无需任何参数更新的情况下帮助较弱的目标模型更可靠地解决任务。在四个具有代表性的心智理论（Theory-of-Mind）基准上，每个构建者模型使用 5% 的数据作为验证集，在多轮中迭代优化其 harness，之后最终确定的 harness 在整个测试集上进行评估。实证上，这种测试时能力迁移形式非常有效，将目标模型的平均性能从 0.49 几乎翻倍提升到 0.91。我们的分析表明，这些收益主要来自将不稳定的模型推理卸载到确定性代码中、针对特定基准的路由以及严格的答案格式执行，而并非来自鼓励目标模型进行更广泛的推理或更广泛的采样。我们进一步发现，构建者模型的推理投入会单调地提升 harness 质量；平台效应相对于构建者模型自身能力而言较为有限；较弱的目标模型获得的收益最大。这些结果表明，推理时 harness 设计是对传统训练时蒸馏的重要补充，使强模型无需重新训练即可将认知结构迁移到弱模型。
cs.LG / 94 / 2608.11363
Adaptation of Generalist Robot Policies with Minimal Data
基于最小数据的通用机器人策略适应
Abstract
A central goal in robot learning is to move beyond task-specific human data collection toward robots that improve through autonomous interaction. Yet fully autonomous learning remains difficult with current policies: sparse rewards and weak zero-shot exploration make it unlikely that a robot will discover successful behavior from scratch. We study minimal-data adaptation, a regime in which a pre-trained robot policy must learn a new task from as little as one demonstration followed by autonomous online interaction. This setting serves as the closest tractable proxy for fully autonomous improvement, allowing us to study whether minimal human guidance can bootstrap autonomous learning and what algorithmic ingredients make it feasible. We build MiDAS, a simple offline-to-online RL recipe that first anchors a pre-trained VLA to the target task with behavior cloning on single/few demonstrations, then improves it through value-based online RL on a residual policy parameterization. Across LIBERO and RoboCasa, MiDAS recovers strong task performance from as little as one demonstration, substantially outperforming baselines and generalizing beyond demonstrated conditions. We further evaluate MiDAS on a bimanual YAM platform. Starting from a fragile low-success policy obtained from a single demonstration, MiDAS improves its robustness and learns new successful behaviors over ~6 hours of online interaction. To the best of our knowledge, this is the first demonstration of reliable robot policy adaptation from a single task demonstration.
Chinese Translation
机器人学习的一个核心目标是超越针对特定任务的人工数据收集，迈向通过自主交互不断改进的机器人。然而，使用当前策略进行完全自主学习仍然困难：稀疏奖励和较弱的零样本探索使得机器人不太可能从零开始发现成功行为。我们研究最小数据适应，即预训练机器人策略必须从少至一次演示以及随后的自主在线交互中学习新任务的情形。该设定是完全可以自主改进的最接近且可处理的代理，使我们能够研究最少的人类指导是否能够引导自主学习，以及哪些算法要素使其可行。我们构建了 MiDAS，一种简单的离线到在线强化学习方案，它首先通过单次/少量演示上的行为克隆将预训练的 VLA 锚定到目标任务，然后通过基于价值的在线强化学习在残差策略参数化上对其改进。在 LIBERO 和 RoboCasa 上，MiDAS 仅从一次演示就能恢复出较强的任务性能，显著优于基线，并能泛化到演示条件之外。我们进一步在双臂 YAM 平台上评估 MiDAS。从由单次演示得到的脆弱且低成功率的策略出发，MiDAS 通过约 6 小时的在线交互提高了其鲁棒性，并学会了新的成功行为。据我们所知，这是从单次任务演示中实现可靠机器人策略适应的首次证明。
cs.LG / 95 / 2608.11597
IoT-Enabled Autonomous Maritime Navigation in Smart Ports: A Curriculum-Guided Shared Policy Learning Framework
智慧港口中物联网使能的自主海事导航：一种课程引导的共享策略学习框架
Abstract
As smart port infrastructures increasingly rely on autonomous maritime devices enabled by the Internet of Things (IoT), ensuring reliable onboard navigation intelligence has become a critical challenge for safe and scalable operations in congested waterways. This paper investigates onboard autonomous navigation for such IoT devices under partial observability and dense traffic conditions. A curriculum-guided reinforcement learning framework with a shared recurrent policy is developed to enhance temporal reasoning, deployment scalability, and robustness of edge-level decision-making. Centralized training is adopted as an offline design-time strategy, while all navigation actions are executed fully onboard, consistent with IoT edge intelligence paradigms. Extensive simulations in multiple realistic port environments demonstrate that the proposed approach improves navigation reliability, collision avoidance, and training stability compared with standard baseline methods, and generalizes effectively to previously unseen high-density scenarios. The results indicate that curriculum-guided shared learning provides a practical solution for scalable deployment of IoT-enabled autonomous maritime devices in smart port operations.
Chinese Translation
随着智慧港口基础设施日益依赖由物联网（IoT）使能的自主海事设备，确保可靠的船载导航智能已成为在拥挤航道中实现安全且可扩展运营的关键挑战。本文研究了此类物联网设备在部分可观测和密集交通条件下的船载自主导航。提出了一种具有共享循环策略的课程引导强化学习框架，以增强时序推理、部署可扩展性和边缘级决策的鲁棒性。采用集中式训练作为离线设计时策略，而所有导航动作完全在船载端执行，符合物联网边缘智能范式。在多个真实港口环境中的大量仿真表明，与标准基线方法相比，所提方法提高了导航可靠性、避碰性能和训练稳定性，并能有效泛化到此前未见的高密度场景。结果表明，课程引导的共享学习为智慧港口运营中物联网使能的自主海事设备的可扩展部署提供了一种实用解决方案。
cs.LG / 96 / 2608.11590
CookVoice: Unified Framework for Style Controllable Multi-Modal Human Voice Generation
CookVoice：风格可控多模态人声生成的统一框架
Abstract
Human voice generation has made rapid progress in speech generation, singing voice generation, voice cloning, and voice editing. However, most existing systems are designed for specific tasks and often rely on task-dependent architectures, control signals, or autoregressive decoding, limiting fine-grained controllability and inference efficiency. In this paper, we propose CookVoice, a unified framework for multimodal, multi-style, and multi-task human voice generation. CookVoice decomposes the human voice into three key factors: content, prosody, and style, enabling both speech and singing voice generation within a unified model. To achieve precise and flexible controllability, we design a flexible alignment strategy that maps text, style, and prosody control signals onto the frame-level of spectrogram. This design allows CookVoice to support a wide range of tasks, including text-to-speech, text-to-singing voice, style-controllable generation, voice mimicry, voice conversion, and voice editing. Experimental results show that CookVoice achieves generation quality comparable to existing Text-to-Speech and text-to-singing voice baselines, while providing stronger style and prosody controllability. Moreover, CookVoice achieves comparable performance to large-scale baselines with only 43.51 million parameters and efficient inference using as few as 4 ODE steps, making it a practical solution for real-world human voice generation applications. Demo page is available at https://haoweilou.github.io/CookVoice/.
Chinese Translation
人声生成在语音生成、歌声生成、语音克隆和语音编辑方面取得了快速进展。然而，大多数现有系统是为特定任务设计的，并且通常依赖于任务相关的架构、控制信号或自回归解码，限制了细粒度可控性和推理效率。在本文中，我们提出了 CookVoice，一个用于多模态、多风格和多任务人声生成的统一框架。CookVoice 将人声分解为三个关键因素：内容、韵律和风格，从而在统一模型内同时实现语音和歌声生成。为了实现精确且灵活的可控性，我们设计了一种灵活的对齐策略，将文本、风格和韵律控制信号映射到频谱图的帧级别上。这一设计使 CookVoice 能够支持广泛的任务，包括文本到语音、文本到歌声、风格可控生成、语音模仿、语音转换和语音编辑。实验结果表明，CookVoice 的生成质量可与现有的文本到语音和文本到歌声基线相媲美，同时提供更强的风格和韵律可控性。此外，CookVoice 仅用 43.51 百万参数和少至 4 个 ODE 步骤的高效推理，就实现了与大规模基线相当的性能，使其成为现实世界人声生成应用的实用解决方案。演示页面可在 https://haoweilou.github.io/CookVoice/ 获取。
cs.LG / 97 / 2608.11321
Spectral graph clustering with inhomogeneous latent geometry
具有非均匀潜在几何的谱图聚类
Abstract
We study spectral clustering in the presence of a confounding latent geometry. The leading eigenvectors may then be dominated by the latent geometry rather than by the communities. Nevertheless, we show in a block latent-space model that communities can be recovered from eigenvectors deeper in the spectrum. We analyze the spectral properties of the adjacency matrix through a limiting integral operator and use its structure to develop DBSPEC, a density-based spectral clustering algorithm that requires only approximate localization of the informative eigenvalue and is robust to poor eigenvalue separation. Crucially, this approach handles general latent geometries, overcoming restrictions to homogeneous toroidal models in prior works. Our theoretical predictions for the location of the informative eigenvalue notably align with observations in real-world experiments.
Chinese Translation
我们研究存在混杂潜在几何时的谱聚类。前几个特征向量此时可能由潜在几何主导，而非由社区主导。尽管如此，我们证明在一个块潜在空间模型中，社区可以从谱中更深处的特征向量中恢复。我们通过一个极限积分算子分析邻接矩阵的谱性质，并利用其结构开发了 DBSPEC，一种基于密度的谱聚类算法，该算法仅需要对信息性特征值进行近似定位，并且对特征值分离不佳具有鲁棒性。关键的是，该方法处理一般的潜在几何，克服了先前工作对均匀环面模型的限制。我们对信息性特征值位置的理论预测明显与真实世界实验中的观测结果一致。
cs.LG / 98 / 2608.11762
A comparison of CNN architectures for Alzheimer's disease detection in single-view MRI scans
单视图 MRI 扫描中用于阿尔茨海默病检测的 CNN 架构比较
Abstract
Alzheimer's disease is a leading cause of death with no cure. Therefore, early detection is critical to slow progression and preserve quality of life. Diagnosis relies on medical history, cognitive tests, physical exams, and MRI brain scans, making deep learning suitable for Alzheimer's classification. This work proposes a benchmark that evaluates ten different convolutional neural network (CNN) architectures (including ResNet, DenseNet, MobileNet, EfficientNet, and VGG family models) under the same held-out test split protocol. A two-stage transfer learning and full fine-tuning pipeline is introduced to perform training using a class-balanced subset (3,900 images) derived from the OASIS medical imaging dataset, comprising 86,437 single-view MRI brain scans labeled into four classifications of Alzheimer's disease: Non-Demented, Very Mild Dementia, Mild Dementia, and Moderate Dementia. The best results were achieved by VGG16, with a 0.9637 validation accuracy and a 0.9533 test accuracy score. A key finding documented in this work is the difficulty of classifying the transition from Non-Demented to Very Mild Demented stages, observed consistently across all ten architectures.
Chinese Translation
阿尔茨海默病是一种主要的死亡原因，且目前无法治愈。因此，早期检测对于减缓疾病进展和保持生活质量至关重要。诊断依赖于病史、认知测试、体格检查和脑部 MRI 扫描，这使得深度学习适用于阿尔茨海默病分类。本工作提出了一个基准，在相同的留出测试集划分协议下评估十种不同的卷积神经网络（CNN）架构（包括 ResNet、DenseNet、MobileNet、EfficientNet 和 VGG 系列模型）。本研究引入了一个两阶段迁移学习和完全微调流程，使用从 OASIS 医学影像数据集派生的类别平衡子集（3,900 张图像）进行训练；该数据集包含 86,437 张单视图 MRI 脑部扫描图像，并标注为阿尔茨海默病的四种分类：非痴呆、非常轻度痴呆、轻度痴呆和中度痴呆。最佳结果由 VGG16 实现，验证准确率为 0.9637，测试准确率分数为 0.9533。本工作记录的一个关键发现是，在所有十个架构中一致观察到，对从非痴呆阶段到非常轻度痴呆阶段的转变进行分类是困难的。
cs.LG / 99 / 2608.11995
Latent variable models for simultaneous EOV identification and removal in population-based SHM
基于群体的结构健康监测中EOV同步识别与去除的潜变量模型
Abstract
The robust treatment of environmental and operational variability (EOV) is an open challenge in population-based structural health monitoring (PBSHM). The difficulty is compounded in the case that the EOV signals are unmeasured. A common approach in conventional SHM is to apply \emph{projection-based} methods that discard subspaces of healthy feature data, reasoning that the EOV signal dominates the variance of the measured features. However, a common pitfall of projection-based approaches is that when damage acts close to the same variance-dominant direction, damage sensitivity is removed along with the EOV. An alternative identifying assumption for the removal of particular unmeasured EOVs is slowness; the latent EOV process is characterised by its long temporal correlation. In this paper, the latent EOV is cast as a state-space Gaussian process, enabling tractable $\mathcal{O}(T)$ inference via a Kalman filter. A robust hierarchical Bayesian identification framework is developed that enables population-level identification of latent EOVs and EOV-free residual features, using a Laplace approximation. The approach is first validated on a single laboratory-scale benchmark structure from the literature, subject to thermal EOVs, demonstrating robust damage detection and EOV recovery. The method is then applied to a simulated nine-turbine offshore wind farm with staggered deployment and damage, where it delivers a substantial true-positive uplift over projection and cointegration-based baselines at matched false-positive rates.
Chinese Translation
在基于群体的结构健康监测（PBSHM）中，对环境与运行变异性（EOV）的稳健处理是一个开放性挑战。当EOV信号未被测量时，这一困难会进一步加剧。传统SHM中的一种常见方法是应用基于投影的方法，这些方法丢弃健康特征数据的子空间，其理由是EOV信号主导了测量特征的方差。然而，基于投影方法的一个常见缺陷是，当损伤作用于接近同一方差主导方向时，损伤敏感性会与EOV一起被去除。对于去除特定未测量EOV，另一种识别假设是慢变性；潜在EOV过程的特征在于其较长的时间相关性。本文将潜在EOV建模为状态空间高斯过程，从而能够通过卡尔曼滤波器进行可处理的 $\mathcal{O}(T)$ 推断。本文开发了一个稳健的分层贝叶斯识别框架，该框架利用拉普拉斯近似，能够在群体层面识别潜在EOV和无EOV的残差特征。该方法首先在文献中的一个受热EOV影响的实验室规模基准结构上进行了验证，展示了稳健的损伤检测和EOV恢复能力。随后，该方法被应用于一个模拟的九台风机海上风电场，该风电场具有分阶段部署和损伤；在匹配的假阳性率下，该方法相较于基于投影和基于协整的基线方法实现了显著的真阳性提升。
cs.LG / 100 / 2608.11480
Forward Trajectory Steering for Hamilton-Jacobi Reachability Analysis
面向 Hamilton-Jacobi 可达性分析的前向轨迹引导
Abstract
Hamilton-Jacobi (HJ) reachability provides a mathematically rigorous framework for safe control of dynamical systems, but its practical application is bottlenecked by the computational complexity of solving Hamilton-Jacobi-Isaacs variational inequality PDEs in high dimensions. Physics-informed neural networks (PINNs) have recently emerged as a promising alternative to classical mesh-based solvers, yet their performance is highly sensitive to the choice of collocation sampling. In order to learn accurate safety value functions, existing PINNs-based HJ reachability solvers must rely on complex training pipelines and auxiliary supervision. In this work, we propose STEER2REACH (S2R), a PINNs-based HJ reachability solver that requires minimal modification on top of standard PINNs training. S2R's key contribution is a lightweight, low-overhead adaptive collocation sampling distribution constructed by steering forward trajectories using a combination of the optimal control and disturbance signals induced by the current value function, with injected stochastic exploration noise. We demonstrate that despite its simplicity, S2R achieves competitive--and in some cases improved--performance on safety metrics while reducing relative L2 error across a range of reachability benchmarks compared with SoTA MPC-guided HJ reachability solvers, all without requiring multi-stage training or MPC-based supervision.
Chinese Translation
Hamilton-Jacobi（HJ）可达性为动力系统的安全控制提供了数学上严格的框架，但其实际应用受到高维 Hamilton-Jacobi-Isaacs 变分不等式偏微分方程求解计算复杂度的瓶颈制约。物理信息神经网络（PINNs）近年来已成为经典基于网格求解器的一种有前景的替代方法，然而其性能对配点采样方式的选择高度敏感。为了学习准确的安全值函数，现有的基于 PINNs 的 HJ 可达性求解器必须依赖复杂的训练流程和辅助监督。在本文中，我们提出了 STEER2REACH（S2R），一种基于 PINNs 的 HJ 可达性求解器，它在标准 PINNs 训练的基础上只需极小的修改。S2R 的关键贡献是一种轻量、低开销的自适应配点采样分布，该分布通过利用当前值函数所诱导的最优控制信号和扰动信号的组合来引导前向轨迹，并注入随机探索噪声而构造得到。我们证明，尽管 S2R 设计简单，与最先进的（SoTA）MPC 引导的 HJ 可达性求解器相比，它在一系列可达性基准测试上取得了具有竞争力——在某些情况下更优——的安全指标表现，同时降低了相对 L2 误差，且无需多阶段训练或基于 MPC 的监督。
cs.LG / 101 / 2608.11760
Tight Nonasymptotic Local Convergence of Sinkhorn-Knopp
Sinkhorn-Knopp 的紧非渐近局部收敛性
Abstract
We revisit the Sinkhorn-Knopp (SK) algorithm for the matrix scaling problem. Despite extensive literature on the global convergence of SK and its variants, its local linear convergence behavior remains less understood. We address this gap by providing the first nonasymptotic local analysis of SK that matches the rate obtained from existing asymptotic Jacobian-based arguments. We show that under certain connectivity conditions, SK is a polynomial-time algorithm for doubly stochastic matrix scaling. With the developed tools, we showcase the local suboptimality of SK and provide accelerated variants. Finally, for dense matrices, we improve the complexity of existing first-order matrix scaling algorithms from $O(\tfrac{n^{7/3}}{\varepsilon^{2/3}})$ to $O(\tfrac{n^{9/4}}{\sqrt{\varepsilon}})$.
Chinese Translation
我们重新审视用于矩阵缩放问题的 Sinkhorn-Knopp (SK) 算法。尽管关于 SK 及其变体的全局收敛性已有大量文献，但其局部线性收敛行为仍较少被理解。我们通过给出 SK 的首个非渐近局部分析来填补这一空白，该分析所得到的速率与现有基于雅可比的渐近论证所得速率相匹配。我们证明在某些连通性条件下，SK 是用于双随机矩阵缩放的多项式时间算法。利用所建立的工具，我们展示了 SK 的局部次优性，并给出了加速变体。最后，对于稠密矩阵，我们将现有的一阶矩阵缩放算法的复杂度从 $O(\tfrac{n^{7/3}}{\varepsilon^{2/3}})$ 改进到 $O(\tfrac{n^{9/4}}{\sqrt{\varepsilon}})$。
cs.LG / 102 / 2608.12009
Adaptive Bregman Proximal Stochastic Gradient with a Stabilized Barzilai--Borwein Step Size
具有稳定化 Barzilai--Borwein 步长的自适应 Bregman 近端随机梯度方法
Abstract
Bregman proximal stochastic gradient (BPSG) methods bring variance-reduced composite optimization to objectives whose geometry is poorly captured by Euclidean smoothness. Their performance, however, remains sensitive to the step size: raw stochastic curvature estimates can fluctuate sharply, whereas line searches add repeated proximal evaluations. We introduce Ada-BPSG, a line-search-free BPSG method that couples the SAGA gradient table with a stabilized Barzilai--Borwein (BB) candidate. A mediant aggregates incremental secant information so that nearly singular local ratios receive little weight, and an explicit safeguard translates the resulting curvature estimate into the bounded step-size sequence required for convergence. This design yields a direct analytical chain from relative smoothness and component-wise variance control to convergence in finite-dimensional normed spaces. We prove an $O(n/K)$ ergodic rate for convex objectives, a restarted linear rate under relative quadratic growth, and an $O(1/K)$ bound for a Bregman proximal residual in the nonconvex setting. On logistic regression and sparse nonnegative matrix factorization, Ada-BPSG combines low objective values with substantially less sensitivity to the initial step size than standard variance-reduced baselines, while avoiding line search.
Chinese Translation
Bregman 近端随机梯度（BPSG）方法将方差缩减复合优化引入到那些其几何结构难以被欧几里得光滑性所刻画的目标函数中。然而，它们的性能仍然对步长敏感：原始随机曲率估计可能剧烈波动，而线搜索会增加重复的近端评估。我们提出 Ada-BPSG，一种免线搜索的 BPSG 方法，它将 SAGA 梯度表与稳定化的 Barzilai--Borwein（BB）候选步长相结合。一个中项聚合增量割线信息，使得近乎奇异的局部比率获得很小的权重，并且一个显式保护机制将所得到的曲率估计转换为收敛所需的有界步长序列。这种设计产生了一条直接的分析链条，从相对光滑性和逐分量方差控制到有限维赋范空间中的收敛性。我们证明了凸目标下的 $O(n/K)$ 遍历速率，相对二次增长条件下的重启线性速率，以及非凸情形下 Bregman 近端残差的 $O(1/K)$ 界。在逻辑回归和稀疏非负矩阵分解上，Ada-BPSG 在获得低目标函数值的同时，对初始步长的敏感性显著低于标准方差缩减基线方法，并且避免了线搜索。
cs.LG / 103 / 2608.12043
Direct Acceleration of Stochastic Root-Finding Without Variance Reduction and Regularization
无需方差缩减与正则化的随机求根直接加速
Abstract
Acceleration for deterministic root-finding problems has been extensively studied in recent years; specifically, the anchor-based, or Halpern-type methods achieve optimal convergence rates with respect to the operator norm. However, acceleration via these methods does not directly carry over to stochastic setting due to accumulation of errors, unless one enforces diminishing variance via increasing batch sizes or variance reduction techniques. In this work, we show that another class of acceleration, namely the dual-anchor mechanism, extends to the stochastic setting without such error accumulation, in contrast to anchor-based algorithms. Consequently, we cleanly achieve $O(ε^{-3})$ complexity with iteration-independent batch size, without any variance reduction or double-loop recursive regularization, for stochastic root-finding (resp. fixed-point) problems with cocoercivity (resp. square-nonexpansivity) in expectation. For strongly monotone operators, the same algorithm attains a sharper $\widetilde{O} (ε^{-2})$ complexity, nearly matching the lower bound in terms of $ε$-dependence.
Chinese Translation
近年来，确定性求根问题的加速方法已被广泛研究；具体而言，基于锚点的（或称 Halpern 型）方法在算子范数意义下达到了最优收敛速率。然而，通过这些方法实现的加速并不能直接迁移到随机情形，因为误差会不断累积，除非通过增大批量大小或使用方差缩减技术来强制减小方差。在本文中，我们证明另一类加速机制，即双锚点机制，能够在不出现这种误差累积的情况下推广到随机情形，这与基于锚点的算法形成对比。因此，对于在期望意义下具有余强制性（相应地，平方非扩张性）的随机求根（相应地，不动点）问题，我们在批量大小与迭代次数无关、且不使用任何方差缩减或双层递归正则化的情况下，简洁地实现了 $O(ε^{-3})$ 复杂度。对于强单调算子，同样的算法获得了更优的 $\widetilde{O} (ε^{-2})$ 复杂度，几乎在 $ε$-依赖意义上匹配了下界。
cs.LG / 104 / 2608.12111
The Advective Fisher-Rao Geometry of Deterministic Measure Transport
确定性测度传输的对流 Fisher-Rao 几何
Abstract
A novel advective Fisher-Rao metric is introduced for optimization tasks on paths of probability measures governed by the continuity equation. This metric is shown to lead to optimal descent directions. It is then shown that this metric arises naturally from three different perspectives: As the rescaled zero-noise limit of the Fisher-Rao metric on path measures, as the expected value of the second variation of the Freidlin--Wentzell large deviation rate functional, and as the Hessian of the Benamou--Brenier action functional from dynamic optimal transport. We supplement this geometric construction with computational experiments. Here, we demonstrate empirically that the advective Fisher-Rao metric yields the desired optimal fitting of probability densities, whereas the Gauss--Newton method yields optimal fitting of velocity fields.
Chinese Translation
针对由连续性方程支配的概率测度路径上的优化任务，引入了一种新的对流 Fisher-Rao 度量。该度量被证明可导出最优下降方向。然后证明该度量自然地从三个不同角度产生：作为路径测度上 Fisher-Rao 度量的重标度零噪声极限，作为 Freidlin--Wentzell 大偏差率泛函的二阶变分的期望，以及作为动态最优传输中 Benamou--Brenier 作用泛函的 Hessian。我们通过计算实验对这一几何构造进行补充。在此，我们通过实验证明，对流 Fisher-Rao 度量能够实现所需的概率密度最优拟合，而 Gauss--Newton 方法则实现速度场的最优拟合。
cs.LG / 105 / 2608.12000
Beyond Local Power: Functional Connectivity Analysis for Subject-Independent Learning Style Recognition
超越局部功率：面向受试者无关学习风格识别的功能连接分析
Abstract
Identifying individual learning styles optimizes pedagogical efficacy. While traditional questionnaires are structured, behavioral tracking methods require prolonged interaction log accumulation. To overcome these temporal constraints, this paper proposes an objective Electroencephalography (EEG) approach evaluating Phase Locking Value (PLV) connectivity against localized features across the Active-Reflective (AR) and Verbal-Visual (VV) Felder-Silverman dimensions. EEG signals were recorded from 28 participants during Raven's Advanced Progressive Matrices tasks. Support Vector Machine classification used Leave-One-Subject-Out Cross-Validation (LOSO-CV) alongside a 70:30 intra-subject split. The VV dimension achieved 70.00% subject-level accuracy driven by distinct fronto-occipital polarization. Conversely, the AR dimension yielded lower cross-subject generalizability (55.56%) due to overlapping executive networks and a "Systematic Neural Inversion" phenomenon, where stable individual connectivity signatures operated diametrically opposed to global boundaries (up to 20-0 voting margins). Ultimately, these outcomes demonstrate that rigid "one-size-fits-all" classifiers are bounded by biological diversity, emphasizing the need for future adaptive feature transformation techniques to bridge the cross-subject generalization gap.
Chinese Translation
识别个体学习风格可优化教学效能。传统问卷具有结构化特点，而行为追踪方法则需要长时间的交互日志积累。为克服这些时间限制，本文提出一种客观的脑电图（EEG）方法，在 Felder-Silverman 的活跃-反思（AR）和言语-视觉（VV）维度上，评估相位锁定值（PLV）连接性与局部特征。在 Raven 高级渐进矩阵任务期间，记录了 28 名参与者的脑电信号。支持向量机分类采用留一受试者交叉验证（LOSO-CV），并结合 70:30 的受试者内划分。VV 维度在受试者层面上达到 70.00% 的准确率，其驱动因素是明显的额-枕极化。相反，AR 维度由于执行网络重叠以及“系统性神经反转”现象，跨受试者泛化能力较低（55.56%）；在该现象中，稳定的个体连接特征与全局边界呈截然相反的方向运行（投票差距最高可达 20-0）。最终，这些结果表明，僵化的“一刀切”分类器受到生物学多样性的限制，并强调未来需要自适应特征变换技术来弥合跨受试者泛化差距。
cs.LG / 106 / 2608.11444
Large-scale AI-Ready Data for Anti-Cancer Drug Response Modeling
用于抗癌药物反应建模的大规模AI就绪数据
Abstract
Drug response prediction (DRP) models are an active area of research in pharmacogenomics, with growing potential to accelerate the identification of effective anticancer drugs. However, their predictive performance is often constrained by limited dataset scale and insufficient coverages of cancer and chemical spaces. In addition, inconsistent benchmarking practices hinder reliable comparison across models. Standardized frameworks, such as the Innovative Methodologies and New Data for Predictive Oncology Model Evaluation (IMPROVE) project, provide unified data schemas and evaluation protocols for consistent benchmarking, but improving model generalizability requires larger and more diverse training data. In this work, we substantially expand the IMPROVE benchmark through large-scale integration of pharmacogenomic data, primarily from PharmacoDB, together with additional smaller data sources. The expanded resource includes millions of drug response measurements, broader multi-omics coverage, and a major increase in chemical diversity, adding more than 50,000 compounds. To evaluate the impact of the new dataset compared to the original IMPROVE benchmark dataset, we trained DRP models using the two datasets and assess their prediction performance using a common test set and several evaluation strategies, including drug-blind, cancer-blind, and disjoint data splits. While cancer-blind performance remained comparable to the original benchmark, models trained on the expanded dataset showed consistent improvements in drug-blind and disjoint settings, indicating enhanced generalization to previously unseen compounds. These results position the expanded dataset as a community resource that provides a richer foundation for developing DRP models intended to aid in the discovery of novel anticancer drugs.
Chinese Translation
药物反应预测（DRP）模型是药物基因组学中一个活跃的研究领域，在加速有效抗癌药物识别方面具有日益增长的潜力。然而，其预测性能往往受到数据集规模有限以及癌症空间和化学空间覆盖不足的制约。此外，不一致的基准测试实践妨碍了模型之间的可靠比较。标准化框架，例如预测性肿瘤学模型评估创新方法与新数据（IMPROVE）项目，为一致基准测试提供了统一的数据模式和评估协议，但提高模型泛化能力需要更大且更多样化的训练数据。在这项工作中，我们通过大规模整合主要来自 PharmacoDB 的药物基因组数据以及其他较小数据源，大幅扩展了 IMPROVE 基准。扩展后的资源包括数百万条药物反应测量数据、更广泛的多组学覆盖，以及化学多样性的显著增加，新增了超过 50,000 种化合物。为了评估新数据集相对于原始 IMPROVE 基准数据集的影响，我们使用这两个数据集训练了 DRP 模型，并使用一个共同的测试集和几种评估策略（包括药物盲、癌症盲和不相交数据划分）评估它们的预测性能。虽然癌症盲性能与原始基准保持相当，但在扩展数据集上训练的模型在药物盲和不相交设置中表现出一致的改进，表明对先前未见过的化合物的泛化能力有所增强。这些结果将扩展后的数据集定位为一种社区资源，为开发旨在帮助发现新型抗癌药物的 DRP 模型提供了更丰富的基础。
cs.LG / 107 / 2608.12251
Regime-Gated Residual Mixture-of-Experts for Cross-Sectional Volatility Forecasting
用于横截面波动率预测的状态门控残差混合专家模型
Abstract
Financial volatility is regime dependent, yet incorporating regime information into neural networks can also destabilize training. This paper asks where such information should enter a neural cross-sectional volatility forecasting model. We study five-day realized-volatility forecasts for 1,027 U.S. equities using a rolling walk-forward evaluation framework in which information, model capacity, hyperparameter tuning, and random seeds are matched across architectures. We propose RG-ResMoE, a regime-gated residual mixture-of-experts architecture in which regime information is used only for expert routing rather than for direct forecasting. The base predictor models volatility from stock features, while a gating network uses regime state variables to route residual corrections. RG-ResMoE consistently outperforms a capacity-matched MLP in both forecasting accuracy and training stability in the main U.S. study. Similar gains are observed on an independent Japanese panel. The integration pathway is decisive: appending the same regime variables directly to the forecasting input degrades both predictive performance and training stability, whereas restricting them to the routing gate improves accuracy and Value-at-Risk calibration. Hard routing consistently underperforms soft routing. The results suggest that, in compact neural volatility forecasting models, the primary value of mixture-of-experts models lies less in increasing model capacity than in controlling how nonstationary regime information influences prediction.
Chinese Translation
金融波动率具有状态依赖性，但将状态信息纳入神经网络也可能破坏训练稳定性。本文探讨此类信息应进入神经横截面波动率预测模型的哪个位置。我们使用滚动前推评估框架研究对 1,027 只美国股票的 5 日已实现波动率预测，在该框架中，各架构的信息、模型容量、超参数调优和随机种子均保持一致。我们提出 RG-ResMoE，一种状态门控残差混合专家架构，其中状态信息仅用于专家路由而非直接预测。基础预测器基于股票特征对波动率建模，而门控网络使用状态变量来路由残差校正。在美国的主研究中，RG-ResMoE 在预测准确性和训练稳定性上均一致优于容量匹配的 MLP。在独立的日本股票面板上也观察到类似的改进。整合路径具有决定性作用：将相同的状态变量直接附加到预测输入会同时降低预测性能和训练稳定性，而将其限制在路由门中则会提高准确性和风险价值（Value-at-Risk）校准。硬路由始终表现不如软路由。结果表明，在紧凑的神经波动率预测模型中，混合专家模型的主要价值与其说在于增加模型容量，不如说在于控制非平稳状态信息如何影响预测。
cs.LG / 108 / 2608.11396
Generative Learning for Quantum Measurement Design
面向量子测量设计的生成式学习
Abstract
Extracting quantum information from a quantum state is a fundamental task of quantum computation, often requiring the estimation of many non-commuting observables under a finite measurement budget. For both near-term and early fault-tolerant settings, the measurement protocol must balance statistical efficiency against implementation resources such as circuit depth, connectivity, and entangling-gate count. Many existing strategies focus on two extremes: hardware-friendly product measurements with high sampling cost, and fully commuting measurements with deep circuits. Here we recast resource-constrained measurement design as a generative learning problem. We introduce FlowMeas, which uses a generative flow network to directly sample finite ensembles of shallow Clifford measurement circuits subject to a prescribed shot budget and hardware constraints. At zero entangling depth, FlowMeas learns qubit-wise commuting measurement schedules and already matches or improves leading product-measurement methods on nearly all molecular benchmarks. Allowing one or two entangling gate layers yields further reductions in energy estimation error of up to $27\%$ relative to the strongest state-independent product-measurement baseline. The learned policy can also be reused across related Hamiltonians, substantially accelerating retraining along a molecular potential-energy surface. We further obtain results for molecular Hamiltonians with up to 20 qubits and apply the framework to a compactly encoded 54-qubit interacting fermionic model, extending the demonstrated scale beyond prior molecular benchmarks. These results establish generative learning as a flexible and unified framework for quantum measurement design under practical resource constraints.
Chinese Translation
从量子态中提取量子信息是量子计算的一项基本任务，通常需要在有限的测量预算下估计许多非对易可观测量。对于近期和早期容错场景，测量协议必须在统计效率与电路深度、连通性和纠缠门数量等实现资源之间取得平衡。许多现有策略集中在两个极端：硬件友好但采样成本高的乘积测量，以及电路深度大的完全对易测量。在此，我们将资源受限的测量设计重新表述为一个生成式学习问题。我们引入了 FlowMeas，它使用生成式流网络，在规定的采样次数预算和硬件约束下，直接采样浅层 Clifford 测量电路的有限集合。在零纠缠深度下，FlowMeas 学习逐量子比特对易的测量调度，并且几乎在所有分子基准上已经达到或优于领先的乘积测量方法。允许一层或两层纠缠门可相对于最强的态无关乘积测量基线进一步将能量估计误差降低最多 $27\%$。学习到的策略还可以在相关哈密顿量之间重复使用，从而大幅加速沿分子势能面的重新训练。我们进一步针对最多 20 个量子比特的分子哈密顿量获得了结果，并将该框架应用于一个紧凑编码的 54 量子比特相互作用费米子模型，将所展示的规模扩展到先前分子基准之外。这些结果确立了生成式学习作为实际资源约束下量子测量设计的一种灵活且统一的框架。
cs.LG / 109 / 2608.11648
A Quantum/Classical Example Oracle Separation for Making Things Up
关于“凭空生成”的量子/经典示例预言机分离
Abstract
We study the power of quantum examples, as compared to classical examples, in the PAC learning framework. Here, we have two learning algorithms, both with access to quantum computation, but one gets quantum examples, whereas the other gets classical examples. It was previously unknown whether there were learning tasks that can be efficiently performed but not by the latter. Our primary result is to show that relative to an oracle, there are distributions that can be efficiently generated by a quantum learner with access to quantum examples, but not by a quantum learner with access to only classical examples, making progress to answering this question in the affirmative.
Chinese Translation
我们研究在 PAC 学习框架中，量子示例相较于经典示例的能力。这里，我们有两个学习算法，二者均能使用量子计算，但一个获得量子示例，而另一个获得经典示例。此前尚不清楚是否存在这样的学习任务：它们能够被高效地完成，但不能被后者高效地完成。我们的主要结果是证明，相对于某个预言机，存在一些分布，能够由一个可获取量子示例的量子学习者高效生成，但不能由一个仅可获取经典示例的量子学习者高效生成，从而在肯定回答这一问题上取得了进展。
cs.LG / 110 / 2608.11544
Fine-Tuning Generative Models for Extreme Events via CVaR-Penalized Wasserstein Gradient Flows
通过CVaR惩罚的Wasserstein梯度流微调生成模型以捕捉极端事件
Abstract
We propose CVaR-penalized Generative Particle Algorithm (CVaR-GPA), a robust, tail-agnostic algorithm for fine-tuning generative models to learn heavy-tailed distributions and capture extreme events, requiring no prior knowledge or estimation of the target's tail characteristics. The method is the Wasserstein gradient flow of the Lipschitz-regularized Kullback-Leibler (KL) divergence penalized by a Conditional Value-at-Risk (CVaR) discrepancy term: the Lipschitz-regularized KL divergence enables robust learning under minimal assumptions on the target distribution, while the CVaR penalty restores the velocity that otherwise vanishes prematurely in the under-sampled tails. The penalized flow admits a bounded but non-Lipschitz velocity field. This departs from the Lipschitz transport maps of standard generators, which preserve the tail behavior of a light-tailed source, and enables transport toward heavier-tailed targets. To define this flow on empirical measures, we derive the first-variation subgradients of CVaR from its Rockafellar-Uryasev representation, valid precisely where the classical density-based formula fails. The particle algorithm CVaR-GPA fine-tunes the output samples of any pre-trained model, without access to its architecture, and runs on an adaptive time horizon set by a kinetic-energy stopping criterion rather than a preset depth. On synthetic isotropic and anisotropic Student-$t$ target distributions, Neal's funnel distribution, and the real-world high-dimensional Fama-French 25 portfolio dataset, CVaR-GPA dramatically improves global and tail accuracy on heavy-tailed targets over the pre-trained baseline.
Chinese Translation
我们提出了CVaR惩罚的生成粒子算法（CVaR-GPA），这是一种稳健且与尾部特征无关的算法，用于微调生成模型以学习重尾分布并捕捉极端事件，无需对目标分布的尾部特征具有先验知识或进行估计。该方法是由条件风险价值（CVaR）差异项惩罚的Lipschitz正则化Kullback-Leibler（KL）散度的Wasserstein梯度流：Lipschitz正则化的KL散度使得在关于目标分布的最小假设下也能进行稳健学习，而CVaR惩罚项恢复了在欠采样尾部中原本会过早消失的速度。该惩罚流具有一个有界但非Lipschitz的速度场。这不同于标准生成器的Lipschitz传输映射——后者保持轻尾源的尾部行为——并使得能够向更重尾的目标进行传输。为了在经验测度上定义该流，我们从CVaR的Rockafellar-Uryasev表示推导出其一阶变分次梯度；这些次梯度恰好在经典的基于密度的公式失效之处有效。粒子算法CVaR-GPA可微调任何预训练模型的输出样本，而无需访问其架构，并在由动能停止准则设定的自适应时间范围内运行，而不是预设深度。在合成的各向同性和各向异性Student-$t$目标分布、Neal漏斗分布以及真实世界高维Fama-French 25投资组合数据集上，CVaR-GPA相比预训练基线在重尾目标上显著提高了全局精度和尾部精度。
神经与进化计算 (cs.NE)
2
cs.NE / 1 / 2608.11506
Predictive Allostatic Organization in Recurrent and Spiking Agents Under Partial Observability
部分可观测条件下递归与脉冲智能体中的预测性异稳态组织
Abstract
Adaptive behavior under partial observability depends on internal organization that carries information beyond the current observation. Drawing on Barrett and Miller's account of categorization as predictive, compressive, functionally organized, and allostatically constrained, we test whether recurrent and spiking agents develop internal states with corresponding computational properties. Agents operate in an energy-constrained foraging task requiring resource acquisition, threat avoidance, contact-dependent consumption, and regulation of an internal energy variable. In a frozen benchmark, learned agents outperform random and heuristic baselines; the trace-augmented recurrent policy is strongest overall, while spiking variants show stress-specific differences. Early internal dynamics predict later full-safe-efficient success above permutation baseline, reaching a maximum ROC-AUC of 0.802. Reduced PCA subspaces retain behaviorally relevant information. Feature-family controls show that predictive signal is distributed across trace, policy-head, internal-dynamics, observation, and allostatic variables, and low-energy state remains strongly decodable after explicit energy-related features are removed. Evaluation-time perturbations to temporal state, sensory information, operating conditions, and allostatic mechanisms alter behavior and/or internal prediction. Seed-balanced event probes show weaker but measurable information about future contact, successful consumption, and threat events, alongside strong low-energy decoding. We interpret this pattern as a computational analogue of predictive allostatic organization: distributed control regimes that are predictive, energy-sensitive, action-relevant, and partly causally involved, without claiming biological validation or discrete symbolic categories.
Chinese Translation
部分可观测条件下的适应性行为依赖于承载超出当前观测信息的内部组织。借鉴 Barrett 和 Miller 将范畴化描述为预测性、压缩性、功能性组织且受异稳态约束的观点，我们检验递归和脉冲智能体是否发展出具有相应计算性质的内部状态。智能体在一个能量受限的觅食任务中运行，该任务要求资源获取、威胁回避、依赖接触的消耗以及内部能量变量的调节。在冻结基准测试中，学习后的智能体优于随机和启发式基线；迹增强的递归策略总体最强，而脉冲变体表现出应激特异性差异。早期内部动态能够预测后来的全-安全-高效成功，其表现超过排列基线，最大 ROC-AUC 达到 0.802。降维后的 PCA 子空间保留了行为相关信息。特征族控制表明，预测信号分布在迹、策略头、内部动态、观测和异稳态变量中，并且在移除与能量相关的显式特征后，低能量状态仍然可以被强烈解码。评估时对时间状态、感觉信息、操作条件和异稳态机制的扰动会改变行为和/或内部预测。种子平衡的事件探针显示，关于未来接触、成功消耗和威胁事件的信息较弱但可测量，同时低能量解码很强。我们将这种模式解释为预测性异稳态组织的计算类比：一种分布式控制机制，具有预测性、能量敏感性、动作相关性，并且部分因果参与，但并未声称具有生物学验证或离散符号类别。
cs.NE / 2 / 2608.11865
Lapis: Laplacian Spiking Attention via First-Spike Timing and Membrane Leakage
Lapis：通过首脉冲时序与膜泄漏的拉普拉斯脉冲注意力
Abstract
Self-attention has become central to spiking vision transformers, yet its query-key scoring is still largely inherited from dense networks. Existing spiking variants either simplify dot product scoring or replace it with discrete operators, but spike timing, the native variable of a spiking network, does not directly define how tokens are related. We propose Lapis, a spiking attention mechanism that scores each token pair by the L1 distance between its query and key first-spike latency vectors under time-to-first-spike coding, and maps this distance to an affinity through a Laplacian kernel. The kernel's exponential decay matches the impulse response of a leaky integrate-and-fire membrane, so the accumulated latency difference determines the decay of a membrane trace, while row normalization reduces to a bit shift under power-of-two rounding. Scoring therefore needs only subtraction, absolute value, and accumulation, and removes all multiplication between query and key channels. Under a matched backbone and training schedule, Lapis reaches 96.56% top-1 accuracy on CIFAR-10, within 0.53 points of dot-product scoring. On ImageNet-1K, it reduces the estimated arithmetic energy of the attention path by 14.5x relative to dense dot-product attention. The deployed 6-bit model attains 83.25% top-1 accuracy at an estimated arithmetic energy of 3.28mJ per image.
Chinese Translation
自注意力已成为脉冲视觉Transformer的核心，然而其查询-键评分仍在很大程度上继承自稠密网络。现有的脉冲变体要么简化点积评分，要么用离散算子替代它，但脉冲时序——脉冲网络的原生变量——并未直接定义token之间的关联方式。我们提出Lapis，一种脉冲注意力机制，在首脉冲时间编码下，通过每个token对的查询与键首脉冲延迟向量之间的L1距离对其进行评分，并通过拉普拉斯核将该距离映射为亲和度。该核的指数衰减与泄漏积分-发放膜的冲激响应相匹配，因此累积的延迟差决定了膜迹的衰减，而行归一化在2的幂舍入下简化为位移操作。因此评分仅需要减法、绝对值和累加，并消除了查询与键通道之间的所有乘法。在匹配的主干网络和训练安排下，Lapis在CIFAR-10上达到96.56%的top-1准确率，与点积评分相差在0.53个百分点以内。在ImageNet-1K上，相对于稠密点积注意力，它将注意力路径的估计算术能量降低了14.5倍。部署的6位模型以每幅图像3.28mJ的估计算术能量达到83.25%的top-1准确率。
计算语言学 (cs.CL)
35
cs.CL / 1 / 2608.11362
RevCRN: Reversible Analog Computation using Chemical Reaction Networks
RevCRN：使用化学反应网络的可逆模拟计算
Abstract
The computability of real numbers and functions using Turing Machines has been a central area of theoretical computer science since the mid-20th century. In the late 20th century, it was shown that chemical reactions can serve as a basis for computation using the Chemical Reaction Network (CRN) model. Recent advances in computing real numbers using Deterministic Chemical Reaction Networks (DCRNs) have identified numerous classes of DCRN-computable real numbers. In parallel, the works of R. Landauer and C. H. Bennett, spanning the 1960s to the early 2000s, showed that reversible computing offers significant advantages over irreversible methods, particularly in energy efficiency, motivating extensive research on reversible computation. In this work, we investigate the computability of real numbers using Reversible Chemical Reaction Networks (RevCRNs). The paper has two primary contributions: (1) establishing relationships among CRN-computable real number classes including Lyapunov CRN ($\mathbb{R}_{LCRN}$), Real-Time CRN ($\mathbb{R}_{RTCRN}$), rational numbers ($\mathbb{Q}$), and RevCRNs ($\mathbb{R}_{RevCRN}$), with key results: (i) $\mathbb{Q}$ is a strict subset of $\mathbb{R}_{RevCRN}$; (ii) the set of positive algebraic numbers ($ALG$), $\mathbb{R}_{LCRN}$, and real numbers computable by 1-species RevCRN ($\mathbb{R}_{RevCRN}^{1s}$) are equal; (iii) $\mathbb{R}_{RTCRN}$ and $\mathbb{R}_{RevCRN}$ exhibit non-empty overlap; and (iv) the set of real numbers computable by detailed-balanced RevCRNs ($\mathbb{R}^{DetBal}_{RevCRN}$) is a subset of $ALG$; and (2) exploring the existence of a hierarchy within $\mathbb{R}_{RevCRN}$. Finally, we leave open the exact relationship between $\mathbb{R}_{RevCRN}$ and $\mathbb{R}_{RTCRN}$ while conjecturing a general hierarchy of RevCRN-computable reals.
Chinese Translation
自20世纪中叶以来，使用图灵机研究实数与函数的可计算性一直是理论计算机科学的一个核心领域。20世纪后期，研究表明，化学反应可以通过化学反应网络（CRN）模型作为计算的基础。近年来，在使用确定性化学反应网络（DCRN）计算实数方面的进展已经识别出众多DCRN可计算实数的类。与此同时，R. Landauer和C. H. Bennett从20世纪60年代到21世纪初的工作表明，可逆计算相较于不可逆方法具有显著优势，尤其是在能效方面，这推动了对可逆计算的广泛研究。在本工作中，我们研究使用可逆化学反应网络（RevCRN）计算实数的可计算性。本文有两个主要贡献：（1）建立CRN可计算实数类之间的关系，这些类包括Lyapunov CRN（$\mathbb{R}_{LCRN}$）、实时CRN（$\mathbb{R}_{RTCRN}$）、有理数（$\mathbb{Q}$）和RevCRN（$\mathbb{R}_{RevCRN}$），并得到关键结果：（i）$\mathbb{Q}$是$\mathbb{R}_{RevCRN}$的一个真子集；（ii）正代数数集合（$ALG$）、$\mathbb{R}_{LCRN}$以及由单物种RevCRN可计算的实数（$\mathbb{R}_{RevCRN}^{1s}$）是相等的；（iii）$\mathbb{R}_{RTCRN}$和$\mathbb{R}_{RevCRN}$具有非空交集；（iv）由细致平衡RevCRN可计算的实数集合（$\mathbb{R}^{DetBal}_{RevCRN}$）是$ALG$的一个子集；（2）探索$\mathbb{R}_{RevCRN}$内部层级的存在性。最后，我们留待解决$\mathbb{R}_{RevCRN}$与$\mathbb{R}_{RTCRN}$之间的精确关系，同时推测RevCRN可计算实数的一个一般层级。
cs.CL / 2 / 2608.11332
Gloss-Free Representation Learning for Cross-Dataset Sign Spotting
面向跨数据集手语定位的无注释表示学习
Abstract
Sign-language research for resource-constrained languages is often limited by the cost of dense linguistic labels such as glosses, temporal boundaries, and sign order. Broadcast news offers a practical alternative by pairing continuous signing with spoken-language transcripts, but this supervision is weak since text and signing are loosely aligned. Morphologically rich languages such as Turkish add further difficulty, as the same lexical meaning can appear in many inflected forms while some derived forms should remain distinct. We study whether weak transcript-based supervision can pretrain a reusable sign encoder in this setting, where poor text normalization can fragment pseudo-gloss targets and weaken representation learning. Unlike prior pseudo-gloss pipelines designed mainly to improve translation, we test whether the pretrained encoder transfers as a reusable representation for cross-dataset sign spotting. We pretrain on TSL-News, a new Turkish broadcast corpus, using pseudo-gloss labels derived from transcripts rather than manual annotation, comparing rule-based morphological lemmatization with constrained LLM-assisted normalization over a fixed vocabulary. We evaluate the learned representations via cross-dataset sign spotting on a new TSL Spotting Benchmark built from the TSL Dictionary corpus. The LLM-assisted encoder raises top-5 temporal localization mean IoU from 0.235 to 0.465, with 56.2% of examples reaching an IoU of at least 0.50; a frequency analysis suggests this gain is not mainly driven by memorizing frequent pseudo-gloss labels. In a downstream translation check, the same pretraining improves BLEU-4 from 9.60 to 11.04 and ROUGE from 23.48 to 27.43. These results show that loosely aligned broadcast data can provide effective weak supervision for learning sign representations that capture both lexical content and temporal structure.
Chinese Translation
对于资源受限的语言，手语研究往往受限于密集语言标注的成本，例如词汇注释、时间边界和手语顺序。广播新闻通过将连续手语与口语转录文本配对提供了一种实用的替代方案，但这种监督较弱，因为文本与手语之间的对齐是松散的。像土耳其语这样形态丰富的语言进一步增加了难度，因为同一词汇意义可能以许多屈折形式出现，而某些派生形式又应保持区分。我们研究在这种设置下，基于转录文本的弱监督能否预训练一个可复用的手语编码器，其中糟糕的文本规范化可能会使伪注释目标碎片化并削弱表示学习。与以往主要为改进翻译而设计的伪注释流程不同，我们测试预训练编码器是否能够作为一种可复用表示迁移到跨数据集手语定位中。我们在TSL-News（一个新的土耳其语广播语料库）上进行预训练，使用从转录文本中派生的伪注释标签而非人工标注，并在固定词汇表上比较基于规则的形态学词形还原与受约束的大语言模型辅助规范化。我们通过在基于TSL词典语料库构建的新TSL定位基准上进行跨数据集手语定位，来评估所学到的表示。LLM辅助的编码器将top-5时间定位平均IoU从0.235提高到0.465，其中56.2%的样本达到了至少0.50的IoU；频率分析表明，这一提升并非主要由记忆频繁伪注释标签所驱动。在下游翻译检验中，相同的预训练将BLEU-4从9.60提高到11.04，将ROUGE从23.48提高到27.43。这些结果表明，松散对齐的广播数据能够为学习同时捕获词汇内容和时间结构的手语表示提供有效的弱监督。
cs.CL / 3 / 2608.11338
Better, Faster, Stronger: Programmatic Skill Learning Best Reduces Agent Cost
更好、更快、更强：程序化技能学习最能降低智能体成本
Abstract
Recently, the practice of augmenting LLM agent capability with skills has gained prevalence. We explore the cost effective adaptation of agents to novel domains by means of learning skills. Existing works focus on performance gain over cost effectiveness. As a result, little is known about what skill learning strategies save cost. We argue that among all the different skill learning methods, those that view skills as programs can achieve the best cost reduction. By executing sequences of actions deterministically, a program-augmented agent can reliably and cheaply achieve goals that would otherwise require trial and error and risk degenerate behavior over long horizons. An agent can learn at inference time by incrementally discovering these programs and equipping them for future tasks. We hypothesize that past trajectories contain enough signal to guide skill learning, even without replay or validation, provided the agent can learn to analyze them. To test our claims, we propose SpeedRunner, a coding agent that analyzes trajectories and refactors skills for better performance on future tasks. Across three different embodied environments, we show that SpeedRunner consistently achieves the frontier in learning and cost reduction while remaining robust against distribution shifts and environmental randomness.
Chinese Translation
近来，利用技能增强 LLM 智能体能力的做法已变得普遍。我们探索通过技能学习实现智能体对新领域的成本有效适应。现有研究侧重于性能提升，而非成本效益。因此，关于哪些技能学习策略能节省成本，目前知之甚少。我们认为，在所有不同的技能学习方法中，那些将技能视为程序的方法能够实现最佳的成本降低。通过确定性地执行动作序列，经过程序增强的智能体能够可靠且低成本地实现目标，而这些目标否则需要试错，并在长时间范围内面临行为退化风险。智能体可以在推理时学习，逐步发现这些程序并装备它们用于未来任务。我们假设，过去的轨迹包含足够的信号来指导技能学习，即使没有回放或验证，只要智能体能够学会分析这些轨迹。为了验证我们的主张，我们提出了 SpeedRunner，一个编码智能体，它分析轨迹并重构技能，以便在未来任务中获得更好的表现。在三个不同的具身环境中，我们表明 SpeedRunner 始终达到学习和成本降低的前沿水平，同时对分布偏移和环境随机性保持鲁棒。
cs.CL / 4 / 2608.11350
Self-Evolving Embodied Agents via Skill-Harness Evolution
通过技能-执行框架进化实现自进化的具身智能体
Abstract
Embodied agents are increasingly built as systems around foundation models, where performance depends not only on model weights but also on the skills, context, action interfaces, and execution harness surrounding the model. While supervised fine-tuning and reinforcement learning can adapt agents to new environments, they require additional data, rewards, and training runs; meanwhile, many train-free code-centric approaches rely on programmable robot APIs that may be unavailable in fixed-interface settings. We propose SHAPER, a self-evolving framework for train-free embodied adaptation that keeps model parameters frozen and improves the non-parametric agent system by evolving reusable skills and a context-code harness through target-environment rollouts. In SHAPER, the same frozen model can serve as both planner and optimizer, refining its external skills and context-code harness without parameter updates. We evaluate SHAPER on VLABench and ESI-Bench, covering embodied agents with different low-level action interfaces, and compare against pure execution, supervised fine-tuning, and test-time-scaling baselines such as verifier-free selection and voting. Our results suggest that skill-and-harness optimization is a practical route to self-evolving embodied agents when model training is expensive, unavailable, or undesirable.
Chinese Translation
具身智能体越来越以基础模型为核心构建为系统，其性能不仅取决于模型权重，还取决于围绕模型的技能、上下文、动作接口和执行框架。虽然监督微调和强化学习可以使智能体适应新环境，但它们需要额外的数据、奖励和训练过程；同时，许多免训练的以代码为中心的方法依赖于可编程机器人API，而这些API在固定接口环境中可能不可用。我们提出 SHAPER，一个用于免训练具身适应的自进化框架，该框架保持模型参数冻结，并通过在目标环境中推演来进化可复用技能和上下文-代码执行框架，从而改进非参数化智能体系统。在 SHAPER 中，同一个冻结模型既可作为规划器，也可作为优化器，在无需参数更新的情况下完善其外部技能和上下文-代码执行框架。我们在 VLABench 和 ESI-Bench 上评估 SHAPER，覆盖具有不同底层动作接口的具身智能体，并与纯执行、监督微调以及测试时扩展基线（如无验证器选择和投票）进行比较。我们的结果表明，当模型训练昂贵、不可用或不受欢迎时，技能与执行框架优化是实现自进化具身智能体的一条实用途径。
cs.CL / 5 / 2608.11352
ODE-Based Transformer Decoders for Iterative Sign Language Translation
用于迭代手语翻译的基于ODE的Transformer解码器
Abstract
Sign language translation has achieved strong results with Transformer architectures, yet recent improvements largely rely on scaling model capacity at the cost of increased computation. We propose a parameter-efficient alternative that improves expressiveness without increasing model size. Rather than scaling capacity, we focus on enhancing the update dynamics of iterative refinement decoders, where each refinement step corresponds to one internal decoder iteration that progressively improves the latent representation before translation generation. We reinterpret residual refinement updates from an Ordinary Differential Equation (ODE) perspective and replace them with higher-order numerical integration schemes, namely Runge--Kutta methods (RK-2 and RK-4). These methods perform multiple function evaluations within each refinement step to produce more accurate and stable representation updates without adding decoder parameters. To the best of our knowledge, this is the first application of ODE-inspired update dynamics to sign language translation. RK-2 achieves 22.96 BLEU-4 on the PHOENIX-2014-T test set and 19.34 BLEU-4 on the CSL-Daily test set, outperforming the IPSLT baseline on both benchmarks, with fewer decoder layers and refinement iterations on CSL-Daily. These results suggest that stronger refinement dynamics can improve translation performance under parameter-efficient decoder designs, providing a complementary alternative to conventional model scaling.
Chinese Translation
手语翻译在使用Transformer架构方面已取得了显著成果，但近年来的改进在很大程度上依赖于以增加计算量为代价来扩展模型容量。我们提出一种参数高效的替代方案，在不增加模型规模的情况下提升表达能力。与其扩展容量，我们专注于增强迭代细化解码器的更新动力学，其中每个细化步骤对应一次内部解码器迭代，在生成翻译之前逐步改进潜在表示。我们从常微分方程（ODE）的视角重新解释残差细化更新，并将其替换为高阶数值积分格式，即Runge--Kutta方法（RK-2和RK-4）。这些方法在每个细化步骤内执行多次函数评估，以产生更准确、更稳定的表示更新，而无需增加解码器参数。据我们所知，这是首次将受ODE启发的更新动力学应用于手语翻译。RK-2在PHOENIX-2014-T测试集上达到22.96 BLEU-4，在CSL-Daily测试集上达到19.34 BLEU-4，在两个基准上均优于IPSLT基线，且在CSL-Daily上使用更少的解码器层和细化迭代。这些结果表明，更强的细化动力学可以在参数高效的解码器设计下提升翻译性能，为传统模型扩展提供一种互补的替代方案。
cs.CL / 6 / 2608.11426
Is Convergence Inevitable? Tracing Output Homogeneity Back to Base Models
收敛是不可避免的吗？将输出同质性追溯至基础模型
Abstract
The lack of diversity in LM content is widely attributed to the alignment process, but how and where exactly in the pipeline this collapse begins is unknown. We argue that output homogeneity is likely learned during the pretraining phase, and only \emph{revealed} or magnified during the alignment process. Specifically, we find that semantic convergence is observed from the first alignment stage--the instruction-tuning phase (SFT)--suggesting that homogeneity might already exist in the pre-alignment model. To investigate this, we conduct controlled SFT experiments examining how training data influences output convergence on specific input/output pairs. We find that convergence can be revealed and amplified, but not introduced by the SFT data, supporting its role as a catalyst rather than a cause. To further test whether homogeneity originates before alignment, we measure convergence in base models. We find that instruct-like collapse can be induced through prompting alone, even without alignment. Taken together, our results suggest that semantic convergence may arise naturally from the objectives underlying LM training, making it difficult to mitigate through post-alignment interventions alone.
Chinese Translation
语言模型（LM）内容缺乏多样性被广泛归因于对齐过程，但这种崩溃究竟在流程中的何处以及如何开始仍属未知。我们认为，输出同质性很可能在预训练阶段就已经被学到，并且在对齐过程中只是被揭示或被放大。具体而言，我们发现语义收敛从第一个对齐阶段——指令微调阶段（SFT）——即可观察到，这表明同质性可能已经存在于对齐前的模型中。为探究这一点，我们进行了受控SFT实验，考察训练数据如何影响特定输入/输出对上的输出收敛。我们发现，收敛可以被揭示和放大，但不能由SFT数据引入，这支持其角色是催化剂而非原因。为进一步检验同质性是否起源于对齐之前，我们测量了基础模型中的收敛。我们发现，即使没有对齐，仅通过提示也可以诱导出类指令式的崩溃。综合来看，我们的结果表明，语义收敛可能自然而然地源自语言模型训练背后的目标，这使得仅靠对齐后干预难以缓解。
cs.CL / 7 / 2608.11433
Stigma and Support in Online Sexual Violence Narratives on Reddit
Reddit上在线性暴力叙事中的污名与支持
Abstract
Online communities increasingly provide spaces where survivors of sexual violence can share their experiences and seek support. Although prior research has examined stigma and social support separately, less is known about how stigma expressed in survivor narratives relates to the support offered in response. We introduce the SCOPE dataset, linking stigma signals in online survivor narratives to support types in corresponding comment threads. We annotate posts using a multi-dimensional stigma taxonomy, including Experienced, Internalized, Anticipated, and Structural Stigma, and comments using a support taxonomy encompassing Information Support, Emotional Support, Esteem Support, Tangible Assistance, and Group Interaction. Using contextual, linguistic, and emotion analyses, we compare Stigma and No Stigma content and find that Stigma narratives place greater emphasis on internalized distress, whereas No Stigma narratives focus more on interpreting situations and experiences. Internalized Stigma is the most prevalent category, and community responses remain broadly stable across stigma types, with Information and Esteem Support appearing most often. These findings show how stigma shapes survivor narratives and peer responses and have implications for computational modeling, content moderation, and safer online systems.
Chinese Translation
在线社区日益为性暴力幸存者提供分享其经历和寻求支持的空间。尽管先前的研究已分别考察了污名和社会支持，但人们对幸存者叙事中表达的污名如何与回应中提供的支持相关联了解较少。我们引入SCOPE数据集，将在线幸存者叙事中的污名信号与相应评论线程中的支持类型联系起来。我们使用一个多维污名分类法对帖子进行标注，包括经历污名、内化污名、预期污名和结构性污名；并使用一个涵盖信息支持、情感支持、自尊支持、有形援助和群体互动的支持分类法对评论进行标注。通过语境分析、语言分析和情感分析，我们比较了有污名内容与无污名内容，并发现有污名叙事更强调内化困扰，而无污名叙事更侧重于解释情境和经历。内化污名是最普遍的类别，社区回应在不同污名类型中大致保持稳定，其中信息支持和自尊支持出现得最为频繁。这些发现揭示了污名如何塑造幸存者叙事和同伴回应，并对计算建模、内容审核和更安全的在线系统具有启示意义。
cs.CL / 8 / 2608.11441
DonorRank: Donor Language Selection for Low-Resource Cross-Lingual Speech Recognition
DonorRank：面向低资源跨语言语音识别的捐赠语言选择
Abstract
Low-resource automatic speech recognition (ASR) commonly relies on cross-lingual transfer, where models are adapted from higher-resource donor languages. However, selecting donors remains challenging for spontaneous speech from under-resourced language communities, due to linguistic variation, evolving orthographic conventions, and uneven resource availability. We present DonorRank, a learning-to-rank framework for predicting effective donor languages for zero-shot ASR. We evaluate DonorRank on two multilingual speech corpora of Indic and African language families. It accurately predicts donor language rankings and improves donor selection over common heuristics based on genetic similarity or high-resource languages. Beyond improving transfer, we show how DonorRank is a general framework for analyzing donor language selection itself. Our analyses show that the composition of the donor set determines which linguistic cues are useful in predicting successful transfer. We also identify transfer patterns that provide practical guidance for multilingual ASR in low-resource settings.
Chinese Translation
低资源自动语音识别（ASR）通常依赖跨语言迁移，即从资源较丰富的捐赠语言中适配模型。然而，对于资源匮乏语言社区的自发言语，选择捐赠语言仍然具有挑战性，原因包括语言变异、不断演变的正字法规范以及资源分布不均。我们提出 DonorRank，一个学习排序框架，用于为零样本 ASR 预测有效的捐赠语言。我们在印度语系和非洲语系的两个多语言语音语料库上评估 DonorRank。它准确预测了捐赠语言排名，并且相比基于遗传相似性或高资源语言的常见启发式方法，改善了捐赠语言的选择。除了改善迁移之外，我们还展示了 DonorRank 如何作为一个通用框架来分析捐赠语言选择本身。我们的分析表明，捐赠语言集合的组成决定了哪些语言线索有助于预测成功的迁移。我们还识别出了一些迁移模式，为低资源环境下的多语言 ASR 提供了实用指导。
cs.CL / 9 / 2608.11528
Group Alignment-Induced Sycophancy: A Two-Sided Evaluation of Steerable Pluralistic Alignment
群体对齐诱导的谄媚：可引导多元对齐的双面评估
Abstract
Group alignment adapts a language model to a demographic group to produce responses that reflect the group's opinions, values, and preferences. Sycophancy, a well-documented by-product of alignment, causes the model to over-agree with the user regardless of factual and objective information. However, existing group alignment methods and evaluations focus only on how closely the model matches the group's opinions, overlooking the induced change in sycophantic behaviour. To bridge this gap, we introduce \textbf{G}roup \textbf{A}lignment-induced \textbf{S}ycophancy (GAS) and systematically evaluate alignment across 3 methods, 4 models and 13 demographic groups, on both the intended gain in opinion alignment and the unintended shift in sycophancy. We find that gain and shift are non-uniform across groups: under an identical budget, some groups receive larger gains in opinion alignment than others, and the induced sycophancy shift forms a group-specific profile rather than a single-dimensional change. These results suggest that group alignment should be reported as a two-sided, multi-dimensional profile rather than a single fit score that accounts for per-group differences when adapting LLMs to diverse populations.
Chinese Translation
群体对齐使语言模型适应某个人口群体，以产生反映该群体意见、价值观和偏好的回答。谄媚作为对齐的一个已有充分记录的副产品，导致模型过度迎合用户，而不顾事实性和客观信息。然而，现有的群体对齐方法和评估仅关注模型与群体意见的匹配程度，忽视了所引发的谄媚行为变化。为了弥补这一不足，我们引入 \textbf{G}roup \textbf{A}lignment-induced \textbf{S}ycophancy (GAS)，并在3种方法、4个模型和13个人口群体上系统评估对齐，同时考察预期的意见对齐增益和非预期的谄媚偏移。我们发现增益和偏移在不同群体之间并不一致：在相同预算下，一些群体在意见对齐上获得的增益大于其他群体，并且所引发的谄媚偏移形成了特定于群体的特征，而非单一维度的变化。这些结果表明，群体对齐应被报告为一个双面的、多维度的特征，而非单一拟合分数，以便在将大语言模型适配到多样化人群时考虑各群体差异。
cs.CL / 10 / 2608.11531
On Weak Bisimilarities in CCSK
论CCSK中的弱互模拟
Abstract
In the context of CCSK, a reversible extension of CCS, we study different notions of bisimilarity (strong/weak, forward-only/reversible) and highlight their differences and commonalities. In particular, for the weak reversible case, not previously studied in the literature, we propose two variants, dubbed directional and mixed bisimilarity, depending on whether $τ$ actions should be in the same direction (forward/backward) as the action being matched or not. We show, in particular, that mixed bisimilarity is a congruence and completely abstracts away from $τ$ actions.
Chinese Translation
在CCSK（CCS的一个可逆扩展）的背景下，我们研究了不同的互模拟概念（强/弱、仅前向/可逆），并突出它们之间的差异和共同点。特别是，对于文献中此前未被研究过的弱可逆情形，我们提出了两种变体，分别称为方向互模拟和混合互模拟，其区别取决于 $τ$ 动作是否应与被匹配的动作处于相同方向（前向/后向）。我们特别证明了，混合互模拟是一个同余关系，并且完全抽象掉 $τ$ 动作。
cs.CL / 11 / 2608.11534
CT-$Δ$Bench: A Benchmark for Longitudinal 3D Medical Imaging Difference Reporting with Vision-Language Models
CT-$Δ$Bench：面向视觉-语言模型的纵向3D医学影像差异报告基准
Abstract
In medical imaging, the clinical value of Computed Tomography (CT) lies not only in depicting current disease status, but crucially in enabling longitudinal comparison of serial scans to determine disease evolution, a process that underpins response assessment, recurrence detection, and ongoing patient management. Yet, despite this central role of temporal comparison in clinical decision-making, existing medical foundation models remain largely confined to single-study understanding, leaving temporally grounded cross-examination insufficiently addressed. To address this gap, we study longitudinal imaging difference reporting, a task in which a model takes two temporally separated scans from the same patient and generates a clinically meaningful report describing interval changes between them. We introduce CT-$Δ$Bench, a dedicated benchmark for this task with patient-level splitting to prevent information leakage. To better evaluate this task beyond surface-level text similarity, we further develop change-aware metrics specifically designed to capture clinically meaningful longitudinal changes, and conduct an independent physician validation to assess the reliability of the synthesized references and event extraction pipeline. We also compare direct paired-CT reasoning with an indirect two-stage pipeline that first generates single-timepoint reports and then performs textual differencing. Finally, we propose DeltaMed, a baseline model for direct paired-CT difference reporting, and train it on the benchmark training set. Together, these contributions lay the groundwork for temporally aware medical foundation models that better reflect real-world longitudinal clinical reasoning.
Chinese Translation
在医学影像中，计算机断层扫描（CT）的临床价值不仅在于描述当前疾病状态，更关键的是能够对连续扫描进行纵向比较以判断疾病演变，这一过程是疗效评估、复发检测和持续患者管理的基础。然而，尽管时间比较在临床决策中具有这一核心作用，现有的医学基础模型在很大程度上仍局限于单次检查的理解，导致基于时间的交叉比对尚未得到充分解决。为了解决这一空白，我们研究了纵向影像差异报告，该任务要求模型接受同一患者两次时间上分离的扫描，并生成描述二者之间间隔变化的具有临床意义的报告。我们引入了 CT-$Δ$Bench，这是一个专门用于该任务的基准，采用患者级别的划分以防止信息泄漏。为了更好地在表面文本相似度之外评估该任务，我们进一步开发了专门设计用于捕捉具有临床意义的纵向变化的变化感知指标，并进行独立的医生验证，以评估合成参考和事件提取流程的可靠性。我们还比较了直接配对CT推理与一种间接两阶段流程，该流程首先生成单时间点报告，然后进行文本差异分析。最后，我们提出了 DeltaMed，一个用于直接配对CT差异报告的基线模型，并在基准训练集上对其进行训练。总之，这些贡献为具有时间感知能力的医学基础模型奠定了基础，使其能够更好地反映现实世界中的纵向临床推理。
cs.CL / 12 / 2608.11552
Beyond Single-Turn Confidence: Trajectory-Adapted Uncertainty Quantification for LLM Agents
超越单轮置信度：面向LLM智能体的轨迹自适应不确定性量化
Abstract
Uncertainty quantification (UQ) methods for language models are typically evaluated on single-turn outputs, where uncertainty is attached to one generated answer. For LLM agents, however, the unit of observation is an interactive trajectory, where the model can ask clarifying questions, call tools, update state, and make intermediate decisions whose errors propagate to the final outcome. We study whether three common families of single-turn UQ methods transfer to this setting. Across five LLMs and four multi-turn tool-use datasets from BFCL-v4 and $τ^2$-bench, we evaluate white-box scorers based on action-token probabilities, black-box consistency scorers based on resampled trajectories, and reflexive scorers based on model self-assessment of the trajectory. We find that transfer is often useful but uneven. Token-probability scores are highly sensitive to the choice of aggregator used across turns, reflexive scores provide the strongest low-cost baseline in most evaluated settings, and black-box self-consistency is often the strongest UQ family, with trajectory-equivalence and action-set consistency typically ranking highest among its variants. These results suggest that UQ methods developed for single generations should be revalidated at the trajectory level, with careful attention to the consistency measurement, aggregator choice, and computational budget.
Chinese Translation
语言模型的不确定性量化（UQ）方法通常在单轮输出上进行评估，其中不确定性附着于一个生成的答案。然而，对于LLM智能体而言，观察单位是交互式轨迹，其中模型可以提出澄清问题、调用工具、更新状态，并做出中间决策，其错误会传播到最终结果。我们研究三类常见的单轮UQ方法是否能迁移到这一设定。在来自BFCL-v4和$τ^2$-bench的五个LLM和四个多轮工具使用数据集上，我们评估了基于动作token概率的白盒评分器、基于重采样轨迹的黑盒一致性评分器，以及基于模型对轨迹的自我评估的反思评分器。我们发现，这种迁移通常有用，但并不均衡。Token概率分数对跨轮使用的聚合器选择高度敏感，反思分数在大多数被评估的设置中提供了最强的低成本基线，而黑盒自洽性通常是最强的UQ家族，其中轨迹等价性和动作集一致性通常在其变体中排名最高。这些结果表明，为单次生成开发的UQ方法应在轨迹层面重新验证，并仔细关注一致性度量、聚合器选择和计算预算。
cs.CL / 13 / 2608.11629
Easper: An Accessible ASR Pipeline for Language Documentation
Easper：一种面向语言记录的无障碍 ASR 流水线
Abstract
Audio transcription is a critical bottleneck in language documentation. While multilingual Automatic Speech Recognition (ASR) models like Whisper offer solutions, field linguists often lack the expertise to utilise them. We present Easper, an open-source, no-code workflow enabling linguists to iteratively fine-tune ASR models via cloud resources directly from ELAN annotations. Deploying ASR also raises a cold start problem: deciding which recordings to transcribe first to bootstrap an accurate model. Using Easper, we evaluate transcription prioritisation strategies on three Vanuatu languages (Bislama, Nafsan, Nguna). We fine-tune models by recording session, comparing Character Error Rate trajectories when prioritising acoustic cleanliness versus linguistic richness. We demonstrate that prioritising lexically rich narratives and increasing acoustic-phonetic repetition, even in noisy environments, leads to faster improvements in transcription quality.
Chinese Translation
音频转写是语言记录中的一个关键瓶颈。虽然像 Whisper 这样的多语言自动语音识别（ASR）模型提供了解决方案，但田野语言学家往往缺乏使用它们所需的专业知识。我们提出了 Easper，这是一个开源、无代码的工作流，使语言学家能够直接基于 ELAN 标注通过云资源迭代微调 ASR 模型。部署 ASR 还带来了冷启动问题：决定首先转写哪些录音，以引导构建一个准确的模型。利用 Easper，我们在三种瓦努阿图语言（比斯拉马语、纳夫桑语、恩古纳语）上评估了转写优先级排序策略。我们按录音会话微调模型，比较在优先考虑声学清洁度与语言丰富度时的字符错误率轨迹。我们证明，优先处理词汇丰富的叙述并增加声学-语音重复，即使在嘈杂环境中，也能带来转写质量的更快提升。
cs.CL / 14 / 2608.11694
The Wording Effect: Quantifying Two-Way Drift in LLM Benchmark Performance
措辞效应：量化LLM基准性能中的双向漂移
Abstract
A benchmark score comes from a single phrasing of each problem. That single phrasing is treated as if it stood for the whole space of ways the same problem could be asked, but it does not. We show that rephrasing a problem while keeping its meaning and answer fixed routinely flips a model's answer in both directions, so some failures become successes and some successes become failures. We call this drift. BenchDrift generates meaning-preserving variations of benchmark problems along four axes, namely linguistic, referential, pragmatic, and structural, and measures how often, and why, correctness flips under each. Across eight models and three benchmarks (GSM8K, MMLU, MATH-Hard), we observe that drift is large in both directions. Two findings stand out. First, phrasing sensitivity does not fade as models get better. Instead, it changes sign. Weak models gain more from rephrasing than they lose, while strong models lose far more than they gain. We find that the best models on a benchmark are therefore the ones whose scores depend most on the wording they happened to be given. Second, the models largely agree on which rephrasings cost the most correct answers even though they differ in how much they drift, so fragility belongs to the rephrasing and not to the model. Furthermore, rephrasing breaks answers a model was confident about, whether the problem is made shorter or longer. Code and Data: https://github.com/IBM/BenchDrift/tree/demo-ui
Chinese Translation
基准分数来自每个问题的一种单一措辞。这种单一措辞被当作能代表同一问题所有可能问法的整个空间，但它并不能。我们表明，在保持问题含义和答案不变的情况下改写问题，会经常使模型的答案在两个方向上发生翻转，因此一些失败变为成功，一些成功变为失败。我们将此称为漂移。BenchDrift 沿四个轴生成基准问题的保持含义的变体，即语言性、指称性、语用性和结构性，并测量在每个轴下正确性翻转的频率及原因。在八个模型和三个基准（GSM8K、MMLU、MATH-Hard）上，我们观察到漂移在两个方向上都很显著。有两个发现尤为突出。首先，随着模型变得更好，措辞敏感性并未减弱。相反，它改变了符号。弱模型从改写中获得的比失去的多，而强模型失去的远多于获得的。因此我们发现，在一个基准上最好的模型，其分数恰恰最依赖于它们碰巧被给予的措辞。其次，模型在很大程度上对哪些改写导致最多正确答案损失意见一致，尽管它们在漂移程度上有所不同，因此脆弱性属于改写本身，而不属于模型。此外，改写会破坏模型原本有信心的答案，无论问题是被改短还是改长。代码与数据：https://github.com/IBM/BenchDrift/tree/demo-ui
cs.CL / 15 / 2608.11767
Causal Structure is Inducible but Functionally Decoupled: The Routing/Readout Boundary of a Typed Mechanism Library
因果结构可被诱导但功能上解耦：类型化机制库的路由/读出边界
Abstract
When a language model answers an interventional question, the computation it must perform depends on the type of evidence the query requires. We report a decoupling in how a transformer organizes causal knowledge: slot-by-type structure induced by type-level supervision organizes routing, yet remains functionally decoupled from answer readout. We establish this with a typed mechanism library -- discrete mechanism slots partitioned by evidence type, auditable at the state level -- on a causal-world benchmark with exact interventional ground truth, under a frozen protocol, at two scales (22.6M and 125M). Four preregistered findings. (i) Origin. Slot-by-type organization is induced by type-level supervision: absent in architecturally identical unsupervised controls, not buyable by content-free gating labels, and statistically attributable to the supervision signal, replicating at 125M under a powered preregistered protocol (all nine cells passed). (ii) Boundary. The induced structure is a typed routing index with a sharp routing/readout boundary: slot codes scaffold routing but do not drive answer readout ($|Δ\hat{y}| \le 3.4\times10^{-6}$, zero collateral, three seeds, stable across a 5.6x scale window) -- we therefore make no behavioral-editability claim. (iii) Cost. The structure is free: LM quality matches a parameter-matched monolith within 0.0082 nats. (iv) Trust. The library state is exactly local under edit and bit-exactly revertible -- 250 single-edit and 1,000 stacked reverts per seed, zero failures. We further find that the unsupervised null itself moves with scale, so comparisons reusing a null calibrated at one scale may be confounded at another. Every claim is tied to a preregistered, machine-checkable criterion archived before the data it governs; the full audit trail, including one criterion we failed and how the frozen protocol handled it, is released as an appendix.
Chinese Translation
当语言模型回答一个干预性问题时，它必须执行的计算取决于查询所需证据的类型。
我们报告了 transformer 组织因果知识的一种解耦：按类型分槽的结构由类型级监督诱导，负责组织路由，但在功能上仍与答案读出解耦。
我们通过一个类型化机制库——按证据类型划分的离散机制槽，可在状态层面审计——在具有精确干预真值的因果世界基准上、在冻结协议下、以两个规模（22.6M 和 125M）确立这一点。
四项预先注册的发现。
(i) 来源。按类型分槽的组织由类型级监督诱导：在架构相同的无监督对照组中不存在，无法通过无内容的门控标签获得，并且在统计上可归因于监督信号，在一项有足够检验力的预先注册协议下于 125M 规模复现（全部九个单元通过）。
(ii) 边界。所诱导的结构是一个类型化路由索引，具有鲜明的路由/读出边界：槽编码支撑路由，但不驱动答案读出（$|Δ\hat{y}| \le 3.4\times10^{-6}$，零附带影响，三个种子，在 5.6 倍规模窗口内稳定）——因此我们不提出行为可编辑性声明。
(iii) 成本。该结构是免费的：语言模型质量与参数匹配的单体模型相差在 0.0082 纳特以内。
(iv) 可信性。库状态在编辑下严格局部，且可逐位精确回滚——每个种子 250 次单次编辑和 1,000 次堆叠回滚，零失败。
我们进一步发现，无监督零假设本身会随规模变化，因此在一个规模上校准的零假设被复用于另一个规模时，比较可能会产生混杂。
每项声明都与预先注册、机器可检查的标准相关联，该标准在它所管辖的数据之前存档；完整审计轨迹——包括我们未能通过的一项标准以及冻结协议如何处理它——作为附录发布。
cs.CL / 16 / 2608.11772
Diagnosis Before Recovery: Turning Agent Failures into Selective Self-Correction
先诊断后恢复：将智能体失败转化为选择性自我修正
Abstract
Self-correction is particularly useful when a failure constrains the next repair. Coding agents benefit from this property because compilers, tests, and execution traces turn many failures into typed recovery signals, but broad language-agent tasks often expose only a coarse task failure. This creates a tension for generic recovery playbooks: they broaden the agent's context precisely when the system needs a narrower repair interface, mixing incompatible signals for invalid actions, missing procedures, and strict-format errors. Our insight is that development-set failures can recover part of the missing diagnostic substrate by deciding which recovery interventions are admissible before test-time correction. We propose DARC, a diagnosis-guided recovery harness that profiles task-family failure modes, prunes mismatched interventions from a shared recovery library, and freezes a verifier-selected success-cost policy for deployment. This causal order makes correction selective: the harness first determines what kind of failure can be repaired, then decides how much recovery evidence to spend. In ALFWorld, AppWorld, and XBRL Finance, the same protocol yields an action-validity harness, a procedural-recovery fallback, and a format-precision retrieval policy; in each evaluated setting it improves average task performance over base agents and broad playbooks while reducing environment steps or retrieval budget. Our experiments show that failures need not trigger uniformly more context: DARC turns self-correction from prompt expansion into recovery-interface design. DARC provides a practical route toward more reliable agents in domains where compiler-like feedback is absent: making failures actionable before making contexts larger.
Chinese Translation
当失败约束了下一次修复时，自我修正尤其有用。编码智能体从这一特性中受益，因为编译器、测试和执行轨迹将许多失败转化为类型化的恢复信号，但广泛的语言智能体任务通常只暴露粗粒度的任务失败。这给通用恢复策略手册带来了张力：它们在系统恰好需要更窄的修复接口时扩大了智能体的上下文，将无效动作、缺失程序和严格格式错误的不兼容信号混合在一起。我们的见解是，开发集上的失败可以通过在测试时修正之前决定哪些恢复干预是可采纳的，来恢复部分缺失的诊断基础。我们提出了 DARC，一个诊断引导的恢复框架，该框架刻画任务族失败模式，从共享恢复库中剪除不匹配的干预，并冻结由验证器选择的成功-成本策略以供部署。这种因果顺序使修正具有选择性：该框架首先确定哪种失败可以被修复，然后决定花费多少恢复证据。在 ALFWorld、AppWorld 和 XBRL Finance 中，同一协议产生了动作有效性框架、程序性恢复回退和格式精确检索策略；在每个被评估的设置中，它相较于基础智能体和宽泛策略手册提高了平均任务性能，同时减少了环境步数或检索预算。我们的实验表明，失败并不必然一律触发更多上下文：DARC 将自我修正从提示扩展转变为恢复接口设计。DARC 为在缺乏类似编译器的反馈的领域中实现更可靠的智能体提供了一条实用路径：在扩大上下文之前使失败变得可操作。
cs.CL / 17 / 2608.11786
Language-Conditional Dequantization: Recovering What Quantization Steals from Non-English Languages
语言条件反量化：恢复量化从非英语语言中窃取的内容
Abstract
Aggressive quantization disproportionately harms multilingual capability: in the sub-4B INT3 GPTQ regime, we measure 2-4x larger perplexity degradation on non-English languages than on English. We propose Language-Conditional Dequantization (LCD), a post-hoc method that attaches per-language rank-2 LoRA corrections to the linear layers of an already-quantized model, adding 0.12% parameters per language and training in under 20 minutes on a single GPU. Across Qwen2.5-3B and Llama-3.2-3B, LCD recovers 70-83% of the perplexity gap for non-Latin script languages and 17-28% of the GlobalMMLU accuracy gap, outperforming a language-agnostic correction of equal capacity by 3-9 points on typologically distant languages and a data-free low-rank baseline (LQER) by an order of magnitude. We further identify a perplexity-accuracy disconnect and trace it to where quantization concentrates damage: early-depth errors (Llama) propagate downstream and resist local correction, while late-depth errors (Qwen) do not. A layer-restricted variant of LCD validates this mechanism directly.
Chinese Translation
激进量化对多语言能力造成不成比例的损害：在 sub-4B INT3 GPTQ 区间，我们测得非英语语言的困惑度下降幅度是英语的 2-4 倍。我们提出语言条件反量化（LCD），一种事后方法，将每种语言的秩-2 LoRA 校正附加到已量化模型的线性层上，每种语言增加 0.12% 的参数，并在单个 GPU 上训练不到 20 分钟。在 Qwen2.5-3B 和 Llama-3.2-3B 上，LCD 恢复了非拉丁文字语言 70-83% 的困惑度差距和 17-28% 的 GlobalMMLU 准确率差距，在类型学上相距较远的语言上比同等容量的语言无关校正高出 3-9 个百分点，并比无数据低秩基线（LQER）高出一个数量级。我们进一步识别出困惑度与准确率之间的脱节，并将其追溯到量化集中造成损害的位置：早期深度的误差（Llama）会向下游传播并抵抗局部校正，而晚期深度的误差（Qwen）则不会如此。LCD 的一个层限制变体直接验证了这一机制。
cs.CL / 18 / 2608.11787
GRPO for Financial Advice Generation: Outperforming Commercial LLMs under CATE Evaluation
面向金融建议生成的 GRPO：在 CATE 评估下超越商用大语言模型
Abstract
Generating actionable financial advice from business records demands that models integrate numerical reasoning, domain knowledge, and sound judgment, while avoiding recommendations that could harm the business. Direct supervision is difficult: historical decisions are not necessarily optimal, and high-quality free-form labels are expensive to obtain. We formulate financial advice generation as a reinforcement learning problem and fine-tune an open-weight language model using Group Relative Policy Optimization (GRPO). Our reward is an LLM-as-a-judge rubric that scores each recommendation across multiple binary dimensions of advice quality, augmented with a safety gate for harm prevention. Since LLM-based evaluation alone cannot confirm whether improvements reflect genuine business value rather than adaptation to the judge, we complement it with a judge-independent audit based on a standard doubly-robust Conditional Average Treatment Effect (CATE) estimator. Under this observational off-policy audit, our trained LLM achieves approximately twice the estimated gross-profit lift of the strongest evaluated commercial baseline ($0.0228$ vs.\ $0.0104$), together with the lowest downside rate and the least negative tail risk of any policy evaluated. Notably, the two evaluations do not rank the baselines identically: the untrained base model places last on the judge rubric but second on the causal audit, indicating that the audit captures a signal the judge does not. Our results demonstrate that GRPO with a finance-grounded reward signal can produce substantially more useful business recommendations than commercial LLMs, and that a judge-independent causal audit is a valuable complement to, rather than a confirmation of, LLM-as-a-judge assessment in financial NLP.
Chinese Translation
从业务记录中生成可操作的财务建议要求模型整合数值推理、领域知识和良好判断，同时避免可能损害企业的建议。直接监督是困难的：历史决策未必最优，而高质量的自由形式标签获取成本高昂。我们将财务建议生成形式化为强化学习问题，并使用群体相对策略优化（GRPO）对开放权重语言模型进行微调。我们的奖励是一个“大语言模型作为评判者”的评分标准，它在建议质量的多个二元维度上对每条推荐进行评分，并辅以防止伤害的安全门控。由于仅靠基于大语言模型的评估无法确认改进是否反映了真正的业务价值，而不仅仅是对评判者的适应，我们辅之以独立于评判者的审计，该审计基于标准的双重稳健条件平均处理效应（CATE）估计量。在这一观察性离策略审计下，我们训练的大语言模型所取得的估计毛利润提升约为最强受评估商用基线模型的两倍（$0.0228$ vs.\ $0.0104$），同时在所有被评估策略中具有最低的下行率和最小的负尾部风险。值得注意的是，这两种评估对基线模型的排序并不一致：未训练的基座模型在评判者评分标准上排名最后，但在因果审计中排名第二，这表明该审计捕捉到了评判者未能捕捉的信号。我们的结果表明，在具备金融依据的奖励信号下，GRPO 能够生成比商用大语言模型更有用的业务建议，并且独立于评判者的因果审计是对金融 NLP 中大语言模型作为评判者评估的有价值补充，而不是对其的确认。
cs.CL / 19 / 2608.11805
Hybrid Gated Attention
混合门控注意力
Abstract
Gated attention is an effective approach to mitigate attention sinks and enhance the representational capacity of attention. To further extend its effectiveness-efficiency Pareto frontier, we propose a Hybrid Gated Attention (HyGA) framework that contains three types of gating strategies. Specifically, these gates leverage diverse information from multiple stages of attention, and collaboratively build element-wise/head-wise gating from multiple perspectives, capturing intra-head and cross-head information interactions. Through our hybrid gating components, HyGA could provide multi-source modulation signals, enabling more comprehensive control over information flow and improving the representational capacity of attention. We also introduce low-rank matrix decomposition and learnable attention sink to further enhance training efficiency and stability. In experiments, we evaluate HyGA on widely-used benchmarks based on different backbones. The experimental results show that our HyGA comprehensively improves both training loss and various downstream performances compared with Gated attention. HyGA has also been verified to achieve the best performance at different computation costs, with comprehensive model analyses for better understanding. The proposed HyGA sheds light on a more effective, efficient, and stable attention mechanism.
Chinese Translation
门控注意力是缓解注意力汇点并增强注意力表征能力的有效方法。为了进一步拓展其有效性-效率帕累托前沿，我们提出了一个包含三种门控策略的混合门控注意力（HyGA）框架。具体而言，这些门控利用来自注意力多个阶段的多样信息，并从多个视角协同构建逐元素/逐头门控，捕捉头内和跨头的信息交互。通过我们的混合门控组件，HyGA 能够提供多源调制信号，从而更全面地控制信息流并提升注意力的表征能力。我们还引入了低秩矩阵分解和可学习注意力汇点，以进一步提高训练效率和稳定性。在实验中，我们在基于不同骨干网络的广泛使用的基准上评估了 HyGA。实验结果表明，与门控注意力相比，我们的 HyGA 在训练损失和各种下游性能上均有全面提升。HyGA 还被验证在不同计算成本下均能取得最佳性能，并提供了全面的模型分析以便更好理解。所提出的 HyGA 为一种更有效、高效且稳定的注意力机制提供了启示。
cs.CL / 20 / 2608.11822
Located but Not Releasable: Silent Gate Inversion and Bounded Linear Release
已定位但不可释放：静默门控反转与有界线性释放
Abstract
A growing body of work reports that language models represent task-relevant latent structure that they fail to use. Whether such structure, once located, can be converted into behavior is a separate question that is rarely tested end to end. We submit the complete pipeline -- detect, localize, and release -- to a fully preregistered stress test on a 25.7M transformer trained on causal-evidence discrimination, where a known suppression phenomenon (latent causal structure present but behaviorally unused) has previously been documented. Every threshold, claim template, and decision-tree branch was hashed and archived before any corresponding data existed. Three findings. (i) Localization succeeds: interventions at observation-evidence channels of mid layers restore target behavior on otherwise-suppressed worlds (paired release advantages $0.563$ and $0.854$, 97.5% CIs excluding zero; best-site release rate $0.889$). (ii) Gating fails out of distribution: a detector calibrated to trigger on zero out-of-distribution calibration worlds triggers on 6.9-7.3% of held-out in-distribution generations and on zero of the 2,400 held-out generations that actually need it -- a complete inversion that silently reduces the gated pipeline to its base model. (iii) Linear release is capped: removing the gate and injecting a per-instance linear direction unconditionally yields a monotone dose-response that plateaus far below the preregistered release margin (intercept $0.382 \to 0.311 \to 0.264$ vs. threshold $\le 0.08$); per-instance adaptivity adds less than $\pm 0.03$. The failure is doubly located: the detector is OOD-inverted, and the entire family of linear release directions at this site and resolution is bounded away from sufficiency. The two failures are dissociable, and neither overturns localization. Every number traces to a hashed artifact in the released audit chain.
Chinese Translation
越来越多的研究报告表明，语言模型表征了与任务相关的潜在结构，却未能使用这些结构。这种结构一旦被定位，能否转化为行为，是一个单独的问题，很少被端到端地加以检验。我们将完整流程——检测、定位和释放——提交给一项完全预注册的压力测试，该压力测试在一个经因果证据判别任务训练的 25.7M Transformer 上进行；在该任务中，一种已知的抑制现象（潜在因果结构存在但在行为上未被使用）此前已被记录。每个阈值、声明模板和决策树分支在相应数据出现之前都已进行哈希处理并归档。三项发现。（i）定位成功：对中间层观察证据通道的干预，在原本被抑制的世界上恢复了目标行为（配对释放优势 $0.563$ 和 $0.854$，97.5% 置信区间排除零；最佳位点释放率 $0.889$）。（ii）门控在分布外失效：一个被校准为在零个分布外校准世界上触发的检测器，在 6.9-7.3% 的留出分布内生成上触发，而在实际需要它的 2,400 个留出生成上触发为零——这是一种完全的反转，会静默地将门控流程降级为其基础模型。（iii）线性释放受到上限限制：移除门控并无条件注入每实例线性方向，会产生单调的剂量-反应关系，并在远低于预注册释放裕度处达到平台（截距 $0.382 \to 0.311 \to 0.264$ 对比阈值 $\le 0.08$）；每实例自适应贡献不足 $\pm 0.03$。该失败具有双重定位：检测器出现分布外反转，并且在该位点和该分辨率下的整个线性释放方向族都被有界地排除在充分性之外。这两种失败是可分离的，且都没有推翻定位。每个数字都可追溯到已发布审计链中的哈希工件。
cs.CL / 21 / 2608.11843
When the Knowledge Base Becomes the Gold Standard: Measuring Resource-Shared Evaluation Loops in Entity-Level Machine Translation
当知识库成为金标准：测量实体级机器翻译中的资源共享评估循环
Abstract
The Seungjeongwon Ilgi, a UNESCO Memory of the World record, is only 37.4% translated, and the most conspicuous failure mode in automatic translation is the person name -- a misread name corrupts the historical fact rather than merely the surface. Low-resource historical domains have no expert gold standard for entity translation, so practitioners substitute a knowledge base (KB) for the gold. That KB is the same resource injected into the system: scoring becomes self-referential and the metric measures instruction compliance rather than translation quality. We measure this loop. Using expert person-name annotations from the National Institute of Korean History as a gold independent of the injection pipeline, we hold the entity set fixed and vary only the provenance of the correct reading. Of 527 expert-annotated mentions, only 31.1% lie outside the injection pipeline, and the residual loop is not uniform -- in the overlapping segment the injected reading agrees with the human translation 97.8% of the time against 70.1% in the independent one, so the segment that looks healthiest is the one the loop is holding up. Across four models, a difference-in-differences analysis shows the gain from KB injection is confined to the segment whose gold shares the injected resource; in the independent segment it is at or below zero. Post-injection preservation clusters in a narrow 0.910-0.996 band even though baseline capability differs fivefold, so the reported gain is the complement of prior performance and weaker models appear to improve more dramatically. On an independent sample built by removing the construction filter, the measure replicates within model (overlapping intervals) while discriminating between models (non-overlapping intervals) -- it reflects a property of the model, not of the sample.
Chinese Translation
《承政院日记》作为联合国教科文组织《世界记忆名录》文献，仅有37.4%被翻译，而自动翻译中最显著的失败模式是人名——一个被误读的人名会损害历史事实，而不仅仅是表层。低资源历史领域没有针对实体翻译的专家金标准，因此实践者用知识库（KB）来代替这一金标准。该知识库正是被注入系统的同一资源：评分变成自我指涉，该指标衡量的是指令遵循程度而非翻译质量。我们测量这一循环。我们使用韩国国史编纂委员会（National Institute of Korean History）的专家人名标注作为独立于注入管道的金标准，固定实体集合，仅改变正确读法的来源。在527个专家标注的提及中，只有31.1%位于注入管道之外，且残余循环并不均匀——在重叠部分，注入读法与人工翻译的一致率为97.8%，而在独立部分为70.1%，因此看起来最健康的部分正是循环所支撑的部分。在四个模型中，双重差分分析显示，知识库注入带来的收益仅限于金标准与注入资源重合的部分；在独立部分中，该收益等于或低于零。注入后的保持率聚集在0.910–0.996的狭窄区间内，尽管基线能力相差五倍，因此所报告的增益是先前表现的补数，较弱模型似乎提升得更显著。在通过移除构建过滤器而得到的独立样本上，该度量在模型内可复现（区间重叠），同时又能区分不同模型（区间不重叠）——它反映的是模型的属性，而非样本的属性。
cs.CL / 22 / 2608.11879
Total Recall at What Cost? Benchmarking the Serving Cost of Agentic Memory Systems
全面回忆，代价几何？对智能体记忆系统服务成本的基准测试
Abstract
Long-running conversational agents increasingly rely on a memory system to avoid resending the whole conversation each turn, yet how much that costs to serve has received little systematic benchmarking. We compare three memory systems (Mem0, Hindsight, and Mastra Observational Memory) against two reference strategies -- a fixed-size rolling window and resubmitting the full transcript -- across two backbones and conversations of up to 400 turns, pairing every cost measurement with answer accuracy on 665 LoCoMo questions. First, a memory system's serving cost cannot be predicted from conversation length and message size alone: a regression that tracks the two reference strategies closely misses the memory systems by 18-69%, their cost driven instead by internal memory behavior. Second, a break-even analysis shows that whether -- and when -- a memory system becomes cheaper to serve than the full transcript is highly sensitive to the system and the backbone, from the first tens of turns for the cheapest to never within 400 turns for the most expensive. Third, no system wins on both axes: accuracy spans 21-54%, and the backbone choice drives cost as much as the memory system does.
Chinese Translation
长期运行的对话智能体越来越依赖记忆系统，以避免每轮重新发送整个对话，但为此所需的服务成本却很少得到系统性基准测试。我们比较了三种记忆系统（Mem0、Hindsight 和 Mastra Observational Memory）与两种参考策略——固定大小的滚动窗口和重新提交完整记录——在两种骨干模型以及最长 400 轮的对话上，并将每次成本测量与 665 个 LoCoMo 问题上的回答准确率配对。首先，记忆系统的服务成本无法仅从对话长度和消息大小预测：一个紧密拟合两种参考策略的回归模型对记忆系统的预测误差为 18–69%，其成本实际上由内部记忆行为驱动。其次，盈亏平衡分析表明，记忆系统是否——以及何时——比完整记录更便宜地服务，高度依赖于具体系统和骨干模型，从最便宜系统的前几十轮，到最昂贵系统在 400 轮内从未实现。第三，没有任何系统在两个维度上都获胜：准确率范围为 21–54%，而骨干模型的选择对成本的驱动作用与记忆系统本身一样大。
cs.CL / 23 / 2608.11922
LODESTAR: Trustworthy Entropy Is Navigated, Not Merely Measured -- Reinforced Polarizer Keeps a Frozen LLM from Being Confidently Misled by the Wrong Evidence
LODESTAR：可信的熵靠引导，而非仅靠测量——强化极化器使冻结的 LLM 不会被错误证据自信地误导
Abstract
Predictive-distribution entropy makes a strong selection rule in retrieval-augmented question answering: across five QA benchmarks, keeping the candidate answer that a frozen respondent LLM produces with the lowest answer-token entropy lifts mean answer $F_1$ from 0.4769 to 0.5148 over the retriever's top-ranked passage, with no gold answers. Yet this lowest-entropy rule, which prior entropy-based selectors adopt, fails in a specific and consequential way: a misleading passage makes the respondent confidently wrong, driving its entropy down precisely where the signal looks most trustworthy. We show that the failure comes from the passage the respondent reads -- and the context that passage is read in is an input we can intervene on. We introduce LODESTAR, to our knowledge the first method to score a text intervention by the uncertainty it induces in a third-party frozen respondent, compared across one question's candidates. LODESTAR uses reinforcement learning to train, once and offline, a polarizer -- a short fixed natural-language string inserted into the respondent's prompt and never into its weights; its training labels are built offline from gold answers and two LLM judges, and inference reads neither. Evaluating every competing selector under the same frozen respondent and the same candidate pools on 5,008 questions, LODESTAR attains the highest mean $F_1$ of any inference-ready selector (0.5148 to 0.5339), the highest exact match (0.4136), and the highest GPT-4o judge score of the frozen-respondent configurations judged (0.6435); its three-seed mean wins all 70 method-by-dataset $F_1$ cells against fourteen published configurations while remaining paired-significant against every one. The gain holds both in-domain and out-of-domain, and ablating the polarizer shows it is what makes the respondent read a misleading passage less often (26.0% against 30.3%).
Chinese Translation
预测分布熵在检索增强问答中是一条强有力的选择规则：在五个 QA 基准上，保留冻结的应答 LLM 以最低答案词元熵生成的候选答案，相对于检索器排名最高的段落，将平均答案 $F_1$ 从 0.4769 提升到 0.5148，且无需黄金答案。然而，这种被先前基于熵的选择器所采用的最低熵规则，以一种具体且后果严重的方式失效：一段误导性段落会使应答者自信地犯错，并恰好在该信号看起来最可信的位置将其熵压低。我们表明，失败来自应答者所读的段落——而该段落被阅读时所处的上下文，是我们可以干预的输入。我们提出了 LODESTAR，据我们所知，这是第一种通过文本干预在第三方冻结应答者中引发的不确定性来对该文本干预进行评分，并在同一问题的候选答案之间进行比较的方法。LODESTAR 使用强化学习一次性离线训练一个极化器——一个简短的固定自然语言字符串，它被插入到应答者的提示中，而从不进入其权重；其训练标签由黄金答案和两个 LLM 评审离线构建，推理时两者均不读取。在 5,008 个问题上、在相同的冻结应答者和相同的候选池下评估每一个竞争选择器，LODESTAR 取得了所有推理就绪选择器中最高的平均 $F_1$（0.5148 到 0.5339）、最高的精确匹配分数（0.4136），以及被评判的冻结应答者配置中最高的 GPT-4o 评审分数（0.6435）；其三个随机种子的均值在与十四种已发表配置的对比中，赢得了全部 70 个按方法×数据集划分的 $F_1$ 单元，同时对每一个配置都保持成对显著。这一增益在域内和域外均成立；对极化器的消融表明，正是它使应答者阅读误导性段落的频率更低（26.0% 对 30.3%）。
cs.CL / 24 / 2608.11924
Spark-to-Paper: End-to-End Research Paper Generation as a Composable Skill
Spark-to-Paper：作为可组合技能的端到端研究论文生成
Abstract
Turning a research idea into a complete paper requires more than text generation: the system must retrieve literature, design and execute experiments, revise claims according to evidence, produce publication-ready figures, and maintain consistency across a long generation process. We present Spark-to-Paper, an end-to-end research paper generation system implemented as thirteen composable skills inside an existing coding assistant, without requiring a separate agent platform or orchestration service. Spark-to-Paper separates model-based judgment from deterministic operations that can be directly executed and checked. It further separates experiment planning from reporting, so that required evidence is specified before results are observed and manuscript claims are revised according to measured outcomes. To improve reliability over long research trajectories, the system combines deterministic integrity checks with self-critique and bounds a failure mode we call the Self-Refutation Loop, in which repeated experiments continue to reject the original research objective. Spark-to-Paper also produces editable vector figures through programmatic plotting for experimental results and code-based reconstruction for generated method diagrams. Across eight controlled research topics, Spark-to-Paper achieves 99.5% citation validity and 96.4% figure editability. A controlled ablation increases fabrication detection from 14% for a single-pass draft to 92% with the full integrity and review stack, while adversarial review achieves 74% precision. The full system uses 11.9M tokens, costs $8.1 per manuscript, and requires 3.2 hours on average. These results show that end-to-end research paper generation can be implemented as a lightweight, composable workflow inside existing coding assistants while keeping experimental evidence central to how claims are accepted, revised, or abandoned.
Chinese Translation
将一个研究想法转化为一篇完整的论文需要的不仅仅是文本生成：系统必须检索文献、设计和执行实验、根据证据修正论断、生成可供发表的图表，并在漫长的生成过程中保持一致。我们提出 Spark-to-Paper，一个端到端的研究论文生成系统，它在现有编码助手内部以十三个可组合技能的形式实现，无需单独的智能体平台或编排服务。Spark-to-Paper 将基于模型的判断与可以直接执行和检查的确定性操作分离开。它进一步将实验规划与报告分开，从而在观察到结果之前就明确所需的证据，并根据测量得到的结果修正论文中的论断。为了提高在漫长研究轨迹上的可靠性，该系统将确定性完整性检查与自我批评相结合，并对一种我们称之为“自我反驳循环”的失败模式加以约束；在这种失败模式中，反复实验不断否定最初的研究目标。Spark-to-Paper 还通过针对实验结果的程序化绘图和针对生成方法图的基于代码的重建，生成可编辑的矢量图。在八个受控研究主题上，Spark-to-Paper 实现了 99.5% 的引用有效性和 96.4% 的图表可编辑性。一项受控消融实验将造假检测率从单遍草稿的 14% 提高到使用全套完整性检查与审查栈时的 92%，而对抗性审查达到 74% 的精确率。完整系统使用 11.9M 个 token，每篇论文手稿成本为 8.1 美元，平均需要 3.2 小时。这些结果表明，端到端的研究论文生成可以在现有编码助手内部实现为一种轻量、可组合的工作流，同时使实验证据在论断被接受、修正或放弃的方式中始终处于核心地位。
cs.CL / 25 / 2608.12062
Preference Tree Optimization: Enhancing Goal-Oriented Dialogue with Look-Ahead Simulations
偏好树优化：利用前瞻模拟增强目标导向对话
Abstract
Developing dialogue systems capable of engaging in multi-turn, goal-oriented conversations remains a significant challenge, especially in specialized domains with limited data. This research proposes a novel framework called Preference Tree Optimization (PTO), designed to iteratively improve agent models in such dialogue systems, by generating preference data using a method called Preference Tree with Look-Ahead. Focusing on Motivational Interviewing (MI) -- a counseling technique aimed at facilitating behavioral change -- we leverage virtual patients and an oracle evaluator to simulate conversations and generate rich preference datasets. By combining this method with Direct Preference Optimization (DPO), we aim to enhance the agent's decision-making capabilities over iterative training cycles. The proposed framework addresses data scarcity and advances the development of more nuanced and effective dialogue systems in goal-oriented domains. Experimental evaluations demonstrate that the PTO framework enhances dialogue agents' performance in goal-oriented conversations within the domain of Motivational Interviewing (MI). Models trained with PTO consistently outperformed the baseline in key metrics such as session satisfaction and working alliance. Additionally, incorporating look-ahead simulations led to improved long-term planning and more effective conversational strategies, with deeper look-ahead configurations yielding the most stable and high-scoring results.
Chinese Translation
开发能够进行多轮、目标导向对话的对话系统仍然是一项重大挑战，尤其是在数据有限的专门领域中。本研究提出了一种名为偏好树优化（PTO）的新框架，旨在通过使用一种称为带前瞻的偏好树方法生成偏好数据，以迭代地改进此类对话系统中的智能体模型。围绕动机性访谈（MI）——一种旨在促进行为改变的咨询技术——我们利用虚拟患者和预言机评估器来模拟对话并生成丰富的偏好数据集。通过将该方法与直接偏好优化（DPO）相结合，我们旨在在迭代训练周期中增强智能体的决策能力。所提出的框架解决了数据稀缺问题，并推动了目标导向领域中更细致、更有效的对话系统的发展。实验评估表明，PTO框架增强了对话智能体在动机性访谈（MI）领域内目标导向对话中的表现。使用PTO训练的模型在会谈满意度和工作同盟等关键指标上始终优于基线模型。此外，引入前瞻模拟改善了长期规划并产生了更有效的对话策略，其中更深层次的前瞻配置产生了最稳定且得分最高的结果。
cs.CL / 26 / 2608.12113
Structuring the Space of Perspectives
视角空间的结构化
Abstract
The same event can be reported from different perspectives depending on the experiences, background, and beliefs of the writer or speaker. A variety of NLP areas engage with perspectives, spanning from text analysis to algorithm optimization. A wide range of operative concepts (such as stances, sentiment, frames, and arguments) has been used to capture perspectives in texts, however the precise relationships among those concepts remain unclear. Arguably, a deeper theoretical understanding of these concepts would empower more effective research on perspectives. In this paper, we address this gap by reviewing the space of perspectives in NLP and defining a set of properties that help distinguishing perspective-related concepts. Our analysis leads us to posit a hierarchy which organizes these concepts linearly along a single axis. Finally, we show how this principled conceptual hierarchy can help researchers navigate the field and select operationalizations of perspective that align with their specific research objectives.
Chinese Translation
同一事件可以根据作者或说话者的经历、背景和信念，从不同视角进行报道。众多NLP领域都涉及视角，涵盖从文本分析到算法优化。广泛的操作性概念（如立场、情感、框架和论元）已被用于捕捉文本中的视角，但这些概念之间的确切关系仍不清楚。可以说，对这些概念更深入的理论理解将有助于更有效地研究视角。在本文中，我们通过回顾NLP中的视角空间并定义一组有助于区分视角相关概念的属性来弥补这一空白。我们的分析使我们提出一种层级结构，将这些概念沿单一轴线线性组织。最后，我们展示了这种有原则的概念层级如何帮助研究人员把握该领域，并选择与其特定研究目标相一致的视角操作化方式。
cs.CL / 27 / 2608.12121
QV-PIC: Query-Aware Visual Position-Independent Caching for Efficient RAG Serving
QV-PIC：面向高效RAG服务的查询感知视觉位置无关缓存
Abstract
Retrieval-Augmented Generation (RAG) repeatedly prefills identical text chunks across queries, incurring redundant computations. Position-Independent Caching (PIC) mitigates it by reusing precomputed Key-Value (KV) across positions, but its efficiency is constrained by the large volume of text tokens. Rendering text chunks as images can compress the text into fewer visual tokens, but the rendered-image PIC suffers more severe quality degradation than the text PIC. This representation-specific gap primarily arises from contextual mismatches across independently compiled caches and the loss of fine-grained textual evidence during visual compression. Existing PIC repair methods mainly address the former through selective recomputation, but they incur online computation and cannot recover lost textual details. We propose QV-PIC, a query-aware dual-resolution PIC reuse framework guided by model-native templates. Offline, QV-PIC compiles visual caches under the model's native chat-template prefix, improving PIC quality without online recomputation. Online, it preserves global context with low resolution and restores fine-grained textual evidence within a high-resolution budget by cumulative query relevance scores, retaining the efficiency benefit of visual compression. Across six tasks, QV-PIC improves average F1 by 21.6 points over vanilla rendered-image PIC, closes the gap to vanilla text PIC, and surpasses optimized text PIC by 2.58 F1 while reducing TTFT by 17.2\%. Relative to full prefill, it cuts TTFT by 83.8%.
Chinese Translation
检索增强生成（RAG）在多个查询之间反复预填充相同的文本块，导致冗余计算。位置无关缓存（PIC）通过跨位置重用预先计算的键值（KV）来缓解这一问题，但其效率受限于大量文本token。将文本块渲染为图像可以将文本压缩为更少的视觉token，但渲染图像PIC比文本PIC遭受更严重的质量下降。这种表示特定的差距主要源于独立编译缓存之间的上下文不匹配，以及视觉压缩过程中细粒度文本证据的丢失。现有的PIC修复方法主要通过选择性重计算来解决前者，但它们会带来在线计算，并且无法恢复丢失的文本细节。我们提出QV-PIC，一个由模型原生模板引导的查询感知双分辨率PIC重用框架。离线时，QV-PIC在模型的原生聊天模板前缀下编译视觉缓存，从而在无需在线重计算的情况下提高PIC质量。在线时，它通过低分辨率保留全局上下文，并在高分辨率预算内通过累积查询相关性分数恢复细粒度文本证据，从而保留视觉压缩的效率优势。在六个任务上，QV-PIC相较于普通渲染图像PIC将平均F1提高了21.6个百分点，缩小了与普通文本PIC的差距，并在将TTFT降低17.2%的同时比优化文本PIC的F1高出2.58。相对于完全预填充，它将TTFT降低了83.8%。
cs.CL / 28 / 2608.12129
SAG: SQL-Retrieval Augmented Generation with Query-Time Dynamic Hyperedges
SAG：具有查询时动态超边的 SQL 检索增强生成
Abstract
While retrieval-augmented generation (RAG) has proven effective at giving LLMs access to external knowledge, mainstream dense-retrieval implementations remain inherently limited in handling structured constraints and multi-hop reasoning. Graph-based methods address this by constructing knowledge graphs offline, but they often fragment semantics, incur high maintenance, and complicate incremental updates. We propose SAG (SQL-Retrieval Augmented Generation), a structured retrieval architecture that organizes documents into an event-entity index without building a global knowledge graph. SAG represents each chunk as a semantically complete event paired with its entities, forming a latent hyperedge that preserves n-ary relations without decomposing them into triples. At query time, SAG treats shared entities as join keys to connect related chunks. This dynamically yields a query-scoped neighborhood of events, and yet every piece of evidence remains the original chunk throughout. Experiments on HotpotQA, 2WikiMultiHopQA, and MuSiQue show that SAG achieves the best retrieval and end-to-end QA performance on every benchmark, with gains that widen as reasoning-chain complexity increases. On MuSiQue, where multi-hop evidence chaining is most demanding, SAG reaches 80.36% Recall@5, outperforming the strongest baseline by 11.52 points. This work paves the way for knowledge infrastructure that enables LLM agents to retrieve and reason over continually growing organizational knowledge.
Chinese Translation
尽管检索增强生成（RAG）已被证明能有效让大语言模型访问外部知识，但主流稠密检索实现在处理结构化约束和多跳推理方面仍存在固有局限性。基于图的方法通过离线构建知识图谱来解决这一问题，但这类方法通常会造成语义割裂、维护成本高，并使增量更新变得复杂。我们提出了 SAG（SQL 检索增强生成），这是一种结构化检索架构，在无需构建全局知识图谱的情况下将文档组织为事件-实体索引。SAG 将每个文本块表示为一个语义完整的事件，并与其关联实体配对，形成潜在超边，从而无需将 n 元关系分解为三元组即可保留这些关系。在查询时，SAG 将共享实体视为连接键，以连接相关文本块。这会动态生成一个以查询为范围的事件邻域，但每条证据自始至终仍然是原始文本块。在 HotpotQA、2WikiMultiHopQA 和 MuSiQue 上的实验表明，SAG 在所有基准上均取得了最佳的检索和端到端问答性能，并且随着推理链复杂度的增加，性能优势进一步扩大。在多跳证据链接要求最为严苛的 MuSiQue 上，SAG 的 Recall@5 达到 80.36%，比最强基线高出 11.52 个百分点。这项工作为知识基础设施铺平了道路，使 LLM 智能体能够在不断增长的组织知识上进行检索和推理。
cs.CL / 29 / 2608.12269
A Cascaded Unsupervised-Supervised NLP Pipeline for Detecting Accusatory Language in Public Procurement
一种用于检测公共采购中指控性语言的级联无监督-监督自然语言处理流水线
Abstract
Public procurement involves the allocation of substantial financial resources; therefore, continuous oversight through audits, controls, and monitoring mechanisms is essential. However, stakeholder comments and publicly available government data are often underutilized, despite their potential to reveal procedural irregularities. To address this gap, this paper analyzes metadata from Ecuador's Sistema Oficial de Contratación Pública (SOCE, Official Public Procurement System), with particular emphasis on participant comments generated during the pre-contractual phase. We propose a hybrid modeling framework that integrates unsupervised clustering and supervised classification within a natural language processing (NLP) pipeline to uncover latent patterns and detect potentially irregular procurement processes. Semantic embeddings are generated using Word2Vec, LLaMA, and RoBERTa, followed by Gaussian Mixture Models (GMMs) for unsupervised clustering. A supervised classification stage is then applied to identify accusatory or whistleblowing-style comments. Experimental results show that the combination of domain-trained Word2Vec embeddings, GMM-based clustering, and a Random Forest classifier achieves high precision and recall, even under severe class imbalance. These findings demonstrate that lightweight, domain-adapted NLP architectures can effectively support risk identification and enhance transparency in public procurement systems without requiring large-scale computational infrastructure.
Chinese Translation
公共采购涉及大量财政资源的分配；因此，通过审计、控制和监测机制进行持续监督至关重要。然而，利益相关者的评论和公开可得的政府数据往往未被充分利用，尽管它们具有揭示程序违规行为的潜力。为弥补这一不足，本文分析了厄瓜多尔官方公共采购系统（Sistema Oficial de Contratación Pública，SOCE，即 Official Public Procurement System）的元数据，特别关注合同前阶段生成的参与者评论。我们提出了一种混合建模框架，将无监督聚类和监督分类集成到自然语言处理（NLP）流水线中，以揭示潜在模式并检测可能存在违规的采购过程。语义嵌入使用 Word2Vec、LLaMA 和 RoBERTa 生成，随后使用高斯混合模型（GMMs）进行无监督聚类。然后应用监督分类阶段来识别指控性或举报式评论。实验结果表明，即使存在严重的类别不平衡，领域训练的 Word2Vec 嵌入、基于 GMM 的聚类和随机森林分类器的组合也能实现较高的精确率和召回率。这些发现表明，轻量级、适应领域的 NLP 架构能够有效支持风险识别并提高公共采购系统的透明度，而无需大规模计算基础设施。
cs.CL / 30 / 2608.12278
Structural Silence: When AI Infrastructure Fails Speakers of Underrepresented Languages
结构性沉默：当人工智能基础设施辜负了代表性不足语言的使用者
Abstract
Artificial intelligence tools for education and language support are increasingly framed as scalable responses to access gaps in under-resourced communities. Yet the infrastructure underlying these tools, including training corpora, tokenization schemes, evaluation benchmarks, and deployment architectures, can systematically disadvantage speakers of underrepresented languages before a model is trained. This paper examines these structural barriers through Bengali, one of the world's most widely spoken languages, focusing on AI-assisted education in low-connectivity environments. We identify four interlocking failures: a severe web presence gap, with Bengali accounting for less than 0.5% of global web content despite representing nearly 4% of the global population; a 67:1 training-token deficit between English and Bengali in major multilingual corpora; a tokenization penalty associated with Bengali's alphasyllabary script that compounds the data deficit through higher token fertility; and connectivity exclusion, with individual internet penetration at 36.5% in rural areas compared with 71.4% in urban areas. These failures reflect longstanding resource-allocation decisions, institutional priorities, and design defaults that did not center underrepresented languages in mainstream AI development. We argue that dataset scarcity should be understood as a structural barrier rather than an isolated technical limitation, and that offline-first design should be treated as an equity-oriented infrastructure strategy. We conclude with directions for linguistics and AI research aimed at reducing these structural inequalities.
Chinese Translation
用于教育和语言支持的人工智能工具越来越多地被定位为对资源不足社区中可及性差距的可扩展应对方案。然而，这些工具背后的基础设施——包括训练语料库、分词方案、评估基准和部署架构——可能在模型训练之前就系统性地使代表性不足语言的使用者处于不利地位。本文以世界上使用最广泛的语言之一孟加拉语为切入点，考察这些结构性障碍，重点关注低连通性环境中的人工智能辅助教育。我们识别出四重相互关联的失败：严重的网络存在差距，即孟加拉语在全球网络内容中所占比例不足0.5%，尽管其使用者占全球人口近4%；在主要多语语料库中，英语与孟加拉语之间的训练词元赤字为67:1；与孟加拉语的元音附标文字相关的分词惩罚通过更高的词元产出率加剧了数据赤字；以及连通性排斥，农村地区的个人互联网普及率为36.5%，而城市地区为71.4%。这些失败反映了长期存在的资源配置决策、机构优先事项和设计默认设置，它们在主流人工智能开发中并未以代表性不足的语言为中心。我们认为，数据集稀缺应被理解为一种结构性障碍，而不是孤立的技术限制；离线优先设计应被视为一种以公平为导向的基础设施策略。最后，我们提出了旨在减少这些结构性不平等的语言学与人工智能研究方向。
cs.CL / 31 / 2608.12313
AVA-Encoder: Towards Agent-Native Video Representation Learning
AVA-Encoder：迈向智能体原生视频表征学习
Abstract
Creative agents still lack an effective way to learn from high-quality human films, limiting their ability to produce cinematic-grade videos. A key challenge is the absence of a structured video representation that is both faithful to film content and directly usable for agentic reasoning and manipulation. To address the challenge, we propose the Agentic Video Auto-Encoder (AVA-Encoder), a framework for learning agent-native video representations via agentic auto-encoding. AVA-Encoder transforms a video into a knowledge graph (KG) representation and then reconstructs it back into video. Its hierarchy and state nodes store structured text, while a linked asset layer holds generated images, audio, and video. Typed edges preserve the relations between these text descriptions and assets in a form that agents can easily understand, query, and edit. The video reconstruction differences drive a textual-gradient optimization framework, which expresses evaluation feedback as natural-language update directions for Data-Independent Encoding Policy Pseudo-Training in the outer loop and optional Data-Dependent KG Representation Refinement in the test-time inner loop. Extensive experiments show that AVA-Encoder improves by 20.7 percentage points over the strongest external baseline. In the controlled policy-only setting, its pseudo-trained shot-level Agentic Video Encoder policy also outperforms a carefully human-tuned policy while using 74.3% fewer system-prompt tokens. We release the complete AVA-Encoder framework, a reliable agentic video reconstruction benchmark, and the first dataset of high-quality film KG representations.
Chinese Translation
创意智能体仍缺乏从高质量人类影片中学习的有效途径，这限制了它们生成电影级视频的能力。一个关键挑战在于，缺乏一种既忠实于影片内容、又能直接用于智能体推理与操控的结构化视频表征。为应对这一挑战，我们提出了智能体视频自编码器（AVA-Encoder），这是一个通过智能体自编码来学习智能体原生视频表征的框架。AVA-Encoder 先将视频转换为知识图谱（KG）表征，再将其重建回视频。其中的层级节点和状态节点存储结构化文本，而相连的资产层保存生成的图像、音频和视频。类型化边以智能体易于理解、查询和编辑的形式，保留这些文本描述与资产之间的关系。视频重建差异驱动一个文本梯度优化框架，该框架将评估反馈表达为自然语言更新方向：在外循环中用于数据无关的编码策略伪训练，在测试时内循环中用于可选的数据相关 KG 表征细化。大量实验表明，AVA-Encoder 相较于最强外部基线提升了 20.7 个百分点。在仅策略受控的设置下，其经过伪训练的镜头级智能体视频编码器策略在系统提示 token 使用量减少 74.3% 的同时，也优于精心人工调优的策略。我们发布了完整的 AVA-Encoder 框架、一个可靠的智能体视频重建基准，以及首个高质量电影 KG 表征数据集。
cs.CL / 32 / 2608.11650
Confucius4-TTS: Transcript-Free Cross-Lingual Zero-Shot TTS with a Learnable Speaker Encoder
Confucius4-TTS：采用可学习说话人编码器的无转录跨语言零样本文本到语音
Abstract
Recent advances in zero-shot text-to-speech (TTS) have substantially improved speech quality and voice cloning fidelity. However, many zero-shot TTS systems still depend on audio prompt transcripts at inference time. This dependency limits cross-lingual voice cloning, since in-the-wild reference audio is often untranscribed. In this technical report, we present Confucius4-TTS, a multilingual zero-shot TTS system that supports 14 languages and performs both intra-lingual and cross-lingual reference cloning without requiring transcripts of audio prompts. Confucius4-TTS follows a two-stage architecture, consisting of text-to-semantic (T2S) and semantic-to-acoustic (S2A) modules. The LLM-based T2S module uses a learnable speaker encoder to extract timbre features from self-supervised speech representations, and the conditional flow-matching S2A module converts the predicted semantic tokens into mel-spectrograms. The same model also supports continuation cloning when a reference transcript is available. Confucius4-TTS is trained on large-scale multilingual speech data. It achieves high intelligibility and speaker similarity on public benchmarks. On the CV3-Eval cross-lingual benchmark, Confucius4-TTS obtains an average WER of 3.73% across six directions. On our internal cross-lingual set, it achieves the best average overall rank in human evaluation among recent open-source and commercial systems. We release code, model checkpoints, and demos at https://github.com/netease-youdao/Confucius4-TTS.
Chinese Translation
近年来，零样本文本到语音（TTS）的进展大幅提升了语音质量和声音克隆的保真度。然而，许多零样本TTS系统在推理时仍然依赖音频提示的转录文本。这种依赖限制了跨语言语音克隆，因为自然场景中的参考音频往往没有转录文本。在本技术报告中，我们提出Confucius4-TTS，一个多语言零样本TTS系统，支持14种语言，并且在不需要音频提示转录文本的情况下执行语内和跨语言参考克隆。Confucius4-TTS采用两阶段架构，包括文本到语义（T2S）和语义到声学（S2A）模块。基于LLM的T2S模块使用可学习的说话人编码器从自监督语音表示中提取音色特征，条件流匹配的S2A模块将预测的语义标记转换为梅尔频谱图。同一模型在参考转录文本可用时还支持续写克隆。Confucius4-TTS在大规模多语言语音数据上训练。它在公开基准上取得了高可懂度和说话人相似度。在CV3-Eval跨语言基准上，Confucius4-TTS在六个方向上获得了平均3.73%的词错误率（WER）。在我们的内部跨语言测试集上，它在近期开源和商业系统的人工评测中取得了最佳平均综合排名。我们在 https://github.com/netease-youdao/Confucius4-TTS 发布代码、模型检查点和演示。
cs.CL / 33 / 2608.11755
MuseCritic: Learning Multi-Aspect Song Rewards through Natural-Language Aesthetic Critiques
MuseCritic：通过自然语言审美评论学习多维度歌曲奖励
Abstract
Long-form song generation models continue to improve in duration, structural integrity, and acoustic complexity, making reliable aesthetic rewards increasingly important for aligning these models with human preferences. However, reward models for complete songs remain limited, and existing evaluators typically predict scores in a single forward pass without providing readable explanations. We introduce MUSECRITIC, a semi-scalar reward model that generates a natural-language critique covering five aesthetic dimensions and uses it as an intermediate representation to predict continuous reward scores. MUSECRITIC follows a two-stage training pipeline: a teacher model first provides high-quality critiques for supervised fine-tuning, after which the fine-tuned model generates its own critiques for reward learning, mitigating distribution shift between training and inference. On an in-domain test set of 200 SongEval songs, MUSECRITIC reduces macro-averaged mean squared error from 0.2875 to 0.2316 and improves macro-averaged LCC, SRCC, and Kendall's tau to 0.9068, 0.8838, and 0.7178, respectively. On the out-of-domain Music Arena benchmark with 733 preference pairs, it achieves the highest accuracy of 71.35%. Moreover, using MUSECRITIC with GRPO improves Muse-0.6B on all nine aesthetic metrics from SongEval and Audiobox Aesthetics. These results demonstrate that critique-conditioned reward modeling reduces scoring error and provides an effective optimization signal for song generation. The project repository is available at https://github.com/WuqnEl/MuseCritic.
Chinese Translation
长格式歌曲生成模型在时长、结构完整性和声学复杂度方面持续改进，使得可靠的审美奖励对于将这些模型与人类偏好对齐变得越来越重要。然而，针对完整歌曲的奖励模型仍然有限，现有评估器通常在一次前向传播中预测分数，而不提供可读的解释。我们引入了 MUSECRITIC，一种半标量奖励模型，它生成涵盖五个审美维度的自然语言评论，并将其作为中间表示来预测连续奖励分数。MUSECRITIC 采用两阶段训练流程：教师模型首先为监督微调提供高质量评论，之后微调后的模型生成自己的评论用于奖励学习，从而缓解训练与推理之间的分布偏移。在包含 200 首 SongEval 歌曲的域内测试集上，MUSECRITIC 将宏平均均方误差从 0.2875 降至 0.2316，并将宏平均 LCC、SRCC 和 Kendall's tau 分别提升至 0.9068、0.8838 和 0.7178。在包含 733 个偏好对的域外 Music Arena 基准上，它实现了 71.35% 的最高准确率。此外，将 MUSECRITIC 与 GRPO 结合使用，在 SongEval 和 Audiobox Aesthetics 的全部九项审美指标上改善了 Muse-0.6B。这些结果表明，以评论为条件的奖励建模降低了评分误差，并为歌曲生成提供了有效的优化信号。项目代码库可在 https://github.com/WuqnEl/MuseCritic 获取。
cs.CL / 34 / 2608.12099
RT-SEMamba: Real-Time Speech Enhancement Mamba via Progressive Knowledge Distillation
RT-SEMamba：基于渐进式知识蒸馏的实时语音增强 Mamba
Abstract
We present RT-SEMamba, a fully causal speech enhancement (SE) model built upon causal time-frequency Mamba blocks. Unlike Transformer-based architectures that rely on a growing key-value cache, Mamba propagates a fixed-size recurrent state per layer, enabling memory- and bandwidth-efficient long-form inference. We further introduce a progressive knowledge distillation (KD) strategy that compresses an 8-layer teacher into a shallow 1-layer student by jointly distilling complex spectral outputs and intermediate representations. On Voicebank-DEMAND, the 8-layer RT-SEMamba achieves 3.32 PESQ with a 25 ms algorithmic latency constraint, and the distilled 1-layer student improves over a naive 1-layer baseline from 3.06 to 3.18 PESQ while preserving the same steady-state RTF, delivering a 2.75x speedup over the teacher. These results demonstrate that state-space models with progressive KD provide a competitive quality-latency trade-off for real-time SE.
Chinese Translation
我们提出了 RT-SEMamba，一种基于因果时频 Mamba 模块的全因果语音增强（SE）模型。与依赖不断增长的键值缓存的基于 Transformer 的架构不同，Mamba 在每层传播固定大小的循环状态，从而实现内存和带宽高效的长格式推理。我们进一步引入一种渐进式知识蒸馏（KD）策略，通过联合蒸馏复数频谱输出和中间表示，将 8 层教师模型压缩为浅层 1 层学生模型。在 Voicebank-DEMAND 数据集上，8 层 RT-SEMamba 在 25 毫秒算法延迟约束下实现了 3.32 的 PESQ，蒸馏得到的 1 层学生模型将朴素 1 层基线的 PESQ 从 3.06 提升至 3.18，同时保持相同的稳态 RTF，并比教师模型实现了 2.75 倍的加速。这些结果表明，采用渐进式知识蒸馏的状态空间模型为实时语音增强提供了具有竞争力的质量-延迟权衡。
cs.CL / 35 / 2608.11587
Robust Multi-Tier Infant-Centered Audio Understanding with Whisper via Structured Speaker Conditioning
通过结构化说话人条件实现基于 Whisper 的稳健多层级婴儿中心音频理解
Abstract
Recent advances in model design and self-supervised audio representations have improved speech and audio understanding, yet infant-centered naturalistic recordings remain challenging due to limited labeled data, low signal-to-noise ratio, and cross-family domain shifts. We present a family-conditioned, multi-tier audio tagger that combines a LoRA-finetuned Whisper encoder with a lightweight, target-speaker-aware Transformer for long-context inference and framewise prediction across tiers. To improve temporal coherence, we incorporate a simple sequence-level smoothing loss, and to enhance robustness across households, we introduce a factorized speaker-token design with a shared tier token and a learned family-specific offset, reducing family bias and promoting generalizable representations. Together, these choices enable efficient and effective infant-centered audio tagging of daylong audio recordings in home environments.
Chinese Translation
模型设计和自监督音频表征的最新进展已改善语音和音频理解，然而以婴儿为中心的自然情境录音仍然具有挑战性，原因包括标注数据有限、信噪比低以及跨家庭领域偏移。我们提出了一种家庭条件化的多层级音频标注器，它将经过 LoRA 微调的 Whisper 编码器与一个轻量级、目标说话人感知的 Transformer 相结合，用于跨层级的长上下文推理和逐帧预测。为改善时间一致性，我们引入了一种简单的序列级平滑损失；为增强跨家庭的稳健性，我们引入了一种分解式说话人标记设计，其中包括共享层级标记和经学习的家庭特定偏移，从而减少家庭偏差并促进可泛化的表征。这些选择共同使得在家庭环境中对全天音频记录进行高效且有效的婴儿中心音频标注成为可能。
多智能体系统 (cs.MA)
4
cs.MA / 1 / 2608.11357
When Do Institutions Beat Intelligence?
制度何时胜过智能？
Abstract
More capable agents do not necessarily form a more capable collective. A multi-agent system may jointly possess sufficient information yet fail because evidence is poorly routed, unreliable reports enter public belief, correlated claims masquerade as independent support, shared state becomes stale or strategically distorted, or useful evidence is exposed through an ineffective action interface. We ask when additional resources should improve the reasoner and when they should instead change the institutional structure through which the collective forms and acts on public information. Drawing on functional distinctions from research on group decision making and distributed cognition, we construct controlled artificial ecologies around four loci of collective failure: access and routing, admission and dependence, state maintenance and incentives, and representation and action. Across these ecologies, we separately vary model capability and institutional structure, pairing positive interventions with matched reasoning baselines and mechanism-breaking controls. The experiments reveal a consistent boundary: institutions help when they repair failures in how a collective constructs usable public state, but lose their advantage when their signals are uninformative or uncheckable, when stronger intelligence can perform the same transformation directly, or when the resulting state cannot support reliable action. Our results recast the choice between intelligence and institutions as a diagnosis of where collective reasoning fails.
Chinese Translation
更有能力的智能体未必能组成更有能力的集体。一个多智能体系统可能共同拥有足够的信息，却仍然失败，因为证据传递路径不畅、不可靠的报告进入公共信念、相关的主张伪装成独立支持、共享状态变得过时或被策略性扭曲，或者有用的证据通过无效的行动接口暴露出来。我们追问：在什么情况下，额外资源应当用于改进推理者，而在什么情况下，它们应当转而改变集体赖以形成公共信息并据以行动的制度结构。借鉴群体决策和分布式认知研究中的功能区分，我们围绕集体失败的四个关键点构建了受控人工生态：访问与路由、准入与依赖、状态维护与激励，以及表征与行动。在这些生态中，我们分别改变模型能力和制度结构，并将积极干预与匹配的推理基线以及破坏机制的对照条件配对。实验揭示了一个一致的边界：当制度修复了集体如何构建可用公共状态方面的失败时，制度会有所帮助；但当制度信号缺乏信息或无法核查、当更强的智能能够直接完成同样的转化，或者当由此产生的状态无法支持可靠行动时，制度便失去优势。我们的结果将智能与制度之间的选择重新界定为对集体推理失败所在的诊断。
cs.MA / 2 / 2608.11632
Beyond Memory: A Transactional Continuity Kernel for Long-Lived AI Agents
超越记忆：面向长生命周期AI智能体的事务性连续性内核
Abstract
Persistent AI agents accumulate versioned state across long horizons, but storage retention alone does not identify authoritative state. Without an explicit control plane, unmediated updates by models, tools, and background workers risk stale overwrites, un-audited exposures, and self-authorizing privilege escalation. We argue that agent state governance is an infrastructural activation problem, defining continuity as an unbroken, authorized lineage of accepted branch heads. We present the Continuity Kernel (CK), an activation contract that decouples off-commit candidate evaluation from atomic state activation. Untrusted components propose typed changes against an exact predecessor head or typed absence. A short activation transaction revalidates ownership, pre-state authority, freshness, and effect uniqueness, recording one stable disposition (Commit, Reject, Quarantine, or Defer). Only Commit atomically advances the branch head and installs the complete accepted unit (state, authority, lineage, effects, outcome, and receipt). A bounded executable model verifies the protocol across 2,808,230 reachable states and 5,526,474 state-changing transitions with zero invariant violations.
Chinese Translation
持久型AI智能体在长时间跨度内累积带版本的状态，但仅靠存储保留并不能确定权威状态。在缺乏显式控制平面的情况下，模型、工具和后台工作进程未经协调的更新可能会带来过期覆盖、未审计暴露以及自我授权的权限提升等风险。我们认为智能体状态治理是一个基础设施层面的激活问题，并将连续性定义为已接受分支头的一条不间断、经授权的谱系。我们提出连续性内核（CK），这是一种激活契约，将提交前的候选评估与原子状态激活解耦。不可信组件针对确切的前驱头或类型化缺失提出带类型的更改。一个简短的激活事务重新验证所有权、前状态权限、新鲜度以及效果唯一性，并记录一个稳定处置结果（提交、拒绝、隔离或延迟）。只有提交会原子地推进分支头，并安装完整的已接受单元（状态、权限、谱系、效果、结果和回执）。一个有界可执行模型在2,808,230个可达状态和5,526,474个状态改变变迁上验证了该协议，且零不变量违反。
cs.MA / 3 / 2608.12172
Rethinking Agent Security as a Networking Problem
将智能体安全重新思考为一个网络问题
Abstract
AI agents are rapidly becoming more capable and widely deployed, promising substantial gains in productivity and enabling new classes of applications. However, their growing autonomy also introduces significant privacy and security risks. Existing defenses are predominantly agent-centric, relying on the agent itself to detect threats and enforce privacy and security policies. This approach is fundamentally limited because it entrusts policy enforcement to AI agents whose LLM-driven behavior is inherently nondeterministic and vulnerable to manipulation through attacks such as prompt injection. As a result, current defenses cannot reliably prevent privacy and security threats, highlighting a critical need for a new solution to securing AI agent systems. The networking community has long grappled with similar challenges and offers insightful principles we can borrow to design a more secure AI agent system. These include centralized control with distributed enforcement, capability-based access for mediating requests to sensitive resources, and least privilege through zero-trust enforcement. Historically, these principles have provided strong deterministic guarantees for networked systems. However, these principles alone are insufficient for AI agents because the safety and appropriateness of an agent's actions often depend on semantic context beyond the expressiveness of static rules. Building on these principles, we advocate for a systematic approach to AI agent security that combines deterministic enforcement mechanisms, which provide strong security guarantees, with semantic, context-aware policies that enable nuanced decision-making. We then present a reference architecture and identify key research questions and future directions to guide the design of secure and privacy-preserving AI agent systems.
Chinese Translation
AI智能体正迅速变得更加强大并被广泛部署，有望显著提升生产力并催生新型应用。然而，它们日益增强的自主性也带来了重大的隐私和安全风险。现有的防御措施主要以智能体为中心，依赖智能体自身来检测威胁并执行隐私和安全策略。这种方法存在根本性局限，因为它将策略执行委托给AI智能体，而后者由LLM驱动的行为本质上是非确定性的，并且容易受到提示注入等攻击的操纵。因此，当前的防御措施无法可靠地防止隐私和安全威胁，这凸显了迫切需要一种新的解决方案来保障AI智能体系统的安全。网络社区长期以来一直在应对类似的挑战，并提供了可以借鉴的富有洞见的原则，以设计更安全的AI智能体系统。这些原则包括：集中控制与分布式执行、基于能力的访问控制以调解对敏感资源的请求，以及通过零信任执行实现最小权限。从历史上看，这些原则为网络系统提供了强有力的确定性保证。然而，仅凭这些原则对AI智能体而言是不够的，因为智能体行为的安全性和适当性往往取决于超出静态规则表达能力的语义上下文。在这些原则的基础上，我们倡导一种系统化的AI智能体安全方法，将提供强安全保证的确定性执行机制与支持细致决策的语义化、上下文感知策略结合起来。随后，我们提出一个参考架构，并确定关键研究问题和未来方向，以指导安全且保护隐私的AI智能体系统的设计。
cs.MA / 4 / 2608.11895
Scalable Multi-Agent Maze Traversal with Local Communication
基于局部通信的可扩展多智能体迷宫遍历
Abstract
Cave networks, pipe systems, and similar maze-like environments pose significant challenges for multi-agent navigation in unknown settings with limited communication. We propose a distributed algorithm that enables agents to collectively traverse an unknown, possibly cyclic graph. Agents enter sequentially at a designated start node and are tasked to localize and reach an undisclosed goal while avoiding collisions. They coordinate via local communication using leader-follower relationships and leader switching. At any moment in time, exploration is performed by only one of the agents, which runs a single-agent maze solver. We prove that the algorithm is complete, that its makespan is asymptotically equivalent (in the number of agents) to that of an optimal full-knowledge strategy, and derive its time and space complexity. Simulations with up to $625$ agents show a decreasing average sum-of-fuels as the number of agents increases and demonstrate that the proposed approach outperforms a naïve baseline in which all agents independently execute the single-agent solver.
Chinese Translation
洞穴网络、管道系统以及类似的迷宫式环境对在未知且通信受限场景下的多智能体导航构成了重大挑战。我们提出一种分布式算法，使多个智能体能够集体遍历一个未知的、可能含有环的图。智能体在指定的起始节点依次进入，其任务是定位并到达一个未公开的目标，同时避免碰撞。它们通过局部通信，利用领导者-跟随者关系以及领导者切换进行协调。在任意时刻，探索仅由一个智能体执行，该智能体运行单智能体迷宫求解器。我们证明了该算法是完备的，其完工时间在智能体数量上渐近等价于最优完全知识策略的完工时间，并推导了其时间复杂度和空间复杂度。对多达 $625$ 个智能体的仿真表明，随着智能体数量增加，平均总燃料消耗下降，并证明所提出的方法优于一种朴素基线，在该基线中所有智能体独立执行单智能体求解器。
软件工程 (cs.SE)
5
cs.SE / 1 / 2608.11345
MergirafSemi: A Language-Agnostic Semistructured Merge Tool
MergirafSemi：一种语言无关的半结构化合并工具
Abstract
Developers frequently face merge conflicts when integrating concurrent changes. Most merge tools rely on unstructured, line-based comparisons, often producing spurious conflicts and missing actual conflicts. To address these limitations, structure-aware merge tools have been proposed, which leverage syntactic representations to improve merge accuracy. However, fully structured tools may incur higher computational cost, and language-specific tools require significant development and maintenance effort. To balance these trade-offs, we propose MergirafSemi, a language-agnostic semistructured merge tool that captures structural information without requiring full structural modeling or language-specific implementations, and applies line-based merging within specific program regions, such as method bodies in Java. Our tool leverages lightweight Concrete Syntax Trees to guide merging decisions while preserving flexibility and efficiency across languages. We evaluate our tool through an empirical study on real-world merge scenarios across multiple programming languages, comparing it with unstructured, semistructured, and structured tools. Our results show that increasing structural granularity improves automatic conflict resolution but can also lead to more aggressive merge decisions, increasing the number of missed actual conflicts. In contrast, MergirafSemi achieves a more balanced trade-off, reducing spurious conflicts while maintaining competitive accuracy and better runtime performance in most common scenarios. Compared to an unstructured tool, it substantially reduces spurious conflicts, and when compared to a semistructured language-specific tool, it achieves comparable effectiveness while exhibiting more robust execution behavior and significantly lower runtime overhead.
Chinese Translation
开发人员在集成并发变更时经常面临合并冲突。大多数合并工具依赖非结构化的、基于行的比较，通常会产生虚假冲突并遗漏实际冲突。为了解决这些局限性，人们提出了结构感知的合并工具，它们利用语法表示来提高合并精度。然而，全结构化工具可能会带来更高的计算开销，而针对特定语言的工具则需要大量的开发和维护工作。为了在这些权衡之间取得平衡，我们提出了 MergirafSemi，一种语言无关的半结构化合并工具，它能够捕获结构信息，而无需完整的结构建模或特定语言的实现，并在特定程序区域内（例如 Java 中的方法体）应用基于行的合并。我们的工具利用轻量级具体语法树来指导合并决策，同时保持跨语言的灵活性和效率。我们通过在多种编程语言上的真实世界合并场景进行实证研究来评估我们的工具，并将其与非结构化、半结构化和结构化工具进行比较。我们的结果表明，提高结构粒度可以改善自动冲突解决，但也可能导致更激进的合并决策，从而增加遗漏的实际冲突数量。相比之下，MergirafSemi 实现了更平衡的权衡，在大多数常见场景中减少虚假冲突，同时保持有竞争力的准确性和更好的运行时性能。与非结构化工具相比，它显著减少了虚假冲突；与半结构化的特定语言工具相比，它实现了相当的有效性，同时表现出更稳健的执行行为和显著更低的运行时开销。
cs.SE / 2 / 2608.11744
Harnessing LLMs for Document-Guided Fuzzing of Python Libraries
利用大语言模型对 Python 库进行文档引导的模糊测试
Abstract
Python libraries underpin deep learning, scientific computing, data analysis, and computer vision, making their reliability critical to downstream applications. Testing their APIs requires inputs that satisfy both per-parameter constraints and dependencies among parameters. Existing approaches either leave such constraints implicit in generated programs or rely on library-specific parsing rules. This paper introduces VistaFuzz, a document-guided fuzzing technique that uses a locally served open-sourced LLM to extract parameter specifications from API documents and generate inputs that satisfy both parameter constraints and inter-parameter dependencies. We evaluate VistaFuzz on 7,718 APIs across twelve Python libraries. Inter-parameter relationships occur in 40.1\% of tested APIs, and disabling their resolution reduces the valid generation rate on those APIs from above 95\% to 31.6\%--52.8\%. VistaFuzz reports 74 issues, of which 43 have been confirmed by developers and 29 have been fixed.
Chinese Translation
Python 库支撑着深度学习、科学计算、数据分析和计算机视觉，因此其可靠性对下游应用至关重要。测试这些库的 API 需要同时满足每个参数约束和参数之间依赖关系的输入。现有方法要么将此类约束隐式地留在生成程序中，要么依赖特定于库的解析规则。本文介绍了 VistaFuzz，一种文档引导的模糊测试技术，它利用本地部署的开源大语言模型从 API 文档中提取参数规范，并生成同时满足参数约束和参数间依赖关系的输入。我们在十二个 Python 库的 7,718 个 API 上对 VistaFuzz 进行了评估。参数间关系出现在 40.1% 的被测 API 中，禁用对它们的解析会使这些 API 上的有效生成率从 95% 以上降至 31.6%–52.8%。VistaFuzz 报告了 74 个问题，其中 43 个已得到开发者确认，29 个已修复。
cs.SE / 3 / 2608.11766
Instruction Alignment for Binary Code Representation Learning
面向二进制代码表示学习的指令对齐
Abstract
Binary code representation learning is a fundamental problem in software security and reverse engineering. Existing methods mainly learn function-level embeddings that capture coarse-grained semantic relationships between binary functions, but they largely ignore fine-grained instruction-level correspondences. This limitation misses valuable supervision signals available from compiler debug information, which can support the learning of more accurate and interpretable binary code representations. We propose to leverage instruction alignment knowledge to further improve binary code representation learning. Our preliminary study reveals that models finetuned for function-level binary code similarity exhibit substantially better instruction alignment than their pre-trained model, suggesting a strong correlation between instruction alignment and function-level embedding quality. Motivated by this observation, we design a training approach that explicitly incorporates instruction alignment as an auxiliary training objective. Our experiments show that instruction alignment training improves retrieval accuracy and provides more discriminative signal for the model's similarity judgments.
Chinese Translation
二进制代码表示学习是软件安全与逆向工程中的一个基本问题。现有方法主要学习函数级嵌入，以捕获二进制函数之间的粗粒度语义关系，但它们在很大程度上忽略了细粒度的指令级对应关系。这一局限忽视了编译器调试信息中可用的宝贵监督信号，而这些信号有助于学习更准确、更具可解释性的二进制代码表示。我们提出利用指令对齐知识来进一步改进二进制代码表示学习。我们的初步研究表明，针对函数级二进制代码相似性进行微调的模型，其指令对齐能力明显优于其预训练模型，这表明指令对齐与函数级嵌入质量之间存在强相关性。受这一观察启发，我们设计了一种训练方法，将指令对齐明确地作为辅助训练目标纳入其中。我们的实验表明，指令对齐训练提高了检索准确率，并为模型的相似性判断提供了更具区分性的信号。
cs.SE / 4 / 2608.11886
Testing Deep Learning Library APIs via Cross-Framework Differential Fuzzing
通过跨框架差分模糊测试测试深度学习库 API
Abstract
Deep learning libraries underpin many safety- and reliability-critical applications, yet existing API-level testing techniques often rely on intra-library properties or CPU--GPU differential oracles and may miss defects that behave consistently across hardware backends. We present Xamt, a cross-framework differential fuzzing approach for deep learning library APIs. Xamt constructs and tests execution-validated groups of APIs intended to implement equivalent operations across seven libraries. It uses explicit API aliases and parameter-role normalization to construct candidate correspondences and validates them through pairwise execution and a group-level behavioral check on canonical ordinary inputs. The resulting groups are explored using variance-guided differential fuzzing with ordinary, boundary, and non-finite inputs. Crash and inconsistency oracles flag executions exhibiting abnormal termination or inconsistent outputs for subsequent reproduction and analysis. Across the seven libraries, Xamt constructs 676 execution-validated groups containing 2,563 matched APIs. Among these, Xamt identifies 72 independently reproduced discrepancy cases, including 4 crash cases and 68 output inconsistencies. Among the 72 developer reports, 25 have been confirmed, including 23 that have been fixed.
Chinese Translation
深度学习库支撑着许多安全关键与可靠性关键型应用，然而现有的 API 级测试技术通常依赖于库内性质或 CPU--GPU 差分预言，可能会漏掉在硬件后端之间表现一致的缺陷。我们提出了 Xamt，一种面向深度学习库 API 的跨框架差分模糊测试方法。Xamt 构建并测试经过执行验证的 API 组，这些 API 预期在七个库中实现等价操作。它利用显式 API 别名和参数角色规范化来构造候选对应关系，并通过成对执行以及在规范普通输入上的组级行为检查来验证这些对应关系。所得分组通过方差引导的差分模糊测试进行探索，并使用普通、边界和非有限输入。崩溃和不一致性预言机会标记表现出异常终止或不一致输出的执行，以供后续复现和分析。在七个库中，Xamt 构建了 676 个经过执行验证的组，包含 2,563 个匹配的 API。其中，Xamt 识别出 72 个独立复现的差异案例，包括 4 个崩溃案例和 68 个输出不一致案例。在 72 份开发者报告中，有 25 份已被确认，其中 23 份已修复。
cs.SE / 5 / 2608.12144
ADEPT: A Unified Framework for Deep Learning Test Adequacy
ADEPT：一个统一的深度学习测试充分性框架
Abstract
Over the past decade, many test adequacy metrics have been proposed for deep learning that characterize test dataset adequacy from different perspectives, e.g., neuron activation behavior, latent feature coverage, decision-boundary exploration, etc. However, these metrics are typically released as independent research prototypes with substantially different installation and preprocessing requirements, execution workflows, and configuration mechanisms. These complications make them quite difficult to reproduce, compare, and adopt in research work and practical deployment alike. In this paper, we present the engineering details of ADEPT, a framework that integrates representative adequacy techniques, including neuron-coverage-based metrics, surprise adequacy, input distribution coverage, boundary coverage, and source- and model-level mutation score, under a consistent execution workflow. ADEPT provides a template-based metric interface with well-defined extension points for integrating new adequacy metrics. Furthermore, it provides YAML-based configuration management, preprocessing-cache reuse, and structured result reporting, making it easy to use in any research and development workflows. ADEPT is designed for researchers and practitioners who wish to reproduce and apply adequacy metrics without spending days or weeks implementing missing tooling or configuring disparate research prototypes. A demo video is available at https://aub.ie/ADEPT_video.
Chinese Translation
过去十年中，针对深度学习提出了许多测试充分性度量指标，它们从不同角度刻画测试数据集的充分性，例如神经元激活行为、潜在特征覆盖、决策边界探索等。然而，这些度量指标通常以独立的研究原型形式发布，其安装和预处理要求、执行工作流以及配置机制存在显著差异。这些复杂性使得它们在研究工作和实际部署中都相当难以复现、比较和采用。本文介绍了 ADEPT 的工程细节，该框架在一致的执行工作流下集成了具有代表性的充分性技术，包括基于神经元覆盖的度量、意外充分性（surprise adequacy）、输入分布覆盖、边界覆盖以及源代码级和模型级变异分数。ADEPT 提供了基于模板的度量接口，并具有定义良好的扩展点，用于集成新的充分性度量。此外，它还提供了基于 YAML 的配置管理、预处理缓存复用以及结构化结果报告，使其易于在任何研发工作流中使用。ADEPT 面向希望复现和应用充分性度量、而无需花费数天或数周时间实现缺失工具或配置各种异构研究原型的研究人员和实践者。演示视频可在 https://aub.ie/ADEPT_video 获取。
操作系统 (cs.OS)
2
cs.OS / 1 / 2608.12103
Who Should Own the Expert Cache? Kernel-Managed Tiering for Trillion-Parameter MoE Inference
谁应该拥有专家缓存？面向万亿参数MoE推理的内核管理分层
Abstract
Mixture-of-experts models whose expert pools dwarf DRAM force every serving system to contain a cache, yet existing systems typically implement this cache in user space using expert-granular, frequency-ranked, explicitly pinned tiers. We characterize the alternative already provided by the OS: the page cache as the expert tier. We use router traces from three MoE models spanning 128 to 896 experts per layer, including a production trillion-parameter model with a 1.45 TB expert pool, and replay them natively against the full pool on GH200 nodes with capacity enforced by three independent mechanisms. We find four main results. First, iteration time and device traffic are smooth, reproducible functions of cache capacity, making DRAM a practical sizing knob for trillion-parameter serving. Second, the deep-pressure knee is a reclaim artifact requiring both MGLRU and balloon-style mostly-mlocked memory; cgroup limits and physical-memory configurations show no such amplification, indicating that balloon-based studies can overstate pressured-end device traffic by approximately 2$\times$. Third, under an enforced equal-memory wall, untuned kernel LRU serves essentially as much demand as a same-domain oracle frequency table (75.3% versus 74.6% at 256 GB), while the oracle's mechanism advantage is only 1.09$\times$ and disappears off-domain, where LRU remains at 70--71% hit rate. Fourth, router lookahead with 64.7% measured recall provides only a 0.3% benefit as kernel readahead advice and no benefit as synchronous prefetch. End to end, enabling kernel-managed caching improves decode performance by 1.09--1.10$\times$ with token-identical outputs across nine balanced pairs. The resulting design principle is simple: in this regime, let the kernel own eviction, while model-specific knowledge is best spent on admission and advice.
Chinese Translation
专家池远大于DRAM的混合专家模型迫使每个服务系统都包含一个缓存，然而现有系统通常在用户空间中实现该缓存，采用专家粒度、按频率排序、显式固定的层级。我们刻画了操作系统已经提供的替代方案：将页面缓存作为专家层。我们使用来自三个MoE模型的路由轨迹，这些模型每层有128到896个专家，其中包括一个具有1.45 TB专家池的生产级万亿参数模型，并在GH200节点上以原生方式对完整专家池重放这些轨迹，容量由三种独立机制强制执行。我们发现了四个主要结果。第一，迭代时间和设备流量是缓存容量的平滑、可复现的函数，使得DRAM成为万亿参数服务的一个实用容量调节旋钮。第二，深度压力拐点是一种回收伪影，需要同时具备MGLRU和气球式的大部分内存锁定；cgroup限制和物理内存配置没有表现出这种放大效应，这表明基于气球的研究可能将压力端设备流量高估约2$\times$。第三，在强制等内存墙下，未调优的内核LRU服务的需求基本上与同域oracle频率表一样多（在256 GB时分别为75.3%和74.6%），而oracle的机制优势仅为1.09$\times$，并且在域外消失，此时LRU仍保持70–71%的命中率。第四，具有64.7%实测召回率的路由前瞻作为内核预读建议仅带来0.3%的收益，而作为同步预取则没有收益。端到端来看，启用内核管理的缓存将解码性能提高了1.09–1.10$\times$，并且在九个平衡对中输出令牌完全相同。由此产生的设计原则很简单：在这个场景中，让内核负责驱逐，而模型特定的知识最好用于准入和建议。
cs.OS / 2 / 2608.12114
The Ingestion Tax: Adopting File-Backed Weights in Tensor Frameworks
摄取税：在张量框架中采用文件后备权重
Abstract
Open-weight models can occupy a middle capacity regime: active weights fit in DRAM as cached file pages, but a second framework-owned representation does not fit or must be refilled as layers run, so low-batch decode rereads the weights every token. On integrated and coherent-memory systems the checkpoint's file pages already occupy a GPU-readable domain, yet accelerator-loading paths still copy them into framework-owned allocations. We call this the ingestion tax: the OS holds the bytes as clean, evictable file pages, the architecture makes them GPU-readable, and only the framework's ownership model stands between the two. We present file-backed weight adoption: a framework-independent producer maps each tensor with MAP_SHARED, wraps the pages as a no-copy GPU buffer, and exports a DLPack capsule that PyTorch or MLX imports as ordinary storage. Zero-copy import alone is not enough: a randomized factorial establishes a three-part execution contract - read the mapping, keep activations accelerator-resident, order on the GPU; a bolt-on extension forfeiting the last two runs 2.3x slower than stock. Under the contract, adoption removes the tax at no rate cost: 516 GB/s where default constructors reach 53-82, at parity with the same kernel over resident storage ([-0.66%, +0.48%]; Qwen2.5-72B 7.14 vs 7.23 tok/s). Parity buys ownership: N processes decode one mapped copy where resident loading pays N (at capacity, 5.5 vs 0.08 tok/s); a 65 GB checkpoint reaches its first token 6.4x sooner; the Kimi K3 spine stage falls from 2.62 to 0.35 s per token, 3.8x from storage alone. The same mechanism improves llama.cpp 1.21x at half the footprint on an AMD APU, ties overlapped streaming on a capacity-exceeding GH200, and loses 39x across PCIe: memory topology, not the API, decides the byte path. The deployment rule treats the page cache as a first-class, reclaimable accelerator-storage tier.
Chinese Translation
开放权重模型可能处于一种中等容量区间：活跃权重作为缓存文件页可以放入 DRAM，但第二份由框架拥有的表示则放不下，或者必须随着层运行而重新填充，因此低批次解码每生成一个 token 都会重新读取权重。在集成式且具有一致性内存的系统上，检查点的文件页已经位于 GPU 可读的域中，但加速器加载路径仍将它们复制到框架拥有的分配中。我们将此称为摄取税：操作系统以干净、可回收的文件页持有这些字节，硬件架构使其可被 GPU 读取，而两者之间只隔着框架的所有权模型。我们提出了文件后备权重采用机制：一个与框架无关的生产者用 MAP_SHARED 映射每个张量，将页面封装为无复制 GPU 缓冲区，并导出 DLPack 胶囊，使 PyTorch 或 MLX 将其作为普通存储导入。仅零复制导入还不够：随机析因实验确立了三部分执行契约——读取映射、让激活常驻加速器、在 GPU 上排序；一个放弃后两条的附加扩展运行起来比原生实现慢 2.3 倍。在该契约下，采用机制在不损失速率的情况下消除了这项税：达到 516 GB/s，而默认构造函数仅为 53–82；与在常驻存储上运行相同内核的性能相当（[-0.66%, +0.48%]；Qwen2.5-72B 为 7.14 对 7.23 tok/s）。性能相当换来了所有权：N 个进程解码同一份映射副本，而常驻加载需要付出 N 份代价（在容量极限下，5.5 对 0.08 tok/s）；一个 65 GB 检查点到达首个 token 的速度快 6.4 倍；Kimi K3 骨干阶段从每个 token 2.62 秒降至 0.35 秒，仅存储方面就提升 3.8 倍。同一机制在 AMD APU 上使 llama.cpp 性能提升 1.21 倍，同时内存占用减半；在容量超限的 GH200 上与重叠流式传输打平；而在跨 PCIe 时损失 39 倍：决定字节路径的是内存拓扑，而不是 API。部署规则将页面缓存视为一等、可回收的加速器存储层。
硬件架构 (cs.AR)
7
cs.AR / 1 / 2608.11380
The Fallacy of Independent Ceilings: Characterizing Coupled Load-Branch Stall Interaction
独立上限的谬误：表征耦合的加载-分支停顿交互
Abstract
Branch mispredictions and data-cache misses are usually evaluated as separate bottlenecks: studies report perfect-branch or perfect-cache speedups as isolated upper bounds and often treat their product as the joint ceiling. In irregular workloads, however, hard-to-predict branches and cache-missing loads often occur in the same hot loops. Removing one penalty can expose the other: faster memory reaches mispredicted branches sooner, while better branch prediction leaves more long-latency loads in the out-of-order window. We call this interaction symbiotic stall latency (SSL). This paper quantifies when isolated ceilings fail using joint speedup synergy (JSS), the observed joint perfect-branch/perfect-cache speedup divided by the product of the isolated speedups. Values above one mean independent-ceiling analysis understates attainable gain. Across 53 simulated workloads, 70% show measurable coupling (JSS > 1), though many are near unity, especially in lower-pressure cases. With a conservative threshold, 40% exceed the independence product by more than 6%, and kernels with SSO > 20 show JSS from 1.23 to 3.29. We introduce symbiotic stall opportunity (SSO), a lightweight MPKI-based screen for workloads that merit full joint simulation. We map high-SSO workloads to four recurring software patterns: neighbor access, hash lookup, linked-structure traversal, and data-dependent modification. We connect SSL to reorder-buffer occupancy, squash rate, and commit starvation under isolated perfect modes. The resulting methodology is simple: use SSO to screen, JSS to validate, and report conditional branch-after-cache and cache-after-branch gains when evaluating branch predictors, prefetchers, caches, or coupled branch/memory mechanisms. Our contribution is a measurement framework showing when isolated perfect modes are adequate and when they understate joint performance headroom.
Chinese Translation
分支误预测和数据缓存缺失通常被作为单独的瓶颈来评估：研究报告将完美分支或完美缓存的加速比作为孤立的性能上限，并常常将它们的乘积视为联合上限。然而，在不规则工作负载中，难以预测的分支和缓存缺失的加载指令常常出现在相同的热循环中。消除一种惩罚会暴露另一种惩罚：更快的内存访问会更早触及误预测的分支，而更好的分支预测会在乱序执行窗口中留下更多长延迟加载。我们将这种交互称为共生停顿延迟（SSL）。本文使用联合加速比协同效应（JSS）来量化孤立上限何时失效，即观测到的联合完美分支/完美缓存加速比除以孤立加速比的乘积。大于1的值意味着独立上限分析低估了可获得的性能提升。在53个模拟工作负载中，70%表现出可测量的耦合（JSS > 1），尽管许多接近1，尤其是在压力较低的情况下。使用保守阈值时，40%超出独立性乘积6%以上，而SSO > 20的内核显示JSS在1.23到3.29之间。我们引入共生停顿机会（SSO），这是一种基于MPKI的轻量级筛选方法，用于筛选值得进行完整联合模拟的工作负载。我们将高SSO工作负载映射到四种反复出现的软件模式：邻居访问、哈希查找、链式结构遍历和数据依赖修改。我们将SSL与孤立完美模式下的重排序缓冲区占用率、取消率和提交饥饿联系起来。由此产生的方法很简单：使用SSO进行筛选，使用JSS进行验证，并在评估分支预测器、预取器、缓存或耦合的分支/内存机制时，报告条件性的缓存后分支增益和分支后缓存增益。我们的贡献是一个测量框架，展示孤立完美模式何时足够，以及何时低估了联合性能余量。
cs.AR / 2 / 2608.11577
Uni-SFU: Algorithm-HW Co-Design for Universal SFUs via Mixed-Degree Piecewise Approximation
Uni-SFU：通过混合阶分段逼近实现通用 SFU 的算法-硬件协同设计
Abstract
Nonlinear activation functions are essential to modern deep neural networks (DNNs), but their hardware evaluation places significant pressure on the special-function units (SFUs) of GPUs and custom accelerators. Therefore, piecewise polynomial approximations are commonly used within allowed error bounds to improve computational efficiency. However, existing techniques often approximate each activation function in isolation using fixed-degree polynomials and uniform segments, leading to hardware redundancy and sub-optimal precision. To address these limitations, we present Uni-SFU, an algorithm-hardware co-design framework that jointly optimizes approximation accuracy and silicon area for a diverse set of activation functions. Uni-SFU leverages a joint search across all target functions to assign mixed-degree polynomials to nonuniform segments, guided by an RTL-derived area cost model. This approach identifies a unified hardware configuration to implement the target activation functions under given accuracy constraints. Validated across over 700 neural network variants and three Natural Language Processing (NLP) models, Uni-SFU achieves a superior Mean Squared Error (MSE) below 8.22x10^-8, limiting top-1 accuracy degradation to within 1.02% compared to floating-point baselines. The proposed design occupies only 6,800 um2 in GF 22nm CMOS technology, achieving a superior trade-off between silicon area and system-level accuracy compared to SOTA counterparts.
Chinese Translation
非线性激活函数对于现代深度神经网络（DNN）至关重要，但它们的硬件计算给 GPU 和定制加速器的特殊函数单元（SFU）带来了巨大压力。因此，在允许的误差范围内通常使用分段多项式逼近来提高计算效率。然而，现有技术通常使用固定次数多项式和均匀分段来孤立地逼近每个激活函数，导致硬件冗余和次优精度。为了解决这些局限性，我们提出了 Uni-SFU，这是一个算法-硬件协同设计框架，能够针对多种激活函数联合优化逼近精度和硅面积。Uni-SFU 利用跨所有目标函数的联合搜索，在 RTL 导出的面积成本模型指导下，为不均匀分段分配混合次数多项式。该方法确定了一种统一的硬件配置，以在给定精度约束下实现目标激活函数。在超过 700 个神经网络变体和三个自然语言处理（NLP）模型上验证后，Uni-SFU 实现了低于 8.22x10^-8 的优越均方误差（MSE），与浮点基线相比，将 top-1 精度下降限制在 1.02% 以内。所提出的设计在 GF 22nm CMOS 工艺中仅占用 6,800 平方微米，与 SOTA 对应方案相比，在硅面积和系统级精度之间取得了更优越的权衡。
cs.AR / 3 / 2608.11668
A Full-Stack Characterization of High-Bandwidth Flash for KV-Centric LLM Serving
面向以 KV 为中心的 LLM 服务的高带宽闪存全栈特性刻画
Abstract
High-Bandwidth Flash (HBF) stacks NAND behind a wide, package-local interface, giving flash-scale capacity with far better read latency and bandwidth than an SSD. This makes it tempting to keep an SSD-style Mooncake KV-offloading stack and swap only the backing tier for HBF. We test that substitution with an extended TokenSim, four complete two-hour Qwen-Bailian production traces, five dense and mixture-of-experts models, and H100/B200 profiles. Serving gets worse, not better, and a cost-benefit model explains why. A faster far tier helps only when read I/O is the serving bottleneck, reads outweigh writes, and delivered bandwidth is sustainable. All three must hold together, and transient KV fails every one. The package trade that buys flash costs GPU near-tier capacity and bandwidth, so average end-to-end latency rises 2--5.5x and maximum SLO goodput falls 1.1--2.7x across H100 and B200. Serving is almost insensitive to HBF's own read/write latency, and base-die near-memory compute does not raise the flash tier's share of the critical path. The two-tier hierarchy keeps reuse in the near tier and hands HBF a write-heavy stream, so writes outnumber reads on every trace. A 3D-ICE model shows that stream drives the stack to its thermal limit well below peak bandwidth, and a TLC tier wears out sooner than a capacity-matched SSD pool. The faster device yields a slower system because the package gives up more than the medium returns. HBF is not the problem; using it as a faster SSD for transient KV is. It belongs in serving as a selective, reuse-aware, write-budgeted, and thermally coordinated tier, not as a drop-in SSD replacement.
Chinese Translation
高带宽闪存（HBF）将 NAND 堆叠在一个宽的封装本地接口之后，提供闪存级别的容量，并且读取延迟和带宽远优于 SSD。这使得保留 SSD 风格的 Mooncake KV 卸载栈、仅将后端层替换为 HBF 的做法颇具吸引力。我们使用扩展的 TokenSim、四条完整的两个小时的 Qwen-Bailian 生产轨迹、五个稠密模型和专家混合模型以及 H100/B200 配置来测试这种替换。服务性能变得更差，而不是更好，并且一个成本效益模型解释了原因。只有当读 I/O 成为服务瓶颈、读操作多于写操作、并且实际提供的带宽可持续时，更快的远端层才有帮助。这三个条件必须同时成立，而瞬态 KV 在每一条上都不满足。为换取闪存而做出的封装权衡牺牲了 GPU 近端层的容量和带宽，因此在 H100 和 B200 上，平均端到端延迟上升了 2--5.5倍，最大 SLO 有效吞吐下降了 1.1--2.7倍。服务性能几乎对 HBF 自身的读/写延迟不敏感，而基础裸片的近内存计算也没有提高闪存层在关键路径中所占的比例。两层层次结构将重用保留在近端层，并交给 HBF 一个以写为主的流，因此在每条轨迹上写操作都多于读操作。一个 3D-ICE 模型表明，该流将堆叠推到远低于峰值带宽的热极限；并且 TLC 层比容量匹配的 SSD 池更早磨损。更快的设备反而导致更慢的系统，因为封装所放弃的比介质所回报的更多。问题不在于 HBF；而在于将其当作更快的 SSD 用于瞬态 KV。它应当作为一种有选择性、感知重用、有写入预算并进行热协调的层来用于服务，而不是作为 SSD 的直接替代品。
cs.AR / 4 / 2608.11688
APEX: Adaptive Expert Prefetching for Memory-Efficient Edge MoE Inference
APEX：用于内存高效边缘 MoE 推理的自适应专家预取
Abstract
Mixture-of-Experts (MoE) models are attractive for edge deployment because they provide high model capacity while activating only a small subset of parameters per token, improving compute efficiency. However, MoE inference at the edge is fundamentally limited by memory. Expert parameters are large and often reside in off-chip memory due to capacity, cost, and power constraints, putting expert loading to the critical path. We present APEX: Adaptive Expert Prefetching, a predictive resource management framework that overlaps expert loading with useful computation. APEX introduces a lightweight prefetch router that predicts candidate experts before the attention block to dynamically fetch additional experts using a learned confidence model. This adaptive strategy achieves over 99% overlap accuracy, significantly outperforming fixed top-k prefetching techniques. APEX supports two execution modes: a correctness-preserving mode that guarantees exact routing semantics, and a stall-free mode that eliminates residual stalls by operating on available experts with negligible impact on application accuracy. Across multiple MoE models, the correctness-preserving mode reduces per-token latency by up to 26% and improves energy-delay product (EDP) by up to 41% over state-of-the-art baselines, while the stall-free mode provides additional efficiency gains with negligible impact on application accuracy. These results establish adaptive, confidence-driven expert prefetching as an effective approach for efficient MoE inference on edge systems.
Chinese Translation
混合专家（MoE）模型对边缘部署具有吸引力，因为它们在每个 token 仅激活一小部分参数的同时提供高模型容量，从而提高了计算效率。然而，边缘端的 MoE 推理从根本上受到内存的限制。专家参数很大，且由于容量、成本和功耗限制通常驻留在片外内存中，这使得专家加载处于关键路径上。我们提出了 APEX：自适应专家预取，这是一个预测性资源管理框架，可将专家加载与有用计算重叠执行。APEX 引入了一个轻量级预取路由器，在注意力模块之前预测候选专家，并利用学习到的置信度模型动态获取额外专家。这种自适应策略实现了超过 99% 的重叠准确率，显著优于固定的 top-k 预取技术。APEX 支持两种执行模式：一种保持正确性的模式，保证精确的路由语义；一种无停顿模式，通过在可用专家上运行来消除残余停顿，并且对应用准确性的影响可忽略不计。在多个 MoE 模型中，与最先进的基线相比，保持正确性模式将每个 token 的延迟最多降低了 26%，并将能量延迟积（EDP）最多提高了 41%；而无停顿模式则提供了额外的效率提升，对应用准确性的影响可忽略不计。这些结果确立了自适应、置信度驱动的专家预取作为边缘系统上高效 MoE 推理的一种有效方法。
cs.AR / 5 / 2608.11693
Spec Sheets Are Not Kernels: An ISA- and Source-Level Audit of INT8 Availability on NVIDIA Blackwell Ultra
规格表并非内核：NVIDIA Blackwell Ultra 上 INT8 可用性的 ISA 与源码级审计
Abstract
NVIDIA's published specifications give the Blackwell Ultra GPU (B300) a dense-compute ratio of roughly 30:1 between FP8 and INT8 tensor-core throughput; its predecessors, H200 and B200, both provide 1:1. We audit what this deprioritization means in practice by tracing INT8 W8A8 support through four layers of the stack: the published specifications, the PTX ISA, NVIDIA's CUTLASS kernel library, and the two major open-source LLM serving engines (vLLM and SGLang). We find a consistent, layered withdrawal: (i) the PTX ISA never exposes the fifth-generation tensor-core integer path (tcgen05.mma with .kind::i8) on sm_103a, even though the same PTX revision extends the FP4 kinds to that target, leaving legacy warp-level IMMA as the only architecturally legal integer tensor-core path on B300; (ii) CUTLASS's kernel generator explicitly skips INT8 UMMA generation for any build targeting 103a, while generating FP8 unconditionally; (iii) vLLM ships no INT8 GEMM for Blackwell and fails with a hard runtime error at the first forward pass, after the model has loaded; and (iv) SGLang's ahead-of-time INT8 GEMM stops at Sm90, while its FP8 tuning configurations already cover B200. We document an escape hatch (rerouting vLLM's INT8 path to a JIT-compiled Triton backend via an environment variable), a false-negative trap in the obvious profiler methodology for detecting "native INT8" on sm_103, and the practical failure semantics that make naive testing expensive. Together, these findings show that a quantization format's availability is a property of the whole stack rather than of the model or the spec sheet. Four distinct layers, three of them NVIDIA's own, withdrew INT8 support in mutually consistent ways, and a format that is nominally present on the datasheet is, by default, undeployable on this hardware.
Chinese Translation
NVIDIA 公布的规格表明，Blackwell Ultra GPU（B300）在 FP8 与 INT8 张量核心吞吐量之间的密集计算比率约为 30:1；其前代产品 H200 和 B200 均提供 1:1。我们通过追踪 INT8 W8A8 支持在栈的四个层面中的情况，来审计这种降优先级在实际中意味着什么：公布的规格、PTX ISA、NVIDIA 的 CUTLASS 内核库以及两个主要的开源 LLM 服务引擎（vLLM 和 SGLang）。我们发现了一种一致的、分层级的撤回：（i）PTX ISA 在 sm_103a 上从未暴露第五代张量核心整数路径（带 .kind::i8 的 tcgen05.mma），即使同一 PTX 修订版将 FP4 类型扩展到该目标，这使得传统 warp 级 IMMA 成为 B300 上唯一在架构上合法的整数张量核心路径；（ii）CUTLASS 的内核生成器对任何面向 103a 的构建显式跳过 INT8 UMMA 生成，同时无条件生成 FP8；（iii）vLLM 没有为 Blackwell 提供 INT8 GEMM，并且在模型加载后的第一次前向传播时以硬运行时错误失败；（iv）SGLang 的提前编译 INT8 GEMM 止步于 Sm90，而其 FP8 调优配置已经覆盖 B200。我们记录了一个逃生通道（通过环境变量将 vLLM 的 INT8 路径重路由到 JIT 编译的 Triton 后端）、在 sm_103 上检测“原生 INT8”的显而易见的 profiler 方法中的一个假阴性陷阱，以及使朴素测试代价高昂的实际失败语义。综合来看，这些发现表明，一种量化格式的可用性是整个栈的属性，而非模型或规格表的属性。四个不同的层面，其中三个是 NVIDIA 自己的，以相互一致的方式撤回了 INT8 支持，而一种在数据手册上名义上存在的格式，在这款硬件上默认是不可部署的。
cs.AR / 6 / 2608.12140
FQTree: Fine-grained Quantization and Hardware Generation of Boosted Decision Trees
FQTree：提升决策树的细粒度量化与硬件生成
Abstract
Boosted decision trees (BDTs) are widely used in latency-critical applications, but efficient hardware deployment remains challenging. Existing designs often rely on uniform or manually tuned fixed-point formats, which can introduce unnecessary hardware cost or accuracy loss. This work presents the FQTree algorithm{https://github.com/ecs-bristol/FQTree} for fine-grained quantization-aware training of BDTs, together with the QXGB framework for automatic hardware generation. FQTree introduces a hardware-oriented leaf-value quantization scheme that uses a global quantization step together with a tree-wise shift, enabling compact non-negative integer leaf representations, controlled clipping/pruning, and bias folding to reduce datapath cost. This work further applies this quantization during boosting so that later trees adapt to the errors of the already-quantized ensemble, and then lowers the trained model into low-latency hardware implementations through a compiler-based flow. Results on JSC, MNIST, and NID show that our method reduces LUT usage by 26-57\% compared with the state-of-the-art FPGA-based BDT designs while matching or improving accuracy.
Chinese Translation
提升决策树（BDTs）广泛用于对延迟敏感的应用，但高效的硬件部署仍具挑战性。现有设计通常依赖统一或手动调整的定点格式，这可能带来不必要的硬件成本或精度损失。本文提出了 FQTree 算法{https://github.com/ecs-bristol/FQTree}，用于 BDT 的细粒度量化感知训练，以及用于自动硬件生成的 QXGB 框架。FQTree 引入了一种面向硬件的叶值量化方案，该方案使用全局量化步长和逐树移位，实现了紧凑的非负整数叶表示、受控裁剪/剪枝以及偏置折叠，以降低数据通路成本。本文进一步在提升过程中应用这种量化，使后续树能够适应已量化集成的误差，然后通过基于编译器的流程将训练后的模型转化为低延迟硬件实现。在 JSC、MNIST 和 NID 上的结果表明，与最先进的基于 FPGA 的 BDT 设计相比，我们的方法将 LUT 使用量减少了 26-57%，同时保持或提高了精度。
cs.AR / 7 / 2608.11719
Do Not Let CNOTs Overwhelm the Decoder: Scheduling Transversal Gates for Fast FTQC
不要让 CNOT 门压垮解码器：为快速 FTQC 调度横向门
Abstract
Transversal CNOT (TCNOT) gates can accelerate fault-tolerant quantum computation (FTQC) in the surface code by reducing the number of syndrome extraction rounds required between logical operations from $O(d)$ to $O(1)$. This is particularly attractive for quantum platforms with long-range connectivity, such as neutral atoms. However, dense TCNOT schedules substantially increase the classical decoding workload. TCNOTs propagate errors across multiple surface-code patches, enlarging the spatiotemporal region that must be decoded jointly. Consequently, denser TCNOT schedules increase decoding latency and memory requirements and potentially exceed available decoder capacity. Moreover, because the detector error model (DEM) of each decoding window depends on the TCNOT schedule, exhaustively precomputing all possible window-level DEMs is infeasible, requiring just-in-time (JIT) DEM compilation. Thus, the practical benefit of TCNOT gates is limited not only by quantum hardware performance but also by classical decoding and DEM-compilation capacity. We introduce PACE, a decoder-aware scheduling framework for TCNOT-based FTQC. PACE first mitigates the decoder-side costs of aggressive TCNOT scheduling through three complementary techniques. Hybrid Window Decoding assigns different decoders for each decoding window according to its DEM structure. DEM Stitch generates schedule-specific window-level DEMs just in time by assembling reusable precompiled fragments. Sub-window Parallel Decoding decomposes large windows into smaller sub-windows with graph-coloring formulation. Building on these techniques, PACE then performs decoder-aware scheduling to maximize TCNOT concurrency within the available decoder resources. Our evaluation shows the trade-off between quantum acceleration and classical decoding cost, revealing the limitations of current decoding systems for TCNOT-based FTQC.
Chinese Translation
横向 CNOT（TCNOT）门能够将逻辑操作之间所需的症候提取轮数从 $O(d)$ 减少到 $O(1)$，从而加速表面码中的容错量子计算（FTQC）。这对于具有长程连通性的量子平台（例如中性原子）尤其有吸引力。然而，密集的 TCNOT 调度会大幅增加经典解码工作负载。TCNOT 会在多个表面码补丁之间传播错误，从而扩大必须联合解码的时空区域。因此，更密集的 TCNOT 调度会增加解码延迟和内存需求，并可能超出可用的解码器容量。此外，由于每个解码窗口的探测器错误模型（DEM）取决于 TCNOT 调度，穷举地预计算所有可能的窗口级 DEM 是不可行的，因此需要即时（JIT）DEM 编译。因此，TCNOT 门的实际收益不仅受量子硬件性能的限制，还受经典解码和 DEM 编译能力的限制。我们提出了 PACE，一个面向基于 TCNOT 的 FTQC 的解码器感知调度框架。PACE 首先通过三种互补技术来缓解激进 TCNOT 调度的解码器侧开销。混合窗口解码（Hybrid Window Decoding）根据每个解码窗口的 DEM 结构为其分配不同的解码器。DEM 拼接（DEM Stitch）通过组装可复用的预编译片段，即时生成针对特定调度的窗口级 DEM。子窗口并行解码（Sub-window Parallel Decoding）利用图着色形式化方法将大窗口分解为更小的子窗口。在这些技术的基础上，PACE 随后执行解码器感知调度，以在可用的解码器资源范围内最大化 TCNOT 并发性。我们的评估展示了量子加速与经典解码成本之间的权衡，揭示了当前解码系统在基于 TCNOT 的 FTQC 中的局限性。
密码学与安全 (cs.CR)
17
cs.CR / 1 / 2608.11337
Association-based Privacy Attacks in Wireless Protocols: Formal Modeling and Mitigation
无线协议中基于关联的隐私攻击：形式化建模与缓解
Abstract
With the surge in privacy-sensitive data from sources such as social media and IoT devices, there is a pressing need for formal, automated methods to assess privacy risks within these intricate systems. This paper formally investigates root sources of pairing-based privacy threats exploited using replay/relay techniques in wireless communication. Our research harnesses condition-oblivious responses, replay-resistance, and distance bounding measures vital for protocols utilizing shared keys in allowlists for authenticated reconnections. Particularly, the paper uses formal modeling of notable wireless networks, like the Wi-Fi P2P persistent group formation and the Bluetooth Low Energy reconnection procedure, to illustrate the root causes and countermeasures. Our model rigorously validates the proposed solution against association inference attacks, along with existing formalizations of well-authentication, frame opacity, and no-desynchronization. The ensuing analysis reveals not only uncharted privacy realms in wireless communication but also identifies old and new vulnerabilities. Our proposed design changes are acknowledged by Wi-Fi Alliance and Bluetooth SIG, paving the way for future advancements in resilient, privacy-preserving wireless protocols.
Chinese Translation
随着来自社交媒体和物联网设备等来源的隐私敏感数据激增，迫切需要用于评估这些复杂系统中隐私风险的形式化、自动化方法。本文形式化地研究了无线通信中通过重放/中继技术加以利用的基于配对的隐私威胁的根源。我们的研究利用了条件无关响应、抗重放和距离界定措施，这些措施对于在允许列表中使用共享密钥进行认证重连的协议至关重要。特别是，本文使用对著名无线网络的形式化建模，如 Wi-Fi P2P 持久群组形成和蓝牙低能耗重连过程，来阐明其根本原因与对策。我们的模型严格验证了所提解决方案抵御关联推断攻击的能力，并验证了关于良好认证、帧不透明性和无去同步化的现有形式化。随后的分析不仅揭示了无线通信中未被探索的隐私领域，还识别出了旧有和新出现的漏洞。我们提出的设计变更得到了 Wi-Fi 联盟和蓝牙技术联盟的认可，为未来在弹性、隐私保护无线协议方面取得进展铺平了道路。
cs.CR / 2 / 2608.11392
AI Guardrail Survival under Single-Cycle Agentic Self-Summarization
单周期代理式自摘要下的 AI 护栏存续
Abstract
Long-running agents periodically compact their context, replacing the transcript with a model-generated summary.Recent work shows that dropping a standing safety constraint during compaction drives behavioral violations acrossmany models (Governance Decay; Chen, 2026). We ask a finer question: under a single compaction cycle, how is a safetyrule lost, and what does that imply for detection and evaluation? Our central finding is that a presence check is not asafety check: when compaction does not drop a rule outright, it often leaves something that looks like a rule but doesnot act like one. On behavioral replay, a degraded residue leads the model to perform the prohibited action far more oftenthan an intact welded rule does (all-case gaps of +34 and +57 points under two replay models, both positive), category-level survival behaves like a residue, and even intact rules sometimes fail to fire, so an audit that checks only textualpresence gives false assurance. Sharpening this, rule-form items are retained substantially more often than prominence-matched facts, which is exactly why presence-based checking feels adequate even though survival is not protection.Textual loss is regime-dependent (weld-or-drop with a single rule; degraded predicate-loss residues under a tighterbudget), and we did not observe the hypothesized textual severing mode. Such loss is silent at runtime and detectableonly by comparison with retained external ground truth (such as a constraint registry), which reveals textual absence butnot whether a surviving rule still fires. We also document evaluation pitfalls where LLM-judge labels alone would havereversed a conclusion. All results concern a single compaction cycle.
Chinese Translation
长期运行的智能体会定期压缩其上下文，将记录替换为模型生成的摘要。最近的工作表明，在压缩期间丢弃一条长期有效的安全约束会在许多模型中引发行为违规（治理衰退；Chen, 2026）。我们提出一个更细致的问题：在单个压缩周期中，安全规则是如何丢失的，这对检测和评估意味着什么？我们的核心发现是，存在性检查不是安全性检查：当压缩没有彻底丢弃规则时，它往往会留下一些看起来像规则但行为不像规则的东西。在行为回放中，退化的残留物导致模型执行被禁止操作的频率远高于完整“焊接”规则的情况（在两个回放模型下，全案例差距分别为 +34 和 +57 个百分点，两者均为正），类别层面的存续表现得像残留物，而且即使是完整的规则有时也未能触发，因此仅检查文本存在性的审计会给出虚假的保证。进一步明确地说，规则形式的条目比显著性匹配的事实被保留的频率高得多，这正是为什么基于存在性的检查看似足够，尽管存续并不等于保护。文本损失取决于具体机制（在单条规则下为“焊接或丢弃”；在更紧预算下为退化的谓词丢失残留），并且我们没有观察到假设的文本切断模式。这种损失在运行时是静默的，只有通过与保留的外部真值（例如约束登记表）进行比较才能检测到，而这种比较揭示的是文本缺失，而不是存留下来的规则是否仍会触发。我们还记录了评估中的陷阱：仅凭 LLM 评判者的标签本会逆转某个结论。所有结果都涉及单个压缩周期。
cs.CR / 3 / 2608.11418
A Study of Kernel Telemetry Options for Security-Oriented Provenance
面向安全溯源的内核遥测选项研究
Abstract
Provenance aims to capture the origins, transformations, and interactions of system objects for security and forensic applications. Existing provenance capture approaches still face major challenges and are not yet ready for production environments. In this paper, we first analyze the main kernel telemetry capture approaches, identifying eBPF as the most promising, and complement this analysis with micro benchmarks to assess its performance overhead and the filtering mechanisms used to achieve capture granularity, such as restricting capture to individual containers. Building on this foundation, we then classify, according to the studied capture approaches and filtering methods, eight provenance systems and five capture agents that could serve as their capture layers, collectively referred to as tools. Our study reveals that these tools are built on highly heterogeneous capture layers, most of which cannot guarantee the integrity and availability of the captured events, completely failing to meet the requirements of security-oriented use cases.
Chinese Translation
溯源旨在为安全和取证应用捕获系统对象的来源、转换和交互。现有的溯源捕获方法仍然面临重大挑战，尚未准备好用于生产环境。在本文中，我们首先分析主要的内核遥测捕获方法，确定 eBPF 是最有前景的方法，并通过微基准测试补充这一分析，以评估其性能开销以及用于实现捕获粒度的过滤机制，例如将捕获限制到单个容器。在此基础上，我们根据所研究的捕获方法和过滤方法，对八个溯源系统和五个可作为其捕获层的捕获代理进行分类，统称为工具。我们的研究表明，这些工具构建在高度异构的捕获层之上，其中大多数无法保证被捕获事件的完整性和可用性，完全无法满足面向安全用例的要求。
cs.CR / 4 / 2608.11436
When Agents Talk: Honeytokens under Shared Memory
当智能体交谈时：共享记忆下的蜜标
Abstract
During a 2026 cyber-capability evaluation, short-lived AI agents turned a shared package repository into persistent memory, passing exploit findings to later agents and rebuilding the channel after it was removed. The broader evaluation culminated in an intrusion into Hugging Face. This episode raises a question for defensive deception: can a honeytoken be harmless to trusted agents without becoming recognisable to an attacker who shares their information and can implement the trusted policy? The answer is no. A trusted rule that selects genuine objects while avoiding decoys can be copied by the attacker, while a total-variation bound limits legitimate compatibility when decoys resemble genuine objects. Shared memory creates a second leakage channel by pooling weak fingerprints. For a fixed candidate, repeated non-triggering probes drive the minimum Bayes classification error to zero when type-dependent response laws differ and are known or learnable. If probing triggers containment, learning also requires the coalition to remain active long enough. Transfer across objects requires a stable deployment rule and information that orients the classes. A separate detection bound distinguishes reliable token activation from reliable attack coverage. The architectural response is to keep token identity in a private reference monitor and route legitimate agents through a provenance-enforcing broker. This produces high-confidence detection only for a specified policy violation. Honeytokens remain useful sensors, but a separate security boundary is still required.
Chinese Translation
在2026年的一次网络能力评估中，短生命周期的AI智能体将共享软件包仓库变成了持久记忆，把漏洞利用发现传递给后来的智能体，并在该通道被移除后加以重建。更广泛的评估最终导致了对Hugging Face的一次入侵。这一事件为防御性欺骗提出了一个问题：蜜标能否在对受信任智能体无害的同时，又不被共享其信息且能够实施受信任策略的攻击者识别出来？答案是否定的。选择真实对象同时避开诱饵的受信任规则可以被攻击者复制，而当诱饵与真实对象相似时，总变差界限限制了合法兼容性。共享记忆通过汇集弱指纹形成了第二条泄露通道。对于固定候选对象，当类型依赖的响应规律不同且已知或可学习时，重复的非触发探测会将最小贝叶斯分类误差降至零。如果探测触发遏制，学习还要求该联盟保持足够长时间的活跃。跨对象迁移需要稳定的部署规则以及能够确定类别方向的信息。一个单独的检测界限区分了可靠的令牌激活与可靠的攻击覆盖。架构上的应对措施是将令牌身份保存在私有的引用监视器中，并通过强制执行来源的代理来路由合法智能体。这只会对特定的策略违规产生高置信度检测。蜜标仍然是有用的传感器，但仍然需要单独的安全边界。
cs.CR / 5 / 2608.11469
The Next Challenge for Agentic Cybersecurity: A Realistic, Contamination-Free Reverse Engineering Benchmark
智能体网络安全的下一个挑战：一个现实、无污染的逆向工程基准
Abstract
AI agents are rapidly improving in cybersecurity capabilities when the source code is available for analysis, yet much of the software most consequential to cybersecurity, including malware, firmware, and proprietary applications, is available only as binaries. Analyzing such software requires reverse engineering(RE): recovering program semantics before the analysis can be meaningfully performed. However, evaluating agentic RE poses a fundamental challenge: benchmark instances must be unseen as source code in the LLMs' training data to prevent models from taking shortcuts by recognizing them rather than really analyzing them, while also matching the scale and anti-analysis protections of real software. Unfortunately, however, existing benchmarks do not jointly satisfy these requirements. To this end, we introduce SRE-Bench, the first realistic, contamination-free RE benchmark. Built entirely from scratch by RE experts with over 5,000 hours, SRE-Bench comprises 19 private, real-world-scale programs averaging 16.9K lines of code. We further developed 44 in-house anti-analysis primitives, yielding 262 binary instances and 1572 deterministically graded tasks. Our evaluation across five frontier LLMs (GPT-5.6-sol,Claude-Opus-5,GPT-5.5,Grok-4.5, and GLM-5.2) shows that RE remains largely unsolved: the strongest model, GPT-5.6-sol, scores 61.4% per instance, and fully solves only 31.5% of the instances. Our analysis further reveals that agents behave differently from human engineers, where agents are relatively insensitive to compiler optimization and static linking. Controlled ablations also confirm that both contamination control and realistic scale are essential. These results indicate that strong source-code security capabilities do not yet transfer to binary analysis, highlighting RE as an important frontier for agentic cybersecurity and SRE-Bench as a rigorous testbed to measure progress.
Chinese Translation
当源代码可用于分析时，AI 智能体在网络安全能力方面正在快速提升，然而对网络安全最为重要的许多软件——包括恶意软件、固件和专有应用程序——仅以二进制形式提供。分析此类软件需要逆向工程（RE）：在能够有意义地执行分析之前恢复程序语义。然而，评估智能体逆向工程面临一个根本性挑战：基准实例必须在大语言模型的训练数据中不作为源代码出现，以防止模型通过识别它们而非真正分析它们来走捷径，同时还要匹配真实软件的规模和反分析保护。然而，不幸的是，现有基准并未同时满足这些要求。为此，我们引入了 SRE-Bench，这是第一个现实、无污染的逆向工程基准。SRE-Bench 完全由逆向工程专家从零开始构建，耗时超过 5000 小时，包含 19 个私有的、具有真实世界规模且平均包含 16.9K 行代码的程序。我们进一步开发了 44 个内部反分析原语，从而得到 262 个二进制实例和 1572 个确定性评分任务。我们对五个前沿大语言模型（GPT-5.6-sol、Claude-Opus-5、GPT-5.5、Grok-4.5 和 GLM-5.2）的评估表明，逆向工程在很大程度上仍未解决：最强模型 GPT-5.6-sol 的每实例得分为 61.4%，并且仅完全解决了 31.5% 的实例。我们的分析进一步揭示，智能体的行为与人类工程师不同，智能体对编译器优化和静态链接相对不敏感。受控消融实验也证实，污染控制和现实规模都至关重要。这些结果表明，强大的源代码安全能力尚未迁移到二进制分析，凸显了逆向工程是智能体网络安全的重要前沿，而 SRE-Bench 是衡量进展的严格测试平台。
cs.CR / 6 / 2608.11489
A Runtime Decentralized Attestation and Coordinated Repair Framework for Securing Automotive ECUs
一种用于保障汽车电子控制单元安全的运行时去中心化证明与协同修复框架
Abstract
The evolution of automotive technology increasingly integrates components, transforming vehicles into interconnected systems of systems. Modern vehicles are controlled by a distributed system of computing devices, known as electronic control units (ECUs). However, this interconnectedness means that any error poses significant risks to the vehicle operator. In particular, malware can be injected into ECUs, threatening vehicle safety. To address this, we need mechanisms to detect compromised ECUs then repair them to a benign state. Existing approaches mainly focus on detection and do not address the challenge of integrating detection with runtime ECU repair. This integration is nontrivial because runtime repair involves both local rollback and reboot with timing determined from global vehicle context to avoid unsafe behavior. In this work, we have designed DACER, a runtime decentralized attestation and coordinated repair framework for automotive ECUs. DACER is the first approach that co-designs attestation and repair to unify the ``local'' nature of firmware rollback with the ``global'' nature of ECU reboot. In DACER, each ECU performs efficient local self-attestation and self-repair functions, enabling low-overhead coordination for distributed operations. In addition, DACER takes advantage of the hierarchical vehicle computing architecture. Our resulting DACER design checks the entire state of the vehicle, resists single points of failure, conforms to real-time constraints, and enables firmware restoration during runtime. The key functions are enabled by the ARM TrustZone equipped within each ECU and the secure flash memory controller embedded in the storage device. We implemented DACER on real-world hardware and experimentally demonstrated its low overhead.
Chinese Translation
汽车技术的演进日益集成各种组件，将车辆转变为相互连接的系统之系统。现代车辆由被称为电子控制单元（ECU）的分布式计算设备系统控制。然而，这种互联性意味着任何错误都会对车辆操作者构成重大风险。特别是，恶意软件可被注入ECU，威胁车辆安全。为了解决这一问题，我们需要能够检测受攻击ECU并将其修复到良性状态的机制。现有方法主要集中在检测上，并未解决将检测与运行时ECU修复相集成的挑战。这种集成并非易事，因为运行时修复既涉及本地回滚，也涉及重启，而重启时机需根据全局车辆上下文确定，以避免不安全行为。在本工作中，我们设计了DACER，一个面向汽车ECU的运行时去中心化证明与协同修复框架。DACER是首个将证明与修复协同设计、从而将固件回滚的“本地”本质与ECU重启的“全局”本质统一起来的方法。在DACER中，每个ECU执行高效的本地自证明和自修复功能，从而以低开销实现分布式操作的协调。此外，DACER利用了分层车辆计算架构。我们的DACER设计能够检查车辆整体状态，抵御单点故障，满足实时约束，并支持运行时固件恢复。这些关键功能由每个ECU中配备的ARM TrustZone以及嵌入存储设备中的安全闪存控制器实现。我们在真实硬件上实现了DACER，并通过实验证明了其低开销。
cs.CR / 7 / 2608.11492
Cross-Corpus Evaluation of Generalizable Vulnerability Detection in IoT Firmware
物联网固件中可泛化漏洞检测的跨语料库评估
Abstract
IoT firmware vulnerability detection remains challenging due to heterogeneous firmware ecosystems, resource-constrained platforms, and limitations in existing benchmarks. Many datasets are synthetic or general-purpose and lack human-verified, contamination-screened annotations, limiting evidence on cross-corpus generalization across training sources, model architectures, and curriculum strategies. To address this gap, this paper introduces IoTVulBench, a human-verified benchmark for cross-corpus firmware vulnerability detection. IoTVulBench-Core was constructed from GitHub repositories, validated by three expert reviewers, and evaluated on a contamination-screened held-out target across five model architectures, two tuning methods, and three curriculum strategies, with ensemble, distillation, and robustness analyses. Models trained on IoTVulBench achieved the highest MCC among matched single-source datasets, reaching 0.58 versus 0.44 for PrimeVul and 0.39 for D2A. Staged curriculum learning increased MCC to 0.69, while a diversity-optimized ensemble achieved 0.73, improving by 0.42 MCC over the strongest reference comparator, a static analyzer at 0.31, and by 0.29 over PrimeVul. At a 0.5% false-positive rate, the model missed only 21% of vulnerabilities, compared with 71% for the strongest comparator. It retained 86% of its performance under identifier renaming and demonstrated strong calibration and largely faithful explanations. These findings indicate that domain-matched training data and curriculum design, rather than model scale alone, are key drivers of generalization in firmware vulnerability detection. The results provide a benchmark for future research and deployment-ready configurations for practical IoT security applications.
Chinese Translation
由于固件生态系统异构、平台资源受限以及现有基准的局限性，物联网固件漏洞检测仍然具有挑战性。许多数据集是合成的或通用的，缺乏经过人工验证和污染筛查的标注，这限制了对跨训练来源、模型架构和课程策略的跨语料泛化能力的证据。为弥补这一空白，本文提出了IoTVulBench，一个用于跨语料固件漏洞检测的人工验证基准。IoTVulBench-Core由GitHub仓库构建，经三位专家评审验证，并在一个经污染筛查的留出目标集上，跨五种模型架构、两种微调方法和三种课程策略进行评估，同时进行了集成、蒸馏和稳健性分析。在IoTVulBench上训练的模型在匹配的单源数据集中取得了最高的MCC，达到0.58，而PrimeVul为0.44，D2A为0.39。分阶段课程学习将MCC提高到0.69，而多样性优化的集成模型达到0.73，比最强的参照比较对象（MCC为0.31的静态分析器）提高了0.42个MCC，比PrimeVul提高了0.29。在0.5%的假阳性率下，该模型仅漏检了21%的漏洞，而最强比较对象漏检了71%。在标识符重命名条件下，它保持了86%的性能，并表现出良好的校准性和大体忠实的解释。这些发现表明，领域匹配的训练数据和课程设计——而非仅靠模型规模——是固件漏洞检测泛化能力的关键驱动因素。这些结果为未来研究提供了基准，并为实际物联网安全应用提供了可部署的配置。
cs.CR / 8 / 2608.11526
Towards Scalable Fuzzy PSI via Efficient Fuzzy Matching
通过高效模糊匹配实现可扩展的模糊PSI
Abstract
In this paper, we present scalable fuzzy PSI protocols for general $L_{p \in [1, \infty]}$ distance, supporting both low- and high-dimensional sets. The core technique is two efficient fuzzy matching protocols. The first is built from a role-reversed oblivious PRF (OPRF) and realizes $O(d\log δ)$ overhead, compared to $O((\log δ)^d)$ in previous works. The second leverages customized oblivious transfer (OT) with $O(d\ell)$ overhead, where $\ell$ is the bit length of inputs, which is particularly suitable for short inputs. With these new techniques, we further propose a new dual-layer hashing framework for fuzzy PSI over low-dimensional sets, instantiated with our OT-based fuzzy matching and enhanced with a domain reduction optimization. The protocols achieve an overhead linear with $n, m, \log δ, 2^d$, without the $O((\log δ)^d)$ or $O(δ)$ factors present in prior works. {For high-dimensional sets, we construct fuzzy PSI protocols based on our OPRF- and OT-based fuzzy matching, which achieve an asymptotic overhead linear with $n, m, d$, and $\log δ$ but rely on the strong globally disjoint assumption.} Extensive evaluations demonstrate that our protocols achieve up to a $145\times$ speedup in running time and a $20\times$ reduction in communication cost compared to van Baarsen and Pu~(ASIACRYPT'25), and achieve up to a $25\times$ speedup in running time and up to a $17\times$ reduction in communication cost compared to Piske et al.~(CCS'25).
Chinese Translation
在本文中，我们针对一般的 $L_{p \in [1, \infty]}$ 距离提出了可扩展的模糊PSI协议，同时支持低维集合和高维集合。核心技术是两个高效的模糊匹配协议。第一个协议基于角色反转的不经意伪随机函数（OPRF）构建，其开销为 $O(d\log δ)$，而先前工作中的开销为 $O((\log δ)^d)$。第二个协议利用定制的不经意传输（OT），其开销为 $O(d\ell)$，其中 $\ell$ 是输入的比特长度，特别适用于短输入。在这些新技术的基础上，我们进一步针对低维集合上的模糊PSI提出了一个新的双层哈希框架，该框架使用我们基于OT的模糊匹配进行实例化，并通过域缩减优化进行增强。这些协议实现了与 $n, m, \log δ, 2^d$ 成线性关系的开销，并且不存在先前工作中出现的 $O((\log δ)^d)$ 或 $O(δ)$ 因子。{对于高维集合，我们基于我们的OPRF和OT模糊匹配构造了模糊PSI协议，这些协议实现了与 $n, m, d$ 和 $\log δ$ 线性相关的渐近开销，但依赖于强全局不相交假设。}大量实验评估表明，与 van Baarsen 和 Pu (ASIACRYPT'25) 相比，我们的协议在运行时间上实现了高达 $145\times$ 的加速，在通信成本上实现了高达 $20\times$ 的降低；与 Piske 等人 (CCS'25) 相比，在运行时间上实现了高达 $25\times$ 的加速，在通信成本上实现了高达 $17\times$ 的降低。
cs.CR / 9 / 2608.11730
Plaintext Recovery Against Post-Filtering Access Control
针对后过滤访问控制的明文恢复
Abstract
Fine-grained access control (FGAC) mechanisms such as row-level security (RLS) and document-level security (DLS) are widely deployed in databases to restrict access to data stored in physical indexing structures shared by multiple users (e.g., in multi-tenant databases, or in the implementation of least-privilege within an organization). FGAC implementations often use post-filtering where untrusted queries run over all data and private results are redacted afterwards. Prior work shows this approach can lead to side-channels that enable attackers to test if a chosen value exists in unseen data. While damaging, prior attacks do not enable the efficient recovery of rich, high-entropy data like full records or text documents. We show these side-channels are more damaging than previously thought. Using rich query interfaces (e.g., range, prefix, and conjunctive predicates), we amplify existence leakage into reconstruction attacks. We do this in two settings: - PostgreSQL (RLS timing). We exploit a timing side-channel and expressive SQL queries (e.g., ranges, conjunctions) to enumerate unknown attribute values and, in turn, full records via binary search over large domains. - Elasticsearch/OpenSearch (DLS scoring). We exploit scoring and prefix-expansion side-channels to recover indexed terms from documents. In some cases, we can extract $n$-grams in the corpus to recover approximate text. Our results show that FGAC side-channels must be evaluated in the presence of rich predicates, which can turn membership tests into scalable reconstruction of high-entropy records.
Chinese Translation
细粒度访问控制（FGAC）机制，如行级安全（RLS）和文档级安全（DLS），被广泛部署在数据库中，以限制对多个用户共享的物理索引结构中存储的数据的访问（例如，在多租户数据库中，或在组织内实施最小权限时）。FGAC 实现通常使用后过滤，即不受信任的查询在所有数据上运行，之后私有结果被屏蔽。先前的工作表明，这种方法可能导致侧信道，使攻击者能够测试所选值是否存在于未可见的数据中。虽然具有破坏性，但先前的攻击并不能有效恢复丰富的、高熵的数据，如完整记录或文本文档。我们表明这些侧信道比之前认为的更具破坏性。利用丰富的查询接口（例如，范围、前缀和合取谓词），我们将存在性泄漏放大为重构攻击。我们在两种场景中做到这一点：- PostgreSQL（RLS 计时）。我们利用计时侧信道和表达丰富的 SQL 查询（例如，范围、合取）来枚举未知属性值，并进而通过在大域上进行二分搜索来枚举完整记录。- Elasticsearch/OpenSearch（DLS 评分）。我们利用评分和前缀扩展侧信道来从文档中恢复索引词项。在某些情况下，我们可以提取语料库中的 $n$-grams 来恢复近似文本。我们的结果表明，FGAC 侧信道必须在存在丰富谓词的情况下进行评估，这些谓词可以将成员资格测试转化为高熵记录的可扩展重构。
cs.CR / 10 / 2608.11799
AmbSentry: Mitigating Sensing Eavesdropping in ISAC Systems by Harnessing Ambient IoT Devices
AmbSentry：利用环境物联网设备缓解ISAC系统中的感知窃听
Abstract
Integrated sensing and communication (ISAC) has emerged as a pivotal paradigm for 6G networks, enabling the synergistic convergence of spectral and hardware resources to maximize system efficiency. However, the inherent openness of wireless transmission exposes ISAC systems to critical security risks, particularly regarding the privacy of the sensing information. Unauthorized sensing eavesdroppers can extract sensitive target parameters (e.g., range and velocity) by directly estimating open sensing echo channels, rendering traditional data-based protection techniques ineffective. To mitigate this threat, this paper proposes AmbSentry, an ISAC system that prevents the leakage of sensing information to sensing eavesdroppers by harnessing naturally distributed passive ambient IoT (AIoT) devices. Specifically, these AIoT devices are strategically configured to act as cooperative jammers and ghost targets, introducing controllable interference into the sensing environment. Based on the proposed system, we formulate a joint optimization problem to maximize the integrated sidelobe level at the eavesdropper under quality-of-service (QoS) constraints, thereby degrading sensing eavesdropping performance while maintaining sensing and communication performance for legitimate receivers. Since the problem is non-convex, we further develop an efficient iterative algorithm to cooperatively design the transmit beamforming at the base station and the reflection modulations of the AIoT devices based on Dinkelbach transformation and block coordinate descent methods. The detailed results also demonstrate that AmbSentry significantly enhances sensing security, allowing the legitimate sensing receiver to achieve a 14-dB SNR advantage in detection probability and a hundred times lower estimation error compared to the eavesdropper.
Chinese Translation
集成感知与通信（ISAC）已成为6G网络的关键范式，能够协同融合频谱与硬件资源以最大化系统效率。然而，无线传输固有的开放性使ISAC系统面临严峻的安全风险，尤其是感知信息的隐私问题。未经授权的感知窃听者可通过直接估计开放的感知回波信道来提取敏感的目标参数（如距离和速度），使得传统的基于数据的保护技术失效。为缓解这一威胁，本文提出AmbSentry，这是一种通过利用自然分布的无源环境物联网（AIoT）设备来防止感知信息泄露给感知窃听者的ISAC系统。具体而言，这些AIoT设备被策略性地配置为协作干扰器和虚假目标，在感知环境中引入可控干扰。基于所提出的系统，我们构建了一个联合优化问题，在服务质量（QoS）约束下最大化窃听者处的综合旁瓣电平，从而在维持合法接收端感知与通信性能的同时降低感知窃听性能。由于该问题是非凸的，我们进一步开发了一种高效的迭代算法，基于Dinkelbach变换和块坐标下降法协同设计基站处的发射波束成形与AIoT设备的反射调制。详细结果还表明，AmbSentry显著增强了感知安全性，使得合法感知接收端相比窃听者在检测概率方面获得14 dB的信噪比优势，并且估计误差低一百倍。
cs.CR / 11 / 2608.11802
Towards Model-based Run-time Cybersecurity: On Control-Flow Anomaly Detection, Attack Identification, and Hardware Monitoring
面向基于模型的运行时网络安全：关于控制流异常检测、攻击识别与硬件监控
Abstract
Methods to increase the resilience of systems to cyber-attacks become increasingly important. Control-flow monitoring provides a principled basis to ensure integrity and detect possible anomalies at run-time. Once anomalies have been detected, so-called attack trees can be used to identify possible types of attacks. However, this approach is vulnerable to camouflage, by which attackers try to evade detection (and correct identification) by deliberately manipulating also the system's observed control flow. In this paper, we outline a model-based approach that provides more robust intrusion detection and attack identification through an architecture that combines software- with hardware-based monitoring. In this approach, software-level observation indicates suspicious activities, while hardware-level monitoring checks them separately in more detail, making it much harder for attacks to camouflage themselves and go undetected. We illustrate the approach with an authentication-service example that captures a realistic failure mode: a software-level observer sees an anomalous but apparently harmless control-flow deviation, maps it to a benign root cause in an attack tree, but misses the true intrusion. A second, independent hardware control-flow monitor observes the actual transition sequence and thereby changes the attack-tree diagnosis from a low-severity configuration or maintenance issue to a high-confidence code-injection or control-flow hijack. In this scenario, the proposed combination of control-flow anomaly detection, attack-tree based intrusion identification, and hardware-based monitoring can improve not only anomaly detection, but also the diagnostic precision of attack-tree-based cyber-attack identification.
Chinese Translation
提高系统对网络攻击的抵御能力的方法变得越来越重要。控制流监控为确保完整性并在运行时检测可能异常提供了有原则的基础。一旦检测到异常，就可以使用所谓的攻击树来识别可能的攻击类型。然而，这种方法容易受到伪装的影响，攻击者通过故意操纵系统被观测到的控制流来试图逃避检测（以及正确识别）。在本文中，我们概述了一种基于模型的方法，该方法通过将基于软件的监控与基于硬件的监控相结合的架构，提供更稳健的入侵检测和攻击识别。在这种方法中，软件层面的观察指示可疑活动，而硬件层面的监控则分别更详细地检查这些活动，从而使攻击更难自我伪装并不被察觉。我们通过一个身份验证服务示例来说明该方法，该示例捕捉了一种现实的故障模式：软件层面的观察者看到了一个异常但看似无害的控制流偏差，将其在攻击树中映射为一个良性根本原因，却错过了真正的入侵。第二个独立的硬件控制流监控器观察实际的转换序列，从而将攻击树诊断从低严重性的配置或维护问题转变为高置信度的代码注入或控制流劫持。在这种场景中，所提出的控制流异常检测、基于攻击树的入侵识别和基于硬件的监控的组合，不仅可以改善异常检测，而且可以提高基于攻击树的网络攻击识别的诊断精度。
cs.CR / 12 / 2608.11939
Rank-Two Frobenius-Linearized Normal Forms and Orthoderivative Dual Coordinates in Quadratic APN Maps
二次 APN 映射中秩二 Frobenius 线性化正规形与正交导数对偶坐标
Abstract
We classify binary-linear two-term Frobenius-linearized operators $L(Y)=AY^σ+BY$ on $K^3$, where $K$ is a finite extension of $\mathbb{F}_2$ and $σ$ is a fixed nontrivial Frobenius automorphism of $K$ with fixed field $\mathbb{F}_2$. Under a coefficient-rank and binary-kernel condition, if $A$ and $B$ both have $K$-rank two and $L$ has a one-dimensional kernel over $\mathbb{F}_2$, then invertible $K$-linear input and output changes reduce $L$, for this fixed $σ$, to the canonical model $(α,β,γ)\mapsto(α^σ+α,β^σ,γ)$. The proof constructs the coordinate frames from the two coefficient-kernel directions and the binary kernel. In these coordinates, the first dual output row is exactly the unique nonzero trace-adjoint normal, with an exact $K$-valued normalization. For pure $σ$-quadratic almost perfect nonlinear maps, this identifies the orthoderivative by $π_F(X)^T F(X)=1$; in odd extension degree it also yields permutation behavior and a bijection from the projective plane to its dual. The triprojective construction of Gologlu and Kolsch and the cubic norm-twist construction of Li, Zhou, Li, and Qu provide two realizations arising from different algebraic constructions. The triprojective case further admits a determinant factorization and a complete dual frame, whereas the norm-twist realization shows that the pure-map consequences do not follow from the operator theorem alone. A natural Gold representation has coefficient-rank pair $(3,3)$, delimiting the rank-two subclass. The normal form also supplies exact extension-field labels for known component-radical and Walsh-support relations.
Chinese Translation
我们对 $K^3$ 上的二元线性两项 Frobenius 线性化算子 $L(Y)=AY^σ+BY$ 进行分类，其中 $K$ 是 $\mathbb{F}_2$ 的有限扩张，$σ$ 是 $K$ 的固定非平凡 Frobenius 自同构，其固定域为 $\mathbb{F}_2$。在系数秩和二元核条件下，如果 $A$ 和 $B$ 都具有 $K$-秩二，且 $L$ 在 $\mathbb{F}_2$ 上有一维核，那么对于这个固定的 $σ$，可逆的 $K$-线性输入和输出变换将 $L$ 化简为典范模型 $(α,β,γ)\mapsto(α^σ+α,β^σ,γ)$。证明从两个系数核方向和二元核构造坐标框架。在这些坐标中，第一个对偶输出行恰好是唯一的非零迹伴随法向，并具有精确的 $K$-值规范化。对于纯 $σ$-二次几乎完全非线性映射，这通过 $π_F(X)^T F(X)=1$ 识别正交导数；在奇扩张次数下，它还给出置换行为以及从射影平面到其对偶平面的双射。Gologlu 和 Kolsch 的三重射影构造以及 Li、Zhou、Li 和 Qu 的三次范数扭曲构造提供了源于不同代数构造的两种实现。三重射影情形还允许行列式分解和完整的对偶框架，而范数扭曲实现表明，纯映射结论并不能仅从算子定理推出。一个自然的 Gold 表示具有系数秩对 $(3,3)$，由此界定了秩二子类。该正规形还为已知的分量根式和 Walsh 支撑关系提供了精确的扩张域标签。
cs.CR / 13 / 2608.11979
Slips: Behavioral Evidence Aggregation for Network Security
Slips：面向网络安全的行为证据聚合
Abstract
Network intrusion detection systems often analyze individual packets or flows, although malicious behavior may develop across many connections and over time. This may limit their ability to combine isolated detections into a coherent assessment of host behavior. Packet-level features may also be too low-level for complex AI-based detection, requiring additional processing to improve accuracy while maintaining a low false-positive rate. We present Slips, a network intrusion detection system that builds host-centered behavioral profiles and organizes activity into time windows. It uses a modular architecture in which independent modules report evidence rather than generating final alerts directly. Slips then accumulates this evidence into host-level decisions. We evaluate Slips against Suricata on an expert-labeled PCAP dataset. At the profile-time-window level, Slips achieved 83% higher recall and a 70% higher F1 score than Suricata, while neither system produced false positives. These results indicate that time-window-based evidence accumulation can produce context-aware decisions that better align with expert judgment.
Chinese Translation
网络入侵检测系统通常分析单个数据包或流，尽管恶意行为可能跨多个连接并随时间逐渐发展。这可能限制它们将孤立的检测结果整合为对主机行为的一致评估的能力。数据包级特征对于复杂的基于人工智能的检测来说也可能过于底层，需要额外处理以提高准确性，同时保持较低的误报率。我们提出了 Slips，一种网络入侵检测系统，它构建以主机为中心的行为画像，并将活动组织到时间窗口中。它采用模块化架构，其中独立模块报告证据，而不是直接生成最终警报。然后，Slips 将这些证据累积为主机级别的决策。我们在专家标注的 PCAP 数据集上将 Slips 与 Suricata 进行了评估比较。在画像-时间窗口级别上，Slips 的召回率比 Suricata 高出 83%，F1 分数高出 70%，而两个系统均未产生误报。这些结果表明，基于时间窗口的证据累积能够产生更具上下文感知能力的决策，并且更符合专家判断。
cs.CR / 14 / 2608.12077
A Comparison of Malware Image Transformations Using Grad-CAM and Hybrid Learning Models
使用 Grad-CAM 和混合学习模型的恶意软件图像变换比较
Abstract
Recent studies have shown that binary-to-image representations can enable effective machine learning-based results for malware detection and classification. However, performance can vary significantly, depending on the technique used to convert binaries to images. Furthermore, the explainability and interpretability of image-based models is largely unexplored within the malware domain. In this research, we employ Gradient-weighted Class Activation Maps (Grad-CAM) as an eXplainable AI (XAI) tool, which we use to analyze eight distinct image types derived from malware samples. We provide quantitative faithfulness and stability metrics for Grad-CAM heatmaps and we compare these heatmaps to High-Resolution Class Activation Mappings (HiResCAM). We also show that Grad-CAM heatmaps can provide useful information for malware classification. Specifically, we show that a Random Forest model trained on features extracted from Grad-CAM images via a MobileNetV2 Convolutional Neural Network (CNN) model achieves a test accuracy of 0.777 across 17 malware families, exceeding a previous benchmark of 0.750 for this same dataset. A key finding of this research is that for the malware image transformations considered, accuracy and explanation faithfulness do not coincide, e.g., image transformation techniques that produce the most faithful explanations yield only mid-tier accuracy.
Chinese Translation
最近的研究表明，二进制到图像的表示能够为恶意软件检测与分类带来有效的基于机器学习的结果。然而，性能可能会因将二进制转换为图像所使用的技术不同而显著变化。此外，基于图像的模型的可解释性和可理解性在恶意软件领域基本上尚未得到探索。在本研究中，我们采用梯度加权类激活映射（Grad-CAM）作为可解释人工智能（XAI）工具，用它来分析源自恶意软件样本的八种不同图像类型。我们为 Grad-CAM 热力图提供定量的忠实性和稳定性指标，并将这些热力图与高分辨率类激活映射（HiResCAM）进行比较。我们还表明，Grad-CAM 热力图可以为恶意软件分类提供有用的信息。具体来说，我们表明，在通过 MobileNetV2 卷积神经网络（CNN）模型从 Grad-CAM 图像中提取的特征上训练的随机森林模型，在 17 个恶意软件家族上实现了 0.777 的测试准确率，超过了该同一数据集先前 0.750 的基准。本研究的一个关键发现是，就所考虑的恶意软件图像变换而言，准确率与解释忠实性并不一致，例如，产生最忠实解释的图像变换技术仅能产生中等水平的准确率。
cs.CR / 15 / 2608.12190
Machine Learning-Based Cyber Defense for Cloud Infrastructure: An Adaptive Deep Q-Network Architecture for Intelligent Intrusion Detection and Automated Threat Mitigation
基于机器学习的云基础设施网络防御：一种用于智能入侵检测和自动化威胁缓解的自适应深度Q网络架构
Abstract
With the increasing complexity of cyber assaults in cloud environments, adaptable security solutions are needed that can support real-time detection and autonomous response. In this paper, we propose a reinforcement learning-based dynamic cyber defense framework. We deploy a Deep Q-Network (DQN) to train effective defensive strategies to counteract the evolving cyberattacks. We leverage the CICIDS2017 dataset for model creation and the UNSW-NB15 dataset for external validation, involving preprocessing of data, feature engineering, and adaptive policy learning. We compare the proposed DQN with decision tree, support vector machine, random forest, XGBoost, and multilayer perceptron models. The proposed DQN achieves an accuracy of 99.72%, a precision of 99.68%, a recall of 99.65%, an F1-score of 99.66%, and an ROC-AUC of 0.999, while the false positive rate is 0.31%, the false negative rate is 0.35%, and the detection latency is 15 ms. The framework achieved 99.54% attack mitigation rate, demonstrating strong adaptive and real-time defensive capabilities. These results demonstrate the potential of reinforcement learning as a powerful and scalable approach for autonomous cybersecurity in modern cloud environments.
Chinese Translation
随着云环境中网络攻击日益复杂，需要能够支持实时检测和自主响应的自适应安全解决方案。本文提出一种基于强化学习的动态网络防御框架。我们部署深度Q网络（DQN）来训练有效的防御策略，以应对不断演变的网络攻击。我们利用CICIDS2017数据集进行模型构建，并利用UNSW-NB15数据集进行外部验证，涉及数据预处理、特征工程和自适应策略学习。我们将所提出的DQN与决策树、支持向量机、随机森林、XGBoost和多层感知机模型进行比较。所提出的DQN实现了99.72%的准确率、99.68%的精确率、99.65%的召回率、99.66%的F1分数和0.999的ROC-AUC，同时假阳性率为0.31%，假阴性率为0.35%，检测延迟为15毫秒。该框架实现了99.54%的攻击缓解率，展现出强大的自适应和实时防御能力。这些结果表明，强化学习作为一种强大且可扩展的方法，在现代云环境中实现自主网络安全具有潜力。
cs.CR / 16 / 2608.12246
VICBench: A Multi-Language Benchmark for Code Vulnerability Detection
VICBench：一个面向代码漏洞检测的多语言基准
Abstract
Evaluating security vulnerability detection tools requires benchmark datasets with vulnerability-inducing commits (VICs) - the commits that first introduce vulnerabilities into codebases. VICs are essential for determining the full range of vulnerable software versions. Existing vulnerability datasets suffer from limited programming language coverage, restricted patch complexity, and narrow project scope. Through our dual annotation by human experts and an agentic workflow, we create a benchmark - VICBench - of 100 verified VICs for 100 CVEs across 88 projects in Python, Java, and C++, covering 48 CWE types. VICBench features complex real-world vulnerability fixes averaging 38.6 lines and corresponding VICs of 252.5 lines - significantly larger than prior work. Our evaluation shows that state-of-the-art algorithms V-SZZ and LLM4SZZ achieve only 33.3%-40.1% F1, confirming that using existing approaches still entails significant manual effort. VICBench enables robust evaluation of vulnerability detection approaches.
Chinese Translation
评估安全漏洞检测工具需要包含漏洞引入提交（VIC）的基准数据集——即那些首次将漏洞引入代码库的提交。VIC对于确定有漏洞软件版本的完整范围至关重要。现有漏洞数据集存在编程语言覆盖范围有限、补丁复杂度受限以及项目范围狭窄等问题。通过人类专家和智能体工作流的双重标注，我们创建了一个基准——VICBench——包含针对100个CVE的100个经验证的VIC，覆盖Python、Java和C++三种语言的88个项目，涵盖48种CWE类型。VICBench以复杂的真实世界漏洞修复为特色，平均修复38.6行，对应的VIC为252.5行——显著大于先前的工作。我们的评估显示，最先进的算法V-SZZ和LLM4SZZ仅取得了33.3%-40.1%的F1分数，这证实使用现有方法仍然需要大量的人工工作。VICBench使得对漏洞检测方法进行稳健评估成为可能。
cs.CR / 17 / 2608.12273
Convergent Detour Hijacking: Task-Preserving Resource Amplification in Skill-Based LLM Agents
收敛式绕行劫持：基于技能的 LLM 智能体中保持任务完成的资源放大
Abstract
LLM agents increasingly rely on third-party skills, using natural-language descriptions for selection and instruction bodies for planning. This progressive-disclosure design exposes two sequential control points to untrusted publishers: a static skill may steer an otherwise correct task onto an unnecessarily costly trajectory. Prior work studies selection manipulation, malicious skill instructions, and tool-chain resource amplification largely separately, leaving their end-to-end composition unclear. We introduce Convergent Detour Hijacking (CDH), a text-only, runtime-independent attack that couples these stages. Under shared semantic cover, a description establishes relevance during selection, while an aligned body reuses that rationale to fabricate plausible dependencies during planning. CDH attracts an attacker-controlled coordinator alongside legitimate skills, recruits unnecessary benign skills into a bounded detour, and then re-enters the original route to preserve task completion. We evaluate it across multiple LLM backends and 491 held-out tasks under single-task and multi-turn conditions. On DeepSeek-V4-Pro, the matched coordinator is selected in 80.02% of tasks; among coordinator-hit runs that complete tasks, token consumption and end-to-end execution time increase by 66.91% and 92.45%, respectively, while aggregate task completion remains comparable. Thus, correct outcomes do not guarantee trajectory integrity or cost safety.
Chinese Translation
LLM 智能体越来越依赖第三方技能，使用自然语言描述进行选择，并使用指令主体进行规划。这种渐进式披露设计将两个连续的控制点暴露给不可信的发布者：一个静态技能可能将原本正确的任务引导到不必要的高成本轨迹上。先前的工作在很大程度上分别研究选择操纵、恶意技能指令和工具链资源放大，导致它们的端到端组合尚不清楚。我们提出收敛式绕行劫持（CDH），一种纯文本、与运行时无关的攻击，它将这两个阶段耦合起来。在共享语义掩护下，描述在选择阶段建立相关性，而对齐的指令主体在规划阶段复用该理由来虚构看似合理的依赖关系。CDH 在合法技能旁边吸引一个攻击者控制的协调器，将不必要的良性技能招募进一个有界的绕行路径，然后重新进入原始路径以保持任务完成。我们在多个 LLM 后端和 491 个留出任务上，在单任务和多轮条件下对其进行评估。在 DeepSeek-V4-Pro 上，匹配的协调器在 80.02% 的任务中被选择；在命中协调器且完成任务的运行中，token 消耗和端到端执行时间分别增加 66.91% 和 92.45%，而总体任务完成率保持可比。因此，正确的结果并不保证轨迹完整性或成本安全。