|
创作新主题 |
社区所有版块导航
DATA
| docker Elasticsearch |
WEB开发
| linux MongoDB Redis DATABASE NGINX 其他Web框架 web工具 zookeeper tornado NoSql Bootstrap js peewee Git bottle IE MQ Jquery |
机器学习
| 机器学习算法 |
产品
| 短视频 |
印度
| 印度 |
一周十大热门主题
点击阅读原文访问arxivdaily.com，涵盖CS|物理|数学|经济|统计|金融|生物|电气领域，更有搜索、收藏等功能！
cs.LG 方向，今日共计173篇
大模型相关(34篇)
【1】SortedRL: Accelerating RL Training for LLMs through Online Length-Aware Scheduling
标题：SortedRL：通过在线长度感知调度加速LLM的RL训练
链接：https://arxiv.org/abs/2603.23414
摘要：缩放强化学习（RL）在增强大型语言模型（LLM）的推理能力方面表现出了很强的前景，特别是在需要长思维链生成的任务中。然而，RL训练效率通常受到推出阶段的影响，当生成长轨迹时，该阶段可能占总训练时间的70%（例如，16 k令牌），这是由于缓慢的自回归生成以及推出和策略更新之间的同步开销。我们提出了SortedRL，一种在线长度感知调度策略，旨在通过提高推出效率和保持训练稳定性来解决这一瓶颈。SortedRL根据输出长度重新排序卷展栏样本，优先考虑形成早期更新组的短样本。这使得大批量推出，灵活的更新批次，并同时接近政策微课程建设。为了进一步加速流水线，SortedRL采用了一种机制，通过基于缓存的机制来控制非策略训练的程度，并由专用的RL基础设施支持，该基础设施通过有状态控制器和rollout buffer来管理rollout和更新。使用LLaMA-3.1-8B和Qwen-2.5- 32 B进行各种任务的实验，包括逻辑谜题，以及AIME 24，Math 500和Minerval等数学挑战，表明SortedRL将RL训练泡沫比率降低了50%以上，同时在相同数据量的情况下，性能比基线高出3.9%至18.4%。
摘要：Scaling reinforcement learning (RL) has shown strong promise for enhancing the reasoning abilities of large language models (LLMs), particularly in tasks requiring long chain-of-thought generation. However, RL training efficiency is often bottlenecked by the rollout phase, which can account for up to 70% of total training time when generating long trajectories (e.g., 16k tokens), due to slow autoregressive generation and synchronization overhead between rollout and policy updates. We propose SortedRL, an online length-aware scheduling strategy designed to address this bottleneck by improving rollout efficiency and maintaining training stability. SortedRL reorders rollout samples based on output lengths, prioritizing short samples forming groups for early updates. This enables large rollout batches, flexible update batches, and near on-policy micro-curriculum construction simultaneously. To further accelerate the pipeline, SortedRL incorporates a mechanism to control the degree of off-policy training through a cache-based mechanism, and is supported by a dedicated RL infrastructure that manages rollout and update via a stateful controller and rollout buffer. Experiments using LLaMA-3.1-8B and Qwen-2.5-32B on diverse tasks, including logical puzzles, and math challenges like AIME 24, Math 500, and Minerval, show that SortedRL reduces RL training bubble ratios by over 50%, while attaining 3.9% to 18.4% superior performance over baseline given same amount of data.
【2】Off-Policy Value-Based Reinforcement Learning for Large Language Models
标题：大型语言模型的非策略基于价值的强化学习
链接：https://arxiv.org/abs/2603.23355
摘要：提高数据利用效率对于扩展强化学习（RL）用于生成轨迹的长期任务至关重要。然而，LLM的主要RL方法在很大程度上是基于策略的：它们只更新一次每批数据，丢弃它，然后收集新的样本，导致样本效率低下。在这项工作中，我们探索了一种替代的基于值的RL框架，自然地实现了离线学习。我们提出了ReVal，贝尔曼更新为基础的方法，结合逐步信号捕获的内部一致性与来自结果验证的可靠性水平的信号。ReVal自然支持基于回放缓冲区的训练，允许有效地重用过去的轨迹。在标准数学推理基准测试上的实验表明，ReVal不仅收敛速度更快，而且在最终性能上优于GRPO。在DeepSeek-R1-Distill-1.5B上，ReVal提高了训练效率，并在AIME 24中实现了2.7%的改进，在域外基准GPQA中比GRPO提高了4.5%。这些结果表明，基于价值的强化学习是一个实际的替代基于政策的方法LLM培训。
摘要：Improving data utilization efficiency is critical for scaling reinforcement learning (RL) for long-horizon tasks where generating trajectories is expensive. However, the dominant RL methods for LLMs are largely on-policy: they update each batch of data only once, discard it, and then collect fresh samples, resulting in poor sample efficiency. In this work, we explore an alternative value-based RL framework for LLMs that naturally enables off-policy learning. We propose ReVal, a Bellman-update-based method that combines stepwise signals capturing internal consistency with trajectory-level signals derived from outcome verification. ReVal naturally supports replay-buffer-based training, allowing efficient reuse of past trajectories. Experiments on standard mathematical reasoning benchmarks show that ReVal not only converges faster but also outperforms GRPO in final performance. On DeepSeek-R1-Distill-1.5B, ReVal improves training efficiency and achieves improvement of 2.7% in AIME24 and 4.5% in out-of-domain benchmark GPQA over GRPO. These results suggest that value-based RL is a practical alternative to policy-based methods for LLM training.
【3】Not All Tokens Are Created Equal: Query-Efficient Jailbreak Fuzzing for LLMs
标题：并非所有代币都是平等的：LLM的查询高效越狱模糊
链接：https://arxiv.org/abs/2603.23269
摘要：大型语言模型（LLM）被广泛部署，但容易受到越狱提示的影响，导致违反策略的输出。尽管之前的研究已经发现了这些风险，但他们通常在快速突变过程中将所有标记视为同等重要，忽略了单个标记对触发模型拒绝的不同贡献。因此，这些攻击在查询受限的情况下引入了大量冗余搜索，降低了攻击效率并阻碍了全面的漏洞评估。在这项工作中，我们对拒绝行为进行了令牌级分析，并观察到令牌贡献是高度倾斜的，而不是均匀的。此外，我们发现强大的跨模型的一致性拒绝的倾向，使使用代理模型来估计令牌级的贡献目标模型的拒绝。受这些发现的启发，我们提出了TriageFuzz，这是一个令牌感知的越狱模糊框架，它通过一系列定制设计来适应模糊测试方法。TriageFuzz利用代理模型来估计单个令牌对拒绝行为的贡献，从而能够识别提示中的敏感区域。此外，它采用了一个参考引导的进化策略，自适应权重候选提示与轻量级评分器转向绕过安全约束的演变。在六个开源LLM和三个商业API上进行的广泛实验表明，TriageFuzz实现了相当的攻击成功率（ASR），并显着降低了查询成本。值得注意的是，与基线相比，它实现了90%的ASR，查询减少了70%以上。即使在25个查询的极其有限的预算下，TriageFuzz也优于现有方法，将ASR提高了20- 40%。
摘要 ：Large Language Models(LLMs) are widely deployed, yet are vulnerable to jailbreak prompts that elicit policy-violating outputs. Although prior studies have uncovered these risks, they typically treat all tokens as equally important during prompt mutation, overlooking the varying contributions of individual tokens to triggering model refusals. Consequently, these attacks introduce substantial redundant searching under query-constrained scenarios, reducing attack efficiency and hindering comprehensive vulnerability assessment. In this work, we conduct a token-level analysis of refusal behavior and observe that token contributions are highly skewed rather than uniform. Moreover, we find strong cross-model consistency in refusal tendencies, enabling the use of a surrogate model to estimate token-level contributions to the target model's refusals. Motivated by these findings, we propose TriageFuzz, a token-aware jailbreak fuzzing framework that adapts the fuzz testing approach with a series of customized designs. TriageFuzz leverages a surrogate model to estimate the contribution of individual tokens to refusal behaviors, enabling the identification of sensitive regions within the prompt. Furthermore, it incorporates a refusal-guided evolutionary strategy that adaptively weights candidate prompts with a lightweight scorer to steer the evolution toward bypassing safety constraints. Extensive experiments on six open-source LLMs and three commercial APIs demonstrate that TriageFuzz achieves comparable attack success rates (ASR) with significantly reduced query costs. Notably, it attains a 90% ASR with over 70% fewer queries compared to baselines. Even under an extremely restrictive budget of 25 queries, TriageFuzz outperforms existing methods, improving ASR by 20-40%.
【4】SafeSeek: Universal Attribution of Safety Circuits in Language Models
标题：SafeSeek：语言模型中安全电路的普遍属性
链接：https://arxiv.org/abs/2603.23268
摘要：机械可解释性表明，安全关键行为（例如，大型语言模型（LLM）中的对齐、越狱、后门（backdoor）都基于专门的功能组件。然而，现有的安全归因方法由于依赖于启发式的、特定领域的度量和搜索算法而难以推广和可靠。为了解决这个问题，我们提出了\ourmethod，一个统一的安全可解释性框架，确定功能完整的安全电路LLM通过优化。与专注于孤立头或神经元的方法不同，我们的方法引入了可微二进制掩码，通过安全数据集上的梯度下降来提取多粒度电路，同时集成了安全电路调谐，以利用这些稀疏电路进行有效的安全微调。我们在两个关键的场景中验证了我们的方法：（1）后门攻击，识别一个稀疏度为0.42%的后门电路，其消融将攻击成功率（ASR）从100%降低到0.4%，同时保持99%以上的通用性;\textbf{（2）安全对齐}，定位具有3.03\%头部和0.79\%神经元的对齐电路，其移除使ASR从0.8\% $\增加到96.9\%，而在有益性微调期间排除该电路保持96.5%的安全保持率。
摘要：Mechanistic interpretability reveals that safety-critical behaviors (e.g., alignment, jailbreak, backdoor) in Large Language Models (LLMs) are grounded in specialized functional components. However, existing safety attribution methods struggle with generalization and reliability due to their reliance on heuristic, domain-specific metrics and search algorithms. To address this, we propose \ourmethod, a unified safety interpretability framework that identifies functionally complete safety circuits in LLMs via optimization. Unlike methods focusing on isolated heads or neurons, \ourmethod introduces differentiable binary masks to extract multi-granular circuits through gradient descent on safety datasets, while integrates Safety Circuit Tuning to utilize these sparse circuits for efficient safety fine-tuning. We validate \ourmethod in two key scenarios in LLM safety: \textbf{(1) backdoor attacks}, identifying a backdoor circuit with 0.42\% sparsity, whose ablation eradicates the Attack Success Rate (ASR) from 100\% $\to$ 0.4\% while retaining over 99\% general utility; \textbf{(2) safety alignment}, localizing an alignment circuit with 3.03\% heads and 0.79\% neurons, whose removal spikes ASR from 0.8\% $\to$ 96.9\%, whereas excluding this circuit during helpfulness fine-tuning maintains 96.5\% safety retention.
【5】Is AI Catching Up to Human Expression? Exploring Emotion, Personality, Authorship, and Linguistic Style in English and Arabic with Six Large Language Models
标题：人工智能正在追赶人类表情吗？用六种大型语言模型探索英语和阿拉伯语的情感、性格、作者和语言风格
链接：https://arxiv.org/abs/2603.23251
备注：Preprint. Under review
摘要：LLM的进步流畅性提出了关于他们在不同语言和文化背景下模仿复杂人类特征（包括情感表达和个性）的能力的重要问题。本研究调查了LLM是否可以令人信服地模仿英语中的情感细微差别和阿拉伯语中的个性标记，这是一种具有独特语言和文化特征的关键资源不足的语言。我们在六个模型上执行两个任务：Jais，Mistral，LLaMA，GPT-4 o，Gemini和DeepSeek。首先，我们评估机器分类器是否能够可靠地区分人类创作的文本和人工智能生成的文本。其次，我们评估LLM生成的文本表现出与人类相似的情感或人格特征的程度。我们的研究结果表明，人工智能生成的文本与人类创作的文本是可区分的（F1>0.95），尽管分类性能在释义样本上有所下降，表明对表面风格线索的依赖。情感和个性分类实验揭示了显著的泛化差距：在人类数据上训练的分类器在人工智能生成的文本上表现不佳，反之亦然，这表明LLM对情感信号的编码与人类不同。重要的是，用人工智能生成的数据增强训练可以提高阿拉伯人性格分类任务的性能，突出了合成数据在解决资源不足语言挑战方面的潜力。模型特定的分析表明，GPT-4 o和双子座表现出优越的情感连贯性。语言学和心理语言学分析揭示了人类和人工智能文本之间在语气，真实性和文本复杂性方面的可测量差异。这些发现对情感计算、作者归属和负责任的人工智能部署具有影响，特别是在资源不足的语言环境中，生成式人工智能检测和对齐带来了独特的挑战。
摘要：The advancing fluency of LLMs raises important questions about their ability to emulate complex human traits, including emotional expression and personality, across diverse linguistic and cultural contexts. This study investigates whether LLMs can convincingly mimic emotional nuance in English and personality markers in Arabic, a critical under-resourced language with unique linguistic and cultural characteristics. We conduct two tasks across six models:Jais, Mistral, LLaMA, GPT-4o, Gemini, and DeepSeek. First, we evaluate whether machine classifiers can reliably distinguish between human-authored and AI-generated texts. Second, we assess the extent to which LLM-generated texts exhibit emotional or personality traits comparable to those of humans. Our results demonstrate that AI-generated texts are distinguishable from human-authored ones (F1>0.95), though classification performance deteriorates on paraphrased samples, indicating a reliance on superficial stylistic cues. Emotion and personality classification experiments reveal significant generalization gaps: classifiers trained on human data perform poorly on AI-generated texts and vice versa, suggesting LLMs encode affective signals differently from humans. Importantly, augmenting training with AI-generated data enhances performance in the Arabic personality classification task, highlighting the potential of synthetic data to address challenges in under-resourced languages. Model-specific analyses show that GPT-4o and Gemini exhibit superior affective coherence. Linguistic and psycholinguistic analyses reveal measurable divergences in tone, authenticity, and textual complexity between human and AI texts. These findings have implications for affective computing, authorship attribution, and responsible AI deployment, particularly within underresourced language contexts where generative AI detection and alignment pose unique challenges.
【6】Decoding AI Authorship: Can LLMs Truly Mimic Human Style Across Literature and Politics?
标题：解码人工智能作者：法学硕士能否真正模仿文学和政治中的人类风格？
链接：https://arxiv.org/abs/2603.23219
备注：Preprint. Accepted for publication in Digital Scholarship in the Humanities (OUP)
摘要：在生成人工智能模仿特定人类风格的能力不断提高的情况下，本研究调查了最先进的大型语言模型（LLM）的能力，包括GPT-4 o，Gemini 1.5 Pro和Claude Sonnet 3.5，以模仿着名文学和政治人物的作者签名：沃尔特·惠特曼，威廉·华兹华斯，唐纳德·特朗普和巴拉克·奥巴马。利用具有严格主题对齐的zero-shot提示框架，我们生成了通过基于变换的分类（BERT）和可解释机器学习（XGBoost）相结合的互补框架评估的合成语料库。我们的方法集成了语言查询和单词计数（LIWC）标记，困惑和可读性指数，以评估AI生成的文本和人类创作的文本之间的差异。结果表明，人工智能生成的模仿仍然具有高度可检测性，在一组有限的八个风格特征上训练的XGBoost模型实现了与高维神经分类器相当的准确性。特征重要性分析将困惑度确定为主要的判别指标，揭示了与人类书写的较高可变性相比，人工智能输出的随机规律性存在显着差异。虽然LLM在低维启发式特征（如句法复杂性和可读性）上表现出与人类作者的分布收敛，但它们尚未完全复制人类作者语料库中固有的细微差别的情感密度和风格差异。通过隔离当前生成模仿中的特定统计差距，这项研究为LLM文体行为提供了一个全面的基准，并为数字人文和社交媒体中的作者归属提供了重要见解。
摘要 ：Amidst the rising capabilities of generative AI to mimic specific human styles, this study investigates the ability of state-of-the-art large language models (LLMs), including GPT-4o, Gemini 1.5 Pro, and Claude Sonnet 3.5, to emulate the authorial signatures of prominent literary and political figures: Walt Whitman, William Wordsworth, Donald Trump, and Barack Obama. Utilizing a zero-shot prompting framework with strict thematic alignment, we generated synthetic corpora evaluated through a complementary framework combining transformer-based classification (BERT) and interpretable machine learning (XGBoost). Our methodology integrates Linguistic Inquiry and Word Count (LIWC) markers, perplexity, and readability indices to assess the divergence between AI-generated and human-authored text. Results demonstrate that AI-generated mimicry remains highly detectable, with XGBoost models trained on a restricted set of eight stylometric features achieving accuracy comparable to high-dimensional neural classifiers. Feature importance analyses identify perplexity as the primary discriminative metric, revealing a significant divergence in the stochastic regularity of AI outputs compared to the higher variability of human writing. While LLMs exhibit distributional convergence with human authors on low-dimensional heuristic features, such as syntactic complexity and readability, they do not yet fully replicate the nuanced affective density and stylistic variance inherent in the human-authored corpus. By isolating the specific statistical gaps in current generative mimicry, this study provides a comprehensive benchmark for LLM stylistic behavior and offers critical insights for authorship attribution in the digital humanities and social media.
【7】Sparser, Faster, Lighter Transformer Language Models
标题：更精简、更快、更轻的Transformer语言模型
链接：https://arxiv.org/abs/2603.23198
备注：Code and checkpoints available at: https://github.com/SakanaAI/sparser-faster-llms
摘要：缩放自回归大型语言模型（LLM）已经推动了前所未有的进步，但也带来了巨大的计算成本。在这项工作中，我们通过利用LLM前馈层中的非结构化稀疏性来解决这些成本，这些组件占了大部分模型参数和执行FLOP。为了实现这一目标，我们引入了一种新的稀疏打包格式和一组CUDA内核，旨在与现代GPU的优化执行管道无缝集成，从而在LLM推理和训练期间实现高效的稀疏计算。为了证实我们的收益，我们提供了LLM稀疏性的定量研究，证明简单的L1正则化可以诱导超过99%的稀疏性，对下游性能的影响可以忽略不计。当与我们的内核配对时，我们表明这些稀疏水平转化为大量的吞吐量，能源效率和内存使用的好处，随着模型规模的增加。我们将在开源许可证下发布所有代码和内核，以促进采用并加速研究，将稀疏性作为提高现代基础模型效率和可扩展性的实用轴。
摘要：Scaling autoregressive large language models (LLMs) has driven unprecedented progress but comes with vast computational costs. In this work, we tackle these costs by leveraging unstructured sparsity within an LLM's feedforward layers, the components accounting for most of the model parameters and execution FLOPs. To achieve this, we introduce a new sparse packing format and a set of CUDA kernels designed to seamlessly integrate with the optimized execution pipelines of modern GPUs, enabling efficient sparse computation during LLM inference and training. To substantiate our gains, we provide a quantitative study of LLM sparsity, demonstrating that simple L1 regularization can induce over 99% sparsity with negligible impact on downstream performance. When paired with our kernels, we show that these sparsity levels translate into substantial throughput, energy efficiency, and memory usage benefits that increase with model scale. We will release all code and kernels under an open-source license to promote adoption and accelerate research toward establishing sparsity as a practical axis for improving the efficiency and scalability of modern foundation models.
【8】Robust Safety Monitoring of Language Models via Activation Watermarking
标题：通过激活水印对语言模型进行稳健安全监控
链接：https://arxiv.org/abs/2603.23171
备注：20 pages, 17 figures
摘要：大型语言模型（LLM）可能被滥用来泄露敏感信息，例如武器制造指令或编写恶意软件。LLM提供程序依赖于$\employment {monitoring}$来检测和标记推理过程中的不安全行为。一个开放的安全挑战是$\N {adaptive}$对手，他们设计的攻击同时（i）逃避检测，同时（ii）引发不安全的行为。适应性攻击者是一个主要问题，因为LLM提供商无法修补其安全机制，因为他们不知道他们的模型是如何被滥用的。我们将LLM监控作为一个安全游戏，知道监控器的对手试图提取敏感信息，而提供商必须以低误报率准确检测这些对抗性查询。我们的工作（一）表明，现有的LLM监视器容易受到自适应攻击者和（二）设计改进的防御，通过$\{激活水印}$仔细引入不确定性的攻击者在推理过程中。我们发现，$\ldblquote {激活水印}$优于警卫基线高达$52\%$下的自适应攻击者谁知道的监测算法，但没有秘密密钥。
摘要：Large language models (LLMs) can be misused to reveal sensitive information, such as weapon-making instructions or writing malware. LLM providers rely on $\emph{monitoring}$ to detect and flag unsafe behavior during inference. An open security challenge is $\emph{adaptive}$ adversaries who craft attacks that simultaneously (i) evade detection while (ii) eliciting unsafe behavior. Adaptive attackers are a major concern as LLM providers cannot patch their security mechanisms, since they are unaware of how their models are being misused. We cast $\emph{robust}$ LLM monitoring as a security game, where adversaries who know about the monitor try to extract sensitive information, while a provider must accurately detect these adversarial queries at low false positive rates. Our work (i) shows that existing LLM monitors are vulnerable to adaptive attackers and (ii) designs improved defenses through $\emph{activation watermarking}$ by carefully introducing uncertainty for the attacker during inference. We find that $\emph{activation watermarking}$ outperforms guard baselines by up to $52\%$ under adaptive attackers who know the monitoring algorithm but not the secret key.
【9】DAK-UCB: Diversity-Aware Prompt Routing for LLMs and Generative Models
标题：DAK-UCB：LLM和生成模型的多样性感知提示路由
链接：https://arxiv.org/abs/2603.23140
备注：Accepted at ICLR 2026
摘要：生成式AI和LLM服务的扩展强调了对自适应机制的日益增长的需求，以选择适当的可用模型来响应用户的提示。最近的工作已经提出了离线和在线学习公式，以识别输入提示的最佳生成AI模型，仅基于最大化基于模糊度的保真度评估分数，例如，CLIP-文本到图像生成中的得分。然而，这种基于多样性的选择方法忽略了所生成的输出的多样性，因此，它们可能无法解决所生成的响应中的潜在多样性缺陷。在本文中，我们介绍了多样性感知核置信上限（DAK-UCB）方法作为一个上下文强盗算法的在线选择生成模型的多样性考虑。建议DAK-UCB方法将保真度和多样性相关的指标纳入选择过程。我们设计了这个框架的基础上，感知的多样性得分函数，分解到一个基于两个样本的期望在前几代回合的输出对。具体来说，我们说明了我们的框架使用联合核距离和核熵措施的应用。我们的实验结果表明，DAK-UCB的有效性，促进多样性意识的模型选择，同时保持保真度的几代序列的提示。该代码可在https://github.com/Donya-Jafari/DAK-UCB上获得。
摘要：The expansion of generative AI and LLM services underscores the growing need for adaptive mechanisms to select an appropriate available model to respond to a user's prompts. Recent works have proposed offline and online learning formulations to identify the optimal generative AI model for an input prompt, based solely on maximizing prompt-based fidelity evaluation scores, e.g., CLIP-Score in text-to-image generation. However, such fidelity-based selection methods overlook the diversity of generated outputs, and hence, they can fail to address potential diversity shortcomings in the generated responses. In this paper, we introduce the Diversity-Aware Kernelized Upper Confidence Bound (DAK-UCB) method as a contextual bandit algorithm for the online selection of generative models with diversity considerations. The proposed DAK-UCB method incorporates both fidelity and diversity-related metrics into the selection process. We design this framework based on prompt-aware diversity score functions that decompose to a two-sample-based expectation over prompt-output pairs in the previous generation rounds. Specifically, we illustrate the application of our framework using joint kernel distance and kernel entropy measures. Our experimental results demonstrate the effectiveness of DAK-UCB in promoting diversity-aware model selection while maintaining fidelity in the generations for a sequence of prompts. The code is available at https://github.com/Donya-Jafari/DAK-UCB.
【10】Polaris: A Gödel Agent Framework for Small Language Models through Experience-Abstracted Policy Repair
标题：Polaris：通过经验抽象策略修复的小型语言模型的哥德尔代理框架
链接：https://arxiv.org/abs/2603.23129
摘要：Gödel agent实现了递归的自我改进：agent检查自己的策略并跟踪，然后在测试循环中修改该策略。我们介绍Polaris，这是一种用于紧凑模型的哥德尔代理，它通过经验抽象执行策略修复，通过分析、策略形成、抽象和最小代码补丁修复（带有保守检查）的结构化循环将失败转化为策略更新。与响应级别自校正或参数调优不同，Polaris通过小的、可审计的补丁进行策略级别更改，这些补丁在策略中持久存在，并在每个基准测试中的不可见实例上重用。作为循环的一部分，代理参与Meta推理：它解释自己的错误，对自己的策略提出具体的修改，然后更新策略。为了实现累积的策略细化，我们引入了经验抽象，它将失败提炼成紧凑的，可重用的策略，转移到看不见的实例。在MGSM、DROP、GPQA和LitBench（涵盖算术推理、合成推理、研究生水平问题解决和创意写作评估）上，配备Polaris的70亿参数模型在基本策略和竞争基线上实现了一致的收益。
摘要：Gödel agent realize recursive self-improvement: an agent inspects its own policy and traces and then modifies that policy in a tested loop. We introduce Polaris, a Gödel agent for compact models that performs policy repair via experience abstraction, turning failures into policy updates through a structured cycle of analysis, strategy formation, abstraction, and minimal code pat ch repair with conservative checks. Unlike response level self correction or parameter tuning, Polaris makes policy level changes with small, auditable patches that persist in the policy and are reused on unseen instances within each benchmark. As part of the loop, the agent engages in meta reasoning: it explains its errors, proposes concrete revisions to its own policy, and then updates the policy. To enable cumulative policy refinement, we introduce experience abstraction, which distills failures into compact, reusable strategies that transfer to unseen instances. On MGSM, DROP, GPQA, and LitBench (covering arithmetic reasoning, compositional inference, graduate-level problem solving, and creative writing evaluation), a 7-billion-parameter model equipped with Polaris achieves consistent gains over the base policy and competitive baselines.
【11】YOLOv10 with Kolmogorov-Arnold networks and vision-language foundation models for interpretable object detection and trustworthy multimodal AI in computer vision perception
标题：YOLOv 10具有Kolmogorov-Arnold网络和视觉语言基础模型，用于计算机视觉感知中的可解释对象检测和值得信赖的多模式人工智能
链接：https://arxiv.org/abs/2603.23037
备注：14 pages, 23 Figures, 6 Tables
摘要：一种新的Kolmogorov-Arnold网络框架的可解释的对象检测能力在这里进行检查。这种方法是指计算机视觉对自动驾驶汽车感知的一个关键限制。这些系统在视觉退化或模糊场景中提供关于其置信度分数的可靠性的有限透明度。为了解决这一限制，Kolmogorov-Arnold网络被用作可解释的事后代理，使用七个几何和语义特征来对You Only Look Once（Yolov 10）检测的可信度进行建模。Kolmogorov-Arnold网络的基于加性样条的结构可以直接可视化每个特征的影响。这产生了平滑和透明的函数映射，揭示了模型的置信度何时得到很好的支持，何时不可靠。在上下文中的公共对象（COCO）和巴斯大学校园的图像上的实验表明，该框架可以准确地识别模糊，遮挡或低纹理下的低信任预测。这为过滤、审查或下游风险缓解提供了可操作的见解。此外，自举语言图像（BLIP）基础模型生成每个场景的描述性字幕。该工具支持轻量级的多模式界面，而不会影响可解释性层。由此产生的系统提供可解释的对象检测与可信的置信度估计。它为自主和多模态人工智能应用提供了透明和实用的感知组件的强大工具。
摘要：The interpretable object detection capabilities of a novel Kolmogorov-Arnold network framework are examined here. The approach refers to a key limitation in computer vision for autonomous vehicles perception, and beyond. These systems offer limited transparency regarding the reliability of their confidence scores in visually degraded or ambiguous scenes. To address this limitation, a Kolmogorov-Arnold network is employed as an interpretable post-hoc surrogate to model the trustworthiness of the You Only Look Once (Yolov10) detections using seven geometric and semantic features. The additive spline-based structure of the Kolmogorov-Arnold network enables direct visualisation of each feature's influence. This produces smooth and transparent functional mappings that reveal when the model's confidence is well supported and when it is unreliable. Experiments on both Common Objects in Context (COCO), and images from the University of Bath campus demonstrate that the framework accurately identifies low-trust predictions under blur, occlusion, or low texture. This provides actionable insights for filtering, review, or downstream risk mitigation. Furthermore, a bootstrapped language-image (BLIP) foundation model generates descriptive captions of each scene. This tool enables a lightweight multimodal interface without affecting the interpretability layer. The resulting system delivers interpretable object detection with trustworthy confidence estimates. It offers a powerful tool for transparent and practical perception component for autonomous and multimodal artificial intelligence applications.
【12】Can Large Language Models Reason and Optimize Under Constraints?
标题：大型语言模型可以在约束下推理和优化吗？
链接：https://arxiv.org/abs/2603.23004
摘要：大型语言模型（LLM）在各种自然语言任务中表现出了强大的能力，但它们解决具有约束的抽象和优化问题的能力仍然很少被探索。在本文中，我们调查是否LLM可以推理和优化的最优潮流（OPF）问题的物理和操作约束下。我们介绍了一个具有挑战性的评估设置，需要一组基本技能，如推理，结构化的输入处理，算术和约束优化。我们的评估表明，SoTA LLM在大多数任务中失败，推理LLM在最复杂的设置中仍然失败。我们的研究结果突出了LLM在约束条件下处理结构化推理能力的关键差距，这项工作为开发更有能力的LLM助手提供了严格的测试环境，这些助手可以解决现实世界的电网优化问题。
摘要：Large Language Models (LLMs) have demonstrated great capabilities across diverse natural language tasks; yet their ability to solve abstraction and optimization problems with constraints remains scarcely explored. In this paper, we investigate whether LLMs can reason and optimize under the physical and operational constraints of Optimal Power Flow (OPF) problem. We introduce a challenging evaluation setup that requires a set of fundamental skills such as reasoning, structured input handling, arithmetic, and constrained optimization. Our evaluation reveals that SoTA LLMs fail in most of the tasks, and that reasoning LLMs still fail in the most complex settings. Our findings highlight critical gaps in LLMs' ability to handle structured reasoning under constraints, and this work provides a rigorous testing environment for developing more capable LLM assistants that can tackle real-world power grid optimization problems.
【13】TreeTeaming: Autonomous Red-Teaming of Vision-Language Models via Hierarchical Strategy Exploration
标题：TreeTeaming：通过分层策略探索实现视觉语言模型的自主红色团队化
链接：https://arxiv.org/abs/2603.22882
备注：CVPR2026
摘要：视觉语言模型（VLM）的快速发展使其安全漏洞成为焦点。然而，现有的红色组队方法从根本上受到固有的线性探索范式的约束，将它们限制在预定义的策略集内进行优化，并防止发现新颖的、多样化的漏洞。为了超越这一限制，我们引入了TreeTeaming，这是一个自动化的红色团队框架，它将策略探索从静态测试重新构建为动态的进化发现过程。它的核心是一个由大型语言模型（LLM）驱动的战略搜索器，它自主决定是否进化有希望的攻击路径或探索不同的战略分支，从而动态地构建和扩展战略树。然后，多模态执行器的任务是执行这些复杂的策略。在12个主要VLM的实验中，TreeTeaming在11个模型上实现了最先进的攻击成功率，优于现有方法，在GPT-4 o上达到了87.60%。该框架还展示了优于先前公开的越狱策略联盟的战略多样性。此外，生成的攻击表现出平均毒性降低23.09%，展示了它们的隐蔽性和微妙性。我们的工作为自动化漏洞发现引入了一种新的范式，强调了在静态安全性之外进行主动探索以保护前沿AI模型的必要性。
摘要 ：The rapid advancement of Vision-Language Models (VLMs) has brought their safety vulnerabilities into sharp focus. However, existing red teaming methods are fundamentally constrained by an inherent linear exploration paradigm, confining them to optimizing within a predefined strategy set and preventing the discovery of novel, diverse exploits. To transcend this limitation, we introduce TreeTeaming, an automated red teaming framework that reframes strategy exploration from static testing to a dynamic, evolutionary discovery process. At its core lies a strategic Orchestrator, powered by a Large Language Model (LLM), which autonomously decides whether to evolve promising attack paths or explore diverse strategic branches, thereby dynamically constructing and expanding a strategy tree. A multimodal actuator is then tasked with executing these complex strategies. In the experiments across 12 prominent VLMs, TreeTeaming achieves state-of-the-art attack success rates on 11 models, outperforming existing methods and reaching up to 87.60\% on GPT-4o. The framework also demonstrates superior strategic diversity over the union of previously public jailbreak strategies. Furthermore, the generated attacks exhibit an average toxicity reduction of 23.09\%, showcasing their stealth and subtlety. Our work introduces a new paradigm for automated vulnerability discovery, underscoring the necessity of proactive exploration beyond static heuristics to secure frontier AI models.
【14】When AI Shows Its Work, Is It Actually Working? Step-Level Evaluation Reveals Frontier Language Models Frequently Bypass Their Own Reasoning
标题：当人工智能展示它的作用时，它真的有效吗？分步评估揭示前沿语言模型经常绕过自己的推理
链接：https://arxiv.org/abs/2603.22816
摘要：语言模型越来越多地通过在回答之前编写一步一步的推理来“展示他们的工作”。但是，这些推理步骤是真正使用的，还是在模型已经决定之后产生的装饰性叙述？考虑一下：一位医学AI写道：“患者在导管插入术后出现嗜酸性粒细胞增多症和网状青斑，提示胆固醇栓塞综合征。答案：B。“如果我们删除嗜酸性粒细胞增多症的观察，诊断会改变吗？对于大多数前沿模型，答案是否定的-步骤是装饰。 我们引入步骤级评估：每次删除一个推理句子，并检查答案是否发生变化。这个简单的测试只需要API访问--没有模型权重--每个任务每个模型的成本大约为1 -2美元。 测试了10个前沿模型（GPT-5.4，Claude Opus，DeepSeek-V3.2，MiniMax-M2.5，Kimi-K2.5等），涉及情感，数学，主题分类和医学QA（每个N=376-500），大多数产生装饰性推理：删除任何步骤改变的答案不到17%的时间，而任何一个步骤单独恢复答案。这甚至在数学上也成立，其中较小的模型（0.8-8B）显示出真正的步骤依赖性（55%的必要性）。 有两个模型打破了这一模式：情感方面的MiniMax-M2.5（37%的必要性）和主题分类方面的Kimi-K2.5（39%），但这两个模型都简化了其他任务。忠诚是特定于模型和任务的。 我们还发现了“输出刚性”：在相同的医学问题上，Claude Opus编写了11个诊断步骤，而GPT-OSS-120 B输出了一个标记。机制分析（注意模式）证实，CoT的注意力下降更多的后期层的装饰任务（33%）比忠实的（20%）。 含义：来自前沿模型的逐步解释在很大程度上是装饰性的，每个模型每个领域的评估是必不可少的，而训练目标-而不是规模-决定了推理是否真实。
摘要：Language models increasingly "show their work" by writing step-by-step reasoning before answering. But are these reasoning steps genuinely used, or decorative narratives generated after the model has already decided? Consider: a medical AI writes "The patient's eosinophilia and livedo reticularis following catheterization suggest cholesterol embolization syndrome. Answer: B." If we remove the eosinophilia observation, does the diagnosis change? For most frontier models, the answer is no - the step was decorative. We introduce step-level evaluation: remove one reasoning sentence at a time and check whether the answer changes. This simple test requires only API access -- no model weights -- and costs approximately $1-2 per model per task. Testing 10 frontier models (GPT-5.4, Claude Opus, DeepSeek-V3.2, MiniMax-M2.5, Kimi-K2.5, and others) across sentiment, mathematics, topic classification, and medical QA (N=376-500 each), the majority produce decorative reasoning: removing any step changes the answer less than 17% of the time, while any single step alone recovers the answer. This holds even on math, where smaller models (0.8-8B) show genuine step dependence (55% necessity). Two models break the pattern: MiniMax-M2.5 on sentiment (37% necessity) and Kimi-K2.5 on topic classification (39%) - but both shortcut other tasks. Faithfulness is model-specific and task-specific. We also discover "output rigidity": on the same medical questions, Claude Opus writes 11 diagnostic steps while GPT-OSS-120B outputs a single token. Mechanistic analysis (attention patterns) confirms that CoT attention drops more in late layers for decorative tasks (33%) than faithful ones (20%). Implications: step-by-step explanations from frontier models are largely decorative, per-model per-domain evaluation is essential, and training objectives - not scale - determine whether reasoning is genuine.
【15】Caterpillar of Thoughts: The Optimal Test-Time Algorithm for Large Language Models
标题：思想毛毛虫：大型语言模型的最佳测试时间算法
链接：https://arxiv.org/abs/2603.22784
摘要：当允许使用额外的测试时计算时，大型语言模型（LLM）通常可以产生更好的输出，例如采样，思想链，回溯或修改部分解决方案。尽管这些技术在经验上取得了越来越大的成功，但对于推理时间计算应该如何构建，或者什么构成了固定计算预算的最佳使用，理论上的理解有限。 我们将测试时计算建模为与马尔可夫链交互的算法：在任何时候，算法都可以从任何先前观察到的状态恢复生成。也就是说，与被动绘制状态的标准马尔可夫链不同，我们允许算法在任何时候回溯到马尔可夫链的任何先前观察到的状态。许多现有的测试时间算法，诸如思想链（CoT）（Wei等人，2023）、思想树（ToT）（Yao等人，2023），或最佳的-$k$（布朗等人，2024）可以被视为该模型中的特定算法。 我们证明，虽然回溯可以减少代的数量呈指数，一个非常有限的形式回溯理论上是足够的。也就是说，我们表明，最佳算法总是产生一个毛毛虫树。也就是说，如果我们删除的最佳算法产生的状态树的叶子，我们得到一个路径。受我们的最佳算法的特征，我们提出了思想的毛虫（CaT），一个新的测试时间计算算法，减少令牌/状态生成的数量。我们的经验评估表明，与ToT相比，CaT实现了更好的成功率，同时还减少了令牌生成的数量。
摘要：Large language models (LLMs) can often produce substantially better outputs when allowed to use additional test-time computation, such as sampling, chain of thought, backtracking, or revising partial solutions. Despite the growing empirical success of such techniques, there is limited theoretical understanding of how inference time computation should be structured, or what constitutes an optimal use of a fixed computation budget. We model test-time computation as an algorithm interacting with a Markov chain: at any point, the algorithm may resume generation from any previously observed state. That is, unlike standard Markov chains where the states are drawn passively, we allow the algorithm to backtrack to any previously observed state of the Markov chain at any time. Many of the existing test-time algorithms, such as Chain-of-Thought (CoT) (Wei et al., 2023), Tree-of-Thoughts (ToT) (Yao et al., 2023), or Best-of-$k$ (Brown et al., 2024) could be seen as specific algorithms in this model. We prove that while backtracking can reduce the number of generations exponentially, a very limited form of backtracking is theoretically sufficient. Namely, we show that the optimal algorithm always generates a caterpillar tree. That is, if we remove the leaves of the state tree generated by the optimal algorithm, we obtain a path. Motivated by our characterization of the optimal algorithm, we present Caterpillar of Thoughts (CaT), a new test-time computation algorithm, reducing the number of token/state generations. Our empirical evaluation shows that CaT, compared to ToT, achieves a better success rate while also reducing the number of token generations.
【16】KALAVAI: Predicting When Independent Specialist Fusion Works -- A Quantitative Model for Post-Hoc Cooperative LLM Training
标题：KALAVAI：预测独立专家融合何时起作用--一个事后合作LLM训练的定量模型
链接：https://arxiv.org/abs/2603.22755
摘要：独立训练的领域专家可以事后融合成一个单一的模型，该模型优于任何单个专家，并且增益是可预测的：增益= 0.82 x发散度-2.72（R^2 = 0.856，n=6，3-26%发散度）。这使从业者能够在提交计算之前估计合作价值。在KALAVAI协议中，贡献者独立地微调共享检查点的副本，然后提交轻量级MoE路由（500步）。收益是一致的：在410 M（+/-0.02%，3个种子）时+7.72%，在1B（+/-0.01%，3个种子）时+7.49%，在6.9B时+6.53%，每个都超过最好的专家。路由器在<10^{-5} nats内匹配域-oracle路由。跨语言融合（泰米尔语/约鲁巴语/威尔士语/代码）达到+21.76%，约鲁巴语困惑度下降41.9至7.7。一个20个贡献者的联盟达到了+16.71%（+/-0.07pp，3个种子）。共享初始化是必要的：检查点不匹配会降低路由性能。冷冻层在10，000步以下是可选的，超过10，000步是有益的。学习路由是必不可少的：与最佳专家相比，均匀平均会降低-1.2%，而任何经过训练的路由器都能实现Oracle最优分配。
摘要 ：Independently trained domain specialists can be fused post-hoc into a single model that outperforms any individual specialist, and the gain is predictable: gain = 0.82 x divergence - 2.72 (R^2 = 0.856, n=6, 3-26% divergence). This enables practitioners to estimate cooperative value before committing compute. Below ~3.3% divergence, gains approach zero.In the KALAVAI protocol, contributors fine-tune copies of a shared checkpoint independently, then submit for lightweight MoE routing (500 steps). Gains are consistent: +7.72% at 410M (+/-0.02%, 3 seeds), +7.49% at 1B (+/-0.01%, 3 seeds), +6.53% at 6.9B, each over the best specialist. The router matches domain-oracle routing within <10^{-5} nats. Cross-lingual fusion (Tamil/Yoruba/Welsh/Code) achieves +21.76%, with Yoruba perplexity falling 41.9 to 7.7. A 20-contributor federation achieves +16.71% (+/-0.07pp, 3 seeds).Three requirements bound the protocol. Shared initialisation is necessary: checkpoint mismatch degrades routing. Frozen layers are optional below ~10,000 steps and beneficial beyond. Learned routing is essential: uniform averaging degrades by -1.2% vs. best specialist, while any trained router achieves oracle-optimal assignment.
【17】Improving LLM Predictions via Inter-Layer Structural Encoders
标题：通过层间结构编码器改进LLM预测
链接：https://arxiv.org/abs/2603.22665
备注：17 pages, 3 figures. Equal contribution by first two authors
摘要：大型语言模型（LLM）中的标准做法是将预测基于最后一层令牌表示。然而，最近的研究表明，中间层编码大量的信息，其中可能包含更多的任务相关的功能比单独的最后一层表示。重要的是，它表明，对于不同的任务，不同的层可能是最佳的。在这项工作中，我们介绍了层间结构编码器（ILSE），一个强大的结构方法来学习一个有效的表示从LLM的内部层表示在一起。ILSE的核心是Cayley-Encoder，这是一种数学上接地的几何编码器，它利用扩展Cayley图进行有效的层间信息传播。我们在13个分类和语义相似性任务中评估ILSE，其中9个预训练的LLM的参数范围从1400万到80亿。ILSE始终优于基线和现有方法，在准确性方面提高了44%，在相似性指标方面提高了25%。我们进一步表明，ILSE是数据效率在Few-Shot制度，并可以使小LLM与大得多的模型竞争。
摘要：The standard practice in Large Language Models (LLMs) is to base predictions on the final-layer token representations. Recent studies, however, show that intermediate layers encode substantial information, which may contain more task-relevant features than the final-layer representations alone. Importantly, it was shown that for different tasks, different layers may be optimal. In this work we introduce Inter-Layer Structural Encoders (ILSE), a powerful structural approach to learn one effective representation from the LLM's internal layer representations all together. Central to ILSE is Cayley-Encoder, a mathematically grounded geometric encoder that leverages expander Cayley graphs for efficient inter-layer information propagation. We evaluate ILSE across 13 classification and semantic similarity tasks with 9 pre-trained LLMs ranging from 14 million to 8 billion parameters. ILSE consistently outperforms baselines and existing approaches, achieving up to 44% improvement in accuracy and 25% in similarity metrics. We further show that ILSE is data-efficient in few-shot regimes and can make small LLMs competitive with substantially larger models.
【18】Benchmarking Multi-Agent LLM Architectures for Financial Document Processing: A Comparative Study of Orchestration Patterns, Cost-Accuracy Tradeoffs and Production Scaling Strategies
标题：金融文档处理的多代理LLM体系结构基准：编排模式、成本准确性权衡和生产规模策略的比较研究
链接：https://arxiv.org/abs/2603.22651
摘要：采用大型语言模型（LLM）从财务文档中提取结构化信息的速度已经迅速加快，但生产部署面临着基本的架构决策，经验指导有限。我们提出了一个系统的基准比较四个多智能体编排架构：顺序流水线，并行扇出合并，层次管理工人和自反自校正循环。这些是在10，000份SEC文件（10-K，10-Q和8-K表格）的语料库中对五个前沿和开放权重LLM进行评估的。我们的评估涵盖了25个提取字段类型，涵盖治理结构、高管薪酬和财务指标，沿着五个轴进行衡量：字段级F1、文档级准确性、端到端延迟、每份文档成本和令牌效率。我们发现，自反架构实现了最高的字段级F1（0.943），但在2.3倍的成本顺序基线，而分层架构占据成本精度帕累托边界上最有利的位置（F1 0.921在1.4倍的成本）。我们进一步提出消融语义缓存，模型路由和自适应重试策略的研究，表明混合配置可以恢复89%的反射架构的准确性增益只有1.15倍的基线成本。我们从每天1 K到100 K文档的扩展分析揭示了不明显的吞吐量-准确性降级曲线，可为容量规划提供信息。这些发现为从业者在受监管的金融环境中部署多代理LLM系统提供了可操作的指导。
摘要：The adoption of large language models (LLMs) for structured information extraction from financial documents has accelerated rapidly, yet production deployments face fundamental architectural decisions with limited empirical guidance. We present a systematic benchmark comparing four multi-agent orchestration architectures: sequential pipeline, parallel fan-out with merge, hierarchical supervisor-worker and reflexive self-correcting loop. These are evaluated across five frontier and open-weight LLMs on a corpus of 10,000 SEC filings (10-K, 10-Q and 8-K forms). Our evaluation spans 25 extraction field types covering governance structures, executive compensation and financial metrics, measured along five axes: field-level F1, document-level accuracy, end-to-end latency, cost per document and token efficiency. We find that reflexive architectures achieve the highest field-level F1 (0.943) but at 2.3x the cost of sequential baselines, while hierarchical architectures occupy the most favorable position on the cost-accuracy Pareto frontier (F1 0.921 at 1.4x cost). We further present ablation studies on semantic caching, model routing and adaptive retry strategies, demonstrating that hybrid configurations can recover 89\% of the reflexive architecture's accuracy gains at only 1.15x baseline cost. Our scaling analysis from 1K to 100K documents per day reveals non-obvious throughput-accuracy degradation curves that inform capacity planning. These findings provide actionable guidance for practitioners deploying multi-agent LLM systems in regulated financial environments.
【19】OrgForge-IT: A Verifiable Synthetic Benchmark for LLM-Based Insider Threat Detection
标题：OrgForge-IT：基于LLM的内部威胁检测的可验证合成基准
链接：https://arxiv.org/abs/2603.22499
摘要：合成内部威胁基准面临一致性问题：在没有外部事实约束的情况下生成的语料库不能排除跨工件的矛盾。CERT数据集-该领域的规范基准-也是静态的，缺乏跨表面相关场景，并且早于LLM时代。我们提出了OrgForge-IT，一个可验证的合成基准，其中一个确定性的仿真引擎保持地面真相和语言模型只生成表面散文，使跨工件的一致性的架构保证。该语料库跨越51个模拟日，2，904个遥测记录，噪声率为96.4%，以及四种检测场景，旨在击败三种威胁类别和八种可注射行为的单表面和单日分流策略。 十个模型的排行榜揭示了几个发现：（1）分类和判决准确度分离-八个模型实现了相同的分类F1=0.80，但判决F1=1.0和0.80之间的分裂;（2）基线假阳性率是判决F1的必要伴侣，具有相同判决准确度的模型在分类噪声上相差两个数量级;（3）网络钓鱼场景中的受害者归因将层级分开-层级A模型免除了受感染账户持有人的责任，而层级B模型检测到了攻击，但对受害者进行了错误分类;（4）严格的多信号阈值在结构上排除了单表面疏忽的内部人员，证明了并行的、特定于威胁类别的分流管道的必要性;和（5）代理软件工程培训作为一个多天的时间相关性的力量倍增器，但只有当配对的前沿水平参数规模。最后，提示敏感性分析表明，非结构化的提示诱导词汇幻觉，激励双轨评分框架分离的推理能力，及时遵守。 OrgForge-IT在MIT许可下开源。
摘要 ：Synthetic insider threat benchmarks face a consistency problem: corpora generated without an external factual constraint cannot rule out cross-artifact contradictions. The CERT dataset -- the field's canonical benchmark -- is also static, lacks cross-surface correlation scenarios, and predates the LLM era. We present OrgForge-IT, a verifiable synthetic benchmark in which a deterministic simulation engine maintains ground truth and language models generate only surface prose, making cross-artifact consistency an architectural guarantee. The corpus spans 51 simulated days, 2,904 telemetry records at a 96.4% noise rate, and four detection scenarios designed to defeat single-surface and single-day triage strategies across three threat classes and eight injectable behaviors. A ten-model leaderboard reveals several findings: (1) triage and verdict accuracy dissociate - eight models achieve identical triage F1=0.80 yet split between verdict F1=1.0 and 0.80; (2) baseline false-positive rate is a necessary companion to verdict F1, with models at identical verdict accuracy differing by two orders of magnitude on triage noise; (3) victim attribution in the vishing scenario separates tiers - Tier A models exonerate the compromised account holder while Tier B models detect the attack but misclassify the victim; (4) rigid multi-signal thresholds structurally exclude single-surface negligent insiders, demonstrating the necessity of parallel, threat-class-specific triage pipelines; and (5) agentic software-engineering training acts as a force multiplier for multi-day temporal correlation, but only when paired with frontier-level parameter scale. Finally, prompt sensitivity analysis reveals that unstructured prompts induce vocabulary hallucination, motivating a two-track scoring framework separating prompt adherence from reasoning capability. OrgForge-IT is open source under the MIT license.
【20】Functional Component Ablation Reveals Specialization Patterns in Hybrid Language Model Architectures
标题：功能成分消融揭示混合语言模型体系结构中的专业化模式
链接：https://arxiv.org/abs/2603.22473
备注：22 pages, 7 figures, 6 tables. Code and data available at https://github.com/hborobia/hybrid-component-ablation
摘要：将注意力与状态空间模型（SSM）或线性注意力相结合的混合语言模型可以提高效率，但这两个组件是否真正得到利用仍不清楚。我们提出了一个功能组件消融框架，应用于两个亚1B混合模型- Qwen3.5-0.8B（顺序：门控DeltaNet + softmax注意力）和Falcon-H1-0.5B（并行：Mamba-2 +注意力）-与纯Transformer控制（Qwen2.5-0.5B）。通过分组消融、逐层扫描、位置消融、匹配随机对照和五个基准的困惑分析，我们建立了四个发现：（1）两种成分类型都是必不可少的，没有一种被绕过;（2）替代成分（线性注意力或SSM）是主要的语言建模支柱，当删除时会导致>35，000倍的困惑度下降，而注意力则为~ 82倍;（3）组件重要性遵循位置梯度，早期层不成比例地关键;（4）混合架构对随机层移除的弹性比纯Transformers大20- 119倍，揭示了组件类型之间的内置功能冗余。这些结果为混合模型压缩、架构设计和容错部署提供了可操作的指导。
摘要：Hybrid language models combining attention with state space models (SSMs) or linear attention offer improved efficiency, but whether both components are genuinely utilized remains unclear. We present a functional component ablation framework applied to two sub-1B hybrid models -- Qwen3.5-0.8B (sequential: Gated DeltaNet + softmax attention) and Falcon-H1-0.5B (parallel: Mamba-2 + attention) -- with a pure Transformer control (Qwen2.5-0.5B). Through group ablations, layer-wise sweeps, positional ablations, matched random controls, and perplexity analysis across five benchmarks, we establish four findings: (1) both component types are essential and neither is bypassed; (2) the alternative component (linear attention or SSM) is the primary language modeling backbone, causing >35,000x perplexity degradation when removed versus ~82x for attention; (3) component importance follows a positional gradient, with early layers being disproportionately critical; and (4) hybrid architectures exhibit 20-119x greater resilience to random layer removal than pure Transformers, revealing built-in functional redundancy between component types. These results provide actionable guidance for hybrid model compression, architecture design, and fault-tolerant deployment.
【21】SkillRouter: Retrieve-and-Rerank Skill Selection for LLM Agents at Scale
标题：SkillRouter：LLM代理大规模检索和重新排序技能选择
链接：https://arxiv.org/abs/2603.22455
摘要：随着LLM代理生态系统的发展，可用技能（工具，插件）的数量已经达到数万，这使得将所有技能注入代理的上下文变得不可行。这就产生了对技能路由的需求--在给定用户任务的情况下，从一个大的技能池中检索最相关的技能。社区技能库中普遍存在的功能重叠使问题变得更加复杂，其中许多技能具有相似的名称和目的，但在实现细节上有所不同。尽管它的实际重要性，技能路由仍然没有得到充分的探讨。当前的代理架构采用渐进式披露设计-只向代理公开技能名称和描述，同时保持完整的实现主体隐藏-隐式地将元数据视为足以进行选择。我们通过对约8万美元技能和75个专家验证查询的基准进行系统的实证研究来挑战这一假设。我们的关键发现是，技能主体（完整的实现文本）是决定性的信号：删除它会导致所有检索方法的29- 44个百分点的下降，跨编码器的注意力分析显示，91.7%的注意力集中在身体领域。基于这一发现，我们提出了SkillRouter，这是一个两阶段的检索和重新排序管道，总共只有1.2B参数（0.6B编码器+0.6B重新排序器）。SkillRouter实现了74.0%的top-1路由准确性，并在我们评估的紧凑型和zero-shot基线中提供了最强的平均结果，同时仍然可以部署在消费者硬件上。
摘要：As LLM agent ecosystems grow, the number of available skills (tools, plugins) has reached tens of thousands, making it infeasible to inject all skills into an agent's context. This creates a need for skill routing -- retrieving the most relevant skills from a large pool given a user task. The problem is compounded by pervasive functional overlap in community skill repositories, where many skills share similar names and purposes yet differ in implementation details. Despite its practical importance, skill routing remains under-explored. Current agent architectures adopt a progressive disclosure design -- exposing only skill names and descriptions to the agent while keeping the full implementation body hidden -- implicitly treating metadata as sufficient for selection. We challenge this assumption through a systematic empirical study on a benchmark of ~$80K skills and 75 expert-verified queries. Our key finding is that the skill body (full implementation text) is the decisive signal: removing it causes 29--44 percentage point degradation across all retrieval methods, and cross-encoder attention analysis reveals 91.7% of attention concentrating on the body field. Motivated by this finding, we propose SkillRouter, a two-stage retrieve-and-rerank pipeline totaling only 1.2B parameters (0.6B encoder + 0.6B reranker). SkillRouter achieves 74.0% top-1 routing accuracy and delivers the strongest average result among the compact and zero-shot baselines we evaluate, while remaining deployable on consumer hardware.
【22】Sparse but Critical: A Token-Level Analysis of Distributional Shifts in RLVR Fine-Tuning of LLMs
标题：稀疏但关键：LLM WLVR微调中分布转变的代币级分析
链接：https://arxiv.org/abs/2603.22446
备注：Published as a conference paper at the International Conference on Learning Representations (ICLR 2026)
摘要：具有可验证奖励的强化学习（RLVR）显著改善了大型语言模型（LLM）中的推理，但这些改进背后的令牌级机制仍不清楚。我们提出了一个系统的实证研究RLVR的分布效应围绕三个主要分析：（1）令牌级表征的基础和RL模型之间的分布变化，（2）令牌级分布变化的影响序列级推理性能通过交叉采样干预，（3）细粒度的机制，这些变化在令牌级。我们发现，RL微调引起高度稀疏和有针对性的变化，只有一小部分令牌分布表现出有意义的分歧之间的基础和RL的政策。我们进一步表征这些变化的结构和演变，通过分析令牌熵，位置浓度和概率质量的重新分配。为了评估这些稀疏变化的功能重要性，我们进行了交叉采样实验，选择性地在具有不同干预预算的基础模型和RL模型之间交换令牌选择。我们发现，只插入一小部分RL采样令牌到基地代逐步恢复RL性能增益，而注入一个类似的小数目的基地令牌的选择，否则RL生成的序列崩溃的性能基础水平，隔离一小部分令牌级决策直接负责RLVR的性能增益。最后，我们探索了优势信号的差异加权变体作为诊断干预，发现它们可以产生超过基线的改进。总之，我们的研究结果揭示了由RLVR引起的分布变化，并提供了一个细粒度的，令牌级的镜头来理解RLVR微调作为一个有针对性的细化过程。
摘要 ：Reinforcement learning with verifiable rewards (RLVR) has significantly improved reasoning in large language models (LLMs), yet the token-level mechanisms underlying these improvements remain unclear. We present a systematic empirical study of RLVR's distributional effects organized around three main analyses: (1) token-level characterization of distributional shifts between base and RL models, (2) the impact of token-level distributional shifts on sequence-level reasoning performance through cross-sampling interventions, and (3) fine-grained mechanics of these shifts at the token level. We find that RL fine-tuning induces highly sparse and targeted changes, with only a small fraction of token distributions exhibiting meaningful divergence between the base and RL policies. We further characterize the structure and evolution of these shifts through analyses of token entropy, positional concentration, and reallocation of probability mass. To assess the functional importance of these sparse changes, we conduct cross-sampling experiments that selectively swap token choices between the base and RL models with varying intervention budgets. We show that inserting only a small fraction of RL-sampled tokens into base generations progressively recovers RL performance gains, while injecting a similarly small number of base token choices into otherwise RL-generated sequences collapses performance to base levels, isolating a small set of token-level decisions directly responsible for RLVR's performance gains. Finally, we explore divergence-weighted variants of the advantage signal as a diagnostic intervention, finding that they can yield improvements over baselines. Together, our results shed light on the distributional changes induced by RLVR and provide a fine-grained, token-level lens for understanding RLVR fine-tuning as a targeted refinement process.
【23】Large Language Models for Missing Data Imputation: Understanding Behavior, Hallucination Effects, and Control Mechanisms
标题：缺失数据归责的大型语言模型：理解行为、幻觉效应和控制机制
链接：https://arxiv.org/abs/2603.22332
摘要：数据插补是处理真实世界数据集中缺失值的基础技术，这些数据集经常受到缺失的困扰。尽管最近取得了进展，但之前关于基于大型语言模型的插补的研究仍然受到可扩展性挑战、受限的跨模型比较以及对小型或特定领域数据集进行的评估的限制。此外，异构的实验方案和不一致的治疗缺失机制（MCAR，MAR和MNAR）阻碍了跨方法的系统基准。本文研究了大语言模型在表格数据集缺失数据填补中的鲁棒性，使用zero-shot提示工程方法。为此，我们提出了一个全面的基准研究，比较五个广泛使用的LLM对六个国家的最先进的插补基线。实验设计在MCAR，MAR和MNAR机制下的29个数据集（包括9个合成数据集）上评估这些方法，缺失率高达20%。结果表明，领先的LLM，特别是Gemini 3.0 Flash和Claude 4.5 Sonnet，与传统方法相比，在真实世界的开源数据集上始终实现卓越的性能。然而，这种优势似乎与模型在互联网规模语料库的预训练期间学习到的特定领域模式密切相关。相比之下，在合成数据集上，MICE等传统方法优于LLM，这表明LLM的有效性是由语义上下文驱动的，而不是纯粹的统计重建。此外，我们确定了一个明确的权衡：虽然LLM在插补质量方面表现出色，但它们会产生更高的计算时间和金钱成本。总的来说，这项研究提供了一个大规模的比较分析，定位LLM作为有前途的语义驱动的输入复杂的表格数据。
摘要：Data imputation is a cornerstone technique for handling missing values in real-world datasets, which are often plagued by missingness. Despite recent progress, prior studies on Large Language Models-based imputation remain limited by scalability challenges, restricted cross-model comparisons, and evaluations conducted on small or domain-specific datasets. Furthermore, heterogeneous experimental protocols and inconsistent treatment of missingness mechanisms (MCAR, MAR, and MNAR) hinder systematic benchmarking across methods. This work investigates the robustness of Large Language Models for missing data imputation in tabular datasets using a zero-shot prompt engineering approach. To this end, we present a comprehensive benchmarking study comparing five widely used LLMs against six state-of-the-art imputation baselines. The experimental design evaluates these methods across 29 datasets (including nine synthetic datasets) under MCAR, MAR, and MNAR mechanisms, with missing rates of up to 20\%. The results demonstrate that leading LLMs, particularly Gemini 3.0 Flash and Claude 4.5 Sonnet, consistently achieve superior performance on real-world open-source datasets compared to traditional methods. However, this advantage appears to be closely tied to the models' prior exposure to domain-specific patterns learned during pre-training on internet-scale corpora. In contrast, on synthetic datasets, traditional methods such as MICE outperform LLMs, suggesting that LLM effectiveness is driven by semantic context rather than purely statistical reconstruction. Furthermore, we identify a clear trade-off: while LLMs excel in imputation quality, they incur significantly higher computational time and monetary costs. Overall, this study provides a large-scale comparative analysis, positioning LLMs as promising semantics-driven imputers for complex tabular data.
【24】Trained Persistent Memory for Frozen Decoder-Only LLMs
标题：针对冻结的仅解码器LLM的训练持久内存
链接：https://arxiv.org/abs/2603.22329
摘要：仅解码器的语言模型是无状态的：每次向前传递后隐藏的表示都会被丢弃，并且会话之间不会持续任何内容。Jeong（2026 a）表明，经过训练的记忆适配器为冻结的编码器-解码器骨干提供了持久的潜在空间记忆，建立在Jeong（2026 b，c）的横向记忆框架之上。在这里，我们要问的是，同样的原理是否也适用于只有解码器的情况，在这种情况下，不存在交叉注意路径，记忆必须通过自我注意进入。我们采用六种方法-前缀，并行交叉注意，KV扩展，Hebbian内存，上下文门控分支和基于插槽的稀疏写入-到冻结的GPT-2，仅训练一个小适配器$θ_{}$。写规则是共享的;只有读注入从解码器交叉关注变为自关注KV前缀或并行分支。在LoCoMo上，我们发现了一个引人注目的归纳偏差二分法：在1美元乘以容量的情况下，三种具有强结构先验的方法-交叉注意（M.2），Hebbian（M.4）和插槽写入（M.6）-实现了7 -18美元的保留记忆分数和7 -10美元的知识增益ΔK，而其他三种方法失败（< 0.4）。在10倍容量的情况下，所有六个方面都趋于一致，表明差距是架构性的，而不是根本性的。结合Jeong（2026 a）的编码器-解码器结果和Jeong（2026 b，c）的大脑启发模块，这些发现建立了持久潜伏空间记忆作为跨越主要Transformer家族的一般范式。
摘要：Decoder-only language models are stateless: hidden representations are discarded after every forward pass and nothing persists across sessions. Jeong (2026a) showed that trained memory adapters give a frozen encoder-decoder backbone persistent latent-space memory, building on the lateral-memory framework of Jeong (2026b,c). Here we ask whether the same principle transfers to the decoder-only setting, where no cross-attention pathway exists and memory must enter through self-attention alone. We adapt six methods -- prefix, parallel cross-attention, KV extension, Hebbian memory, context-gated branch, and slot-based sparse write -- to a frozen GPT-2, training only a small adapter $θ_{mem}$. The write rule is shared; only the read injection changes from decoder cross-attention to self-attention KV prefix or parallel branch. On LoCoMo we find a striking inductive-bias dichotomy: at $1\times$ capacity, three methods with strong architectural priors -- cross-attention (M.2), Hebbian (M.4), and slot write (M.6) -- achieve retained-memory scores of $7-18\%$ and knowledge gains $ΔK$ of $7-10$, while the other three fail ($< 0.4\%$). At $10\times$ capacity all six converge, showing the gap is architectural, not fundamental. Together with the encoder-decoder results of Jeong (2026a) and the brain-inspired modules of Jeong (2026b,c), these findings establish persistent latent-space memory as a general paradigm spanning major transformer families.
【25】DAQ: Delta-Aware Quantization for Post-Training LLM Weight Compression
标题：DAQ：训练后LLM体重压缩的Delta感知量化
链接：https://arxiv.org/abs/2603.22324
摘要：我们引入了Delta Aware Quantization（DAQ），这是一种无数据的后训练量化框架，可以保留后训练期间获得的知识。标准量化目标最大限度地减少了重建误差，但对基础模型是不可知的，允许量化噪声不成比例地破坏编码后训练行为的小幅度参数增量（$ΔW$）-我们通过量化的镜头分析的效果作为隐式正则化。DAQ用两个增量感知指标（符号保持率和余弦相似性）取代了基于重建的目标，这两个指标直接优化了$ΔW$的方向保真度，只需要基础和训练后的权重矩阵。在FP 8试点研究中，DAQ恢复了标准量化下丢失的特定风格功能，同时保持了一般性能。
摘要：We introduce Delta-Aware Quantization (DAQ), a data-free post-training quantization framework that preserves the knowledge acquired during post-training. Standard quantization objectives minimize reconstruction error but are agnostic to the base model, allowing quantization noise to disproportionately corrupt the small-magnitude parameter deltas ($ΔW$) that encode post-training behavior -- an effect we analyze through the lens of quantization as implicit regularization. DAQ replaces reconstruction-based objectives with two delta-aware metrics -- Sign Preservation Rate and Cosine Similarity -- that directly optimize for directional fidelity of $ΔW$, requiring only the base and post-trained weight matrices. In a pilot FP8 study, DAQ recovers style-specific capabilities lost under standard quantization while maintaining general performance.
【26】CN-Buzz2Portfolio: A Chinese-Market Dataset and Benchmark for LLM-Based Macro and Sector Asset Allocation from Daily Trending Financial News
标题：CN-Buzz 2 Portfolio：来自每日趋势财经新闻的基于LLM的宏观和行业资产配置的中国市场数据集和基准
链接：https://arxiv.org/abs/2603.22305
摘要：大型语言模型（LLM）正在从静态自然语言处理（NLP）任务（包括情感分析和事件提取）快速转变为复杂金融环境中的动态决策代理。然而，LLM演变成自主的金融代理人面临着一个重大的困境，在评估范式。直接实时交易是不可复制的，并且容易因混淆运气和技能而产生结果偏差，而现有的静态基准通常仅限于实体层面的选股，忽视了更广泛的市场关注。为了促进对这些挑战的严格分析，我们引入了CN-Buzz 2 Portfolio，这是一个基于中国市场的可复制基准，将每日趋势新闻映射到宏观和行业资产配置。从2024年到2025年中期，我们的数据集模拟了现实的公众注意力流，要求代理商从高曝光率的叙述中提取投资逻辑，而不是预先过滤的实体新闻。我们提出了一个三阶段CPA代理工作流程，涉及压缩，感知和分配，以评估LLM广泛的资产类别，如交易所交易基金（ETF），而不是个股，从而减少特质波动。对9个LLM的广泛实验揭示了模型如何将宏观层面的叙述转化为投资组合权重的显着差异。这项工作为一般推理和金融决策之间的一致性提供了新的见解，所有数据，代码和实验都被发布，以促进可持续的金融代理研究。
摘要：Large Language Models (LLMs) are rapidly transitioning from static Natural Language Processing (NLP) tasks including sentiment analysis and event extraction to acting as dynamic decision-making agents in complex financial environments. However, the evolution of LLMs into autonomous financial agents faces a significant dilemma in evaluation paradigms. Direct live trading is irreproducible and prone to outcome bias by confounding luck with skill, whereas existing static benchmarks are often confined to entity-level stock picking and ignore broader market attention. To facilitate the rigorous analysis of these challenges, we introduce CN-Buzz2Portfolio, a reproducible benchmark grounded in the Chinese market that maps daily trending news to macro and sector asset allocation. Spanning a rolling horizon from 2024 to mid-2025, our dataset simulates a realistic public attention stream, requiring agents to distill investment logic from high-exposure narratives instead of pre-filtered entity news. We propose a Tri-Stage CPA Agent Workflow involving Compression, Perception, and Allocation to evaluate LLMs on broad asset classes such as Exchange Traded Funds (ETFs) rather than individual stocks, thereby reducing idiosyncratic volatility. Extensive experiments on nine LLMs reveal significant disparities in how models translate macro-level narratives into portfolio weights. This work provides new insights into the alignment between general reasoning and financial decision-making, and all data, codes, and experiments are released to promote sustainable financial agent research.
【27】Sample Transform Cost-Based Training-Free Hallucination Detector for Large Language Models
标题：基于样本转换成本的大型语言模型免训练幻觉检测器
链接：https://arxiv.org/abs/2603.22303
备注：24 pages, 5 figures, 5 tables
摘要：大型语言模型（LLM）中的幻觉仍然是值得信赖的部署的主要障碍，激励检测器准确，轻量级和广泛适用。由于带有提示的LLM定义了条件分布，因此我们认为分布的复杂性是幻觉的指标。然而，分布的密度是未知的，并且样本（即，针对提示生成的响应）是离散分布，这导致在量化分布的复杂性方面的重大挑战。我们建议计算成对样本的令牌嵌入集之间的最优传输距离，这产生了一个Wasserstein距离矩阵，用于测量样本之间的转换成本。这个Wasserstein距离矩阵提供了一种方法来量化由LLM定义的分布的复杂性。基于Wasserstein距离矩阵，我们推导出两个互补的信号：AvgWD，测量平均成本，和EigenWD，测量成本复杂性。这导致了LLM中幻觉的无训练检测器。我们进一步扩展的框架，黑盒LLM通过教师强迫与可访问的教师模型。实验表明，AvgWD和EigenWD与强不确定性基线具有竞争力，并在模型和数据集之间提供互补行为，突出了分布复杂性作为LLM真实性的有效信号。
摘要：Hallucinations in large language models (LLMs) remain a central obstacle to trustworthy deployment, motivating detectors that are accurate, lightweight, and broadly applicable. Since an LLM with a prompt defines a conditional distribution, we argue that the complexity of the distribution is an indicator of hallucination. However, the density of the distribution is unknown and the samples (i.e., responses generated for the prompt) are discrete distributions, which leads to a significant challenge in quantifying the complexity of the distribution. We propose to compute the optimal-transport distances between the sets of token embeddings of pairwise samples, which yields a Wasserstein distance matrix measuring the costs of transforming between the samples. This Wasserstein distance matrix provides a means to quantify the complexity of the distribution defined by the LLM with the prompt. Based on the Wasserstein distance matrix, we derive two complementary signals: AvgWD, measuring the average cost, and EigenWD, measuring the cost complexity. This leads to a training-free detector for hallucinations in LLMs. We further extend the framework to black-box LLMs via teacher forcing with an accessible teacher model. Experiments show that AvgWD and EigenWD are competitive with strong uncertainty baselines and provide complementary behavior across models and datasets, highlighting distribution complexity as an effective signal for LLM truthfulness.
【28】Latent Semantic Manifolds in Large Language Models
标题：大型语言模型中的潜在语义体
链接：https://arxiv.org/abs/2603.22301
摘要：大型语言模型（LLM）在连续的向量空间中执行内部计算，但产生离散的标记-这是一个基本的不匹配，其几何后果仍然知之甚少。我们开发了一个数学框架，将LLM隐藏状态解释为潜在语义流形上的点：配备Fisher信息度量的黎曼子流形，其中令牌对应于划分流形的Voronoi区域。我们定义的可表达性差距，从词汇离散化的语义失真的几何测量，并证明了两个定理：率失真的失真的任何有限的词汇，和线性体积缩放法的可表达性差距通过coarea公式。我们在六个Transformer架构（124 M-1.5 B参数）上验证了这些预测，确认了通用的沙漏固有尺寸轮廓、平滑的曲率结构和斜率为0.87-1.12（R^2 > 0.985）的线性间隙缩放。跨模型的利润率分布揭示了一个持久的硬核边界附近的表示不变的规模，提供了一个几何分解的困惑。我们讨论了架构设计，模型压缩，解码策略和缩放定律的影响
摘要：Large Language Models (LLMs) perform internal computations in continuous vector spaces yet produce discrete tokens -- a fundamental mismatch whose geometric consequences remain poorly understood. We develop a mathematical framework that interprets LLM hidden states as points on a latent semantic manifold: a Riemannian submanifold equipped with the Fisher information metric, where tokens correspond to Voronoi regions partitioning the manifold. We define the expressibility gap, a geometric measure of the semantic distortion from vocabulary discretization, and prove two theorems: a rate-distortion lower bound on distortion for any finite vocabulary, and a linear volume scaling law for the expressibility gap via the coarea formula. We validate these predictions across six transformer architectures (124M-1.5B parameters), confirming universal hourglass intrinsic dimension profiles, smooth curvature structure, and linear gap scaling with slopes 0.87-1.12 (R^2 > 0.985). The margin distribution across models reveals a persistent hard core of boundary-proximal representations invariant to scale, providing a geometric decomposition of perplexity. We discuss implications for architecture design, model compression, decoding strategies, and scaling laws
【29】Between the Layers Lies the Truth: Uncertainty Estimation in LLMs Using Intra-Layer Local Information Scores
标题：层之间谎言真相：使用层内本地信息分数的LLM中的不确定性估计
链接：https://arxiv.org/abs/2603.22299
摘要 ：大型语言模型（LLM）通常是错误的，这使得可靠的不确定性估计（UE）至关重要。基于输出的语义学便宜但脆弱，而探测内部表示是有效的，但高维，难以转移。 我们提出了一个紧凑的，每实例UE的方法，评分跨层协议模式的内部表示使用一个单一的前向通过。 在三个模型中，我们的方法匹配探测分布，平均对角差异最多为$-1.8$ AUPRC百分点和$+4.9$ Brier得分点。在跨数据集传输下，它始终优于探测，实现了高达$+2.86$ AUPRC和$+21.02$ Brier点的非对角增益。在4位仅加权量化下，它保持鲁棒性，平均比探测提高$+1.94$ AUPRC点和$+5.33$ Brier点。 除了性能，检查特定的层与层之间的交互揭示了不同模型如何编码不确定性的差异。总之，我们的UE方法提供了一个轻量级的，紧凑的手段来捕捉LLM中的可转移的不确定性。
摘要：Large language models (LLMs) are often confidently wrong, making reliable uncertainty estimation (UE) essential. Output-based heuristics are cheap but brittle, while probing internal representations is effective yet high-dimensional and hard to transfer. We propose a compact, per-instance UE method that scores cross-layer agreement patterns in internal representations using a single forward pass. Across three models, our method matches probing in-distribution, with mean diagonal differences of at most $-1.8$ AUPRC percentage points and $+4.9$ Brier score points. Under cross-dataset transfer, it consistently outperforms probing, achieving off-diagonal gains up to $+2.86$ AUPRC and $+21.02$ Brier points. Under 4-bit weight-only quantization, it remains robust, improving over probing by $+1.94$ AUPRC points and $+5.33$ Brier points on average. Beyond performance, examining specific layer--layer interactions reveals differences in how disparate models encode uncertainty. Altogether, our UE method offers a lightweight, compact means to capture transferable uncertainty in LLMs.
【30】TIPS: Turn-Level Information-Potential Reward Shaping for Search-Augmented LLMs
标题：TIPS：回合级信息-搜索增强LLM的潜在回报塑造
链接：https://arxiv.org/abs/2603.22293
备注：Code: https://github.com/ucsd-wang-lab-lm/tips
摘要：用强化学习（RL）训练的搜索增强大型语言模型（LLM）在开放域问答（QA）方面取得了很好的效果，但训练仍然是一个重大挑战。优化往往是不稳定的，由于稀疏的奖励和困难的信用分配跨越推理和工具调用。为了解决这个问题，我们引入了回合信息潜在奖励塑造（TIPS），这是一个简单的框架，它根据教师模型下正确答案的可能性增加，为每个推理+工具调用段分配密集的回合奖励。通过利用基于潜力的奖励塑造，TIPS提供了细粒度和策略不变的指导，克服了仅结果优化的局限性。在七个QA基准上进行评估，TIPS始终优于GRPO/PPO基线，并大大提高了培训稳定性。例如，使用Qwen-2.5 7 B指令模型，相对于PPO，TIPS将平均精确匹配分数提高了11.8%，F1提高了13.6%。我们的研究结果表明，回合信息潜在的奖励塑造提供了一个有效的和通用的解决方案，稀疏奖励信用分配多回合LLM推理。
摘要：Search-augmented large language models (LLMs) trained with reinforcement learning (RL) have achieved strong results on open-domain question answering (QA), but training still remains a significant challenge. The optimization is often unstable due to sparse rewards and difficult credit assignments across reasoning and tool calls. To address this, we introduce Turn-Level Information Potential Reward Shaping (TIPS), a simple framework that assigns dense, turn-level rewards to each reasoning + tool-call segment based on the increased likelihood of the correct answer under a teacher model. By leveraging the potential-based reward shaping, TIPS offers fine-grained and policy-invariant guidance that overcomes the limitations of outcome-only optimization. Evaluated on seven QA benchmarks, TIPS consistently outperforms GRPO/PPO baselines and substantially improves training stability. For instance, with a Qwen-2.5 7B Instruct model, TIPS improves the average Exact Match score by 11.8% and F1 by 13.6% relative to PPO. Our results demonstrate that turn-level information-potential reward shaping provides an effective and general solution to sparse-reward credit assignment for multi-turn LLM reasoning.
【31】Evaluating Prompting Strategies for Chart Question Answering with Large Language Models
标题：基于大语言模型的图表问句生成策略评估
链接：https://arxiv.org/abs/2603.22288
摘要：测试策略影响LLM推理性能，但它们在基于图表的QA中的作用仍然未被探索。我们在ChartQA数据集上对GPT-3.5、GPT-4和GPT-4 o上的四种广泛使用的提示范式（Zero-Shot、Few-Shot、Zero-Shot链和Few-Shot思维链）进行了系统评估。我们的框架只在结构化图表数据上运行，将提示结构隔离为唯一的实验变量，并使用两个指标评估性能：准确性和精确匹配。来自1，200个不同ChartQA样本的结果显示，Few-Shot Chain-of-Thought提示始终产生最高的准确率（高达78.2%），特别是在推理密集型问题上，而Few-Shot提示提高了格式依从性。Zero-Shot仅在高容量模型上执行简单任务时表现良好。这些发现为结构化数据推理任务中选择提示策略提供了可操作的指导，并对现实应用中的效率和准确性产生了影响。
摘要：Prompting strategies affect LLM reasoning performance, but their role in chart-based QA remains underexplored. We present a systematic evaluation of four widely used prompting paradigms (Zero-Shot, Few-Shot, Zero-Shot Chain-of-Thought, and Few-Shot Chain-of-Thought) across GPT-3.5, GPT-4, and GPT-4o on the ChartQA dataset. Our framework operates exclusively on structured chart data, isolating prompt structure as the only experimental variable, and evaluates performance using two metrics: Accuracy and Exact Match. Results from 1,200 diverse ChartQA samples show that Few-Shot Chain-of-Thought prompting consistently yields the highest accuracy (up to 78.2\%), particularly on reasoning-intensive questions, while Few-Shot prompting improves format adherence. Zero-Shot performs well only with high-capacity models on simpler tasks. These findings provide actionable guidance for selecting prompting strategies in structured data reasoning tasks, with implications for both efficiency and accuracy in real-world applications.
【32】TTQ: Activation-Aware Test-Time Quantization to Accelerate LLM Inference On The Fly
标题：TTQ：激活感知测试时间量化，以加速LLM动态推理
链接：https://arxiv.org/abs/2603.19296
备注：25 pages
摘要：为了解决大型基础模型的巨大计算需求，已经引入了无需再训练的激活感知压缩技术。然而，由于这些方法高度依赖于校准数据，因此对于看不见的下游任务可能会出现域偏移问题。我们提出了一个测试时间量化（TTQ）框架，压缩大型模型的飞行在推理时间来解决这个问题。通过有效的在线校准，即时激活感知量化可以适应每个提示，而不管下游任务如何，同时实现推理加速。几个实验表明，TTQ可以提高量化性能超过国家的最先进的基线。
摘要：To tackle the huge computational demand of large foundation models, activation-aware compression techniques without retraining have been introduced. However, since these methods highly rely on calibration data, domain shift issues may arise for unseen downstream tasks. We propose a test-time quantization (TTQ) framework which compresses large models on the fly at inference time to resolve this issue. With an efficient online calibration, instant activation-aware quantization can adapt every prompt regardless of the downstream tasks, yet achieving inference speedup. Several experiments demonstrate that TTQ can improve the quantization performance over state-of-the-art baselines.
【33】Prompt Amplification and Zero-Shot Late Fusion in Audio-Language Models for Speech Emotion Recognition
标题：语音情感识别的音频语言模型中的即时放大和Zero-Shot后期融合
链接：https://arxiv.org/abs/2603.23057
摘要 ：音频语言模型（ALM）在理解语音和非语音音频方面取得了长足的进步。然而，领域专家的基础模型（FM）仍然是封闭式语音处理任务（如语音情感识别（SER））的最佳选择。将ALM用于Zero-shot SER是一种流行的选择，但它们与专家合作以实现最先进（SOTA）性能的潜力尚未开发。我们提出了一种后期融合方法，将来自双编码器ALM的zero-shot情感估计与专家FM相结合。为了处理情感的模糊性和对提示选择的敏感性，1）我们使用一个简单的提示集合，2）提出一种称为提示放大的新技术，它重复音频和文本查询，以发现更强的zero-shot能力。我们通过使用三个双编码器ALM和两个FM来评估T-S，并在三个语音情感识别数据集上报告了SOTA基线（如WavLM-Large）的改进来证明我们的技术的有效性。
摘要：Audio-Language Models (ALMs) are making strides in understanding speech and non-speech audio. However, domain-specialist Foundation Models (FMs) remain the best for closed-ended speech processing tasks such as Speech Emotion Recognition (SER). Using ALMs for Zero-shot SER is a popular choice, but their potential to work with specialists to achieve state-of-the-art (SOTA) performance remains unexplored. We propose ZS-Fuse, a late-fusion method that combines zero-shot emotion estimates from a dual-encoder ALM with specialist FMs. To handle ambiguity in emotions and sensitivity to prompt choice, 1) we use a simple prompt ensemble and 2) suggest a novel technique called prompt amplification, which repeats audio and text queries to discover stronger zero-shot capabilities. We demonstrate the efficacy of our technique by evaluating ZS-Fuse with three dual-encoder ALMs and two FMs, and report improvements over SOTA baselines, such as WavLM-Large, on three speech emotion recognition datasets.
【34】Demystifying Low-Rank Knowledge Distillation in Large Language Models: Convergence, Generalization, and Information-Theoretic Guarantees
标题：揭开大型语言模型中低级知识提炼的神秘面纱：收敛、概括和信息理论保证
链接：https://arxiv.org/abs/2603.22355
摘要：知识蒸馏已经成为一种强大的技术，用于将大型语言模型（LLM）压缩成高效的，可部署的架构，同时保留其高级功能。低秩知识蒸馏的最新进展，特别是像低秩克隆（LRC）这样的方法，已经证明了显着的经验成功，实现了与全参数蒸馏相当的性能，大大减少了训练数据和计算开销。然而，这些方法背后的理论基础仍然知之甚少。在本文中，我们建立了一个严格的理论框架，在语言模型的低秩知识蒸馏。我们证明，在温和的假设下，低秩投影保持优化动力学，产生明确的收敛速度为O（1/\sqrt{T}）$。我们推导出的泛化边界的特征模型压缩和泛化能力之间的基本权衡，泛化误差尺度与秩参数为O（r（m+n）/\sqrt{n}）$。此外，我们提供了一个信息理论分析的激活克隆机制，揭示其作用，最大限度地提高教师和学生的中间表征之间的互信息。我们的理论结果提供了原则性的指导原则秩选择，数学上建议最佳秩r^* = O（\sqrt{n}）$其中$n$是样本大小。标准语言建模基准的实验验证证实了我们的理论预测，表明经验收敛，排名缩放和泛化行为与我们的界限密切相关。
摘要：Knowledge distillation has emerged as a powerful technique for compressing large language models (LLMs) into efficient, deployable architectures while preserving their advanced capabilities. Recent advances in low-rank knowledge distillation, particularly methods like Low-Rank Clone (LRC), have demonstrated remarkable empirical success, achieving comparable performance to full-parameter distillation with significantly reduced training data and computational overhead. However, the theoretical foundations underlying these methods remain poorly understood. In this paper, we establish a rigorous theoretical framework for low-rank knowledge distillation in language models. We prove that under mild assumptions, low-rank projection preserves the optimization dynamics, yielding explicit convergence rates of $O(1/\sqrt{T})$. We derive generalization bounds that characterize the fundamental trade-off between model compression and generalization capability, showing that the generalization error scales with the rank parameter as $O(r(m+n)/\sqrt{n})$. Furthermore, we provide an information-theoretic analysis of the activation cloning mechanism, revealing its role in maximizing the mutual information between the teacher's and student's intermediate representations. Our theoretical results offer principled guidelines for rank selection, mathematically suggesting an optimal rank $r^* = O(\sqrt{n})$ where $n$ is the sample size. Experimental validation on standard language modeling benchmarks confirms our theoretical predictions, demonstrating that the empirical convergence, rank scaling, and generalization behaviors align closely with our bounds.
Graph相关(图学习|图神经网络|图优化等)(12篇)
【1】Graph Energy Matching: Transport-Aligned Energy-Based Modeling for Graph Generation
标题：图能量匹配：用于图生成的传输对齐的基于能量的建模
链接：https://arxiv.org/abs/2603.23398
摘要：离散域的基于能量的模型，如图，显式地捕获相对可能性，自然地实现可组合的概率推理任务，如条件生成或在测试时强制执行约束。然而，离散的基于能量的模型通常难以实现高效和高质量的采样，因为不支持的区域通常包含虚假的局部最小值，从而捕获采样器并导致训练不稳定。这在历史上导致了相对于离散扩散模型的保真度差距。我们介绍了图能量匹配（GEM），一个生成框架的图形，关闭这个保真度的差距。受Jordan-Kinderlehrer-Otto（JKO）方案的传输图优化视角的启发，GEM学习了一种置换不变的势能，同时提供了从噪声到数据的传输对齐指导，并在高数据可能性区域内细化样本。此外，我们引入了一种采样协议，该协议利用基于能量的开关无缝地桥接：（i）快速，梯度引导的高概率区域传输到（ii）用于探索学习图形分布的混合机制。在分子图基准上，GEM匹配或超过强离散扩散基线。除了样本质量之外，相对似然的显式建模还可以在推理时进行有针对性的探索，促进成分生成，属性约束采样和图之间的测地线插值。
摘要：Energy-based models for discrete domains, such as graphs, explicitly capture relative likelihoods, naturally enabling composable probabilistic inference tasks like conditional generation or enforcing constraints at test-time. However, discrete energy-based models typically struggle with efficient and high-quality sampling, as off-support regions often contain spurious local minima, trapping samplers and causing training instabilities. This has historically resulted in a fidelity gap relative to discrete diffusion models. We introduce Graph Energy Matching (GEM), a generative framework for graphs that closes this fidelity gap. Motivated by the transport map optimization perspective of the Jordan-Kinderlehrer-Otto (JKO) scheme, GEM learns a permutation-invariant potential energy that simultaneously provides transport-aligned guidance from noise toward data and refines samples within regions of high data likelihood. Further, we introduce a sampling protocol that leverages an energy-based switch to seamlessly bridge: (i) rapid, gradient-guided transport toward high-probability regions to (ii) a mixing regime for exploration of the learned graph distribution. On molecular graph benchmarks, GEM matches or exceeds strong discrete diffusion baselines. Beyond sample quality, explicit modeling of relative likelihood enables targeted exploration at inference time, facilitating compositional generation, property-constrained sampling, and geodesic interpolation between graphs.
【2】A One-Inclusion Graph Approach to Multi-Group Learning
标题：多群体学习的一包含图方法
链接：https://arxiv.org/abs/2603.23208
摘要：我们证明了多组学习的样本复杂度的最严格的已知上界。我们的算法扩展了一个包含图预测策略，使用二分$b$-匹配的推广。在组可实现的设置，我们提供了一个下界确认我们的算法的$\log n / n$的收敛速度是最佳的一般。如果一个放松的学习目标，使我们被评估的组被选择的样本obligatory，那么我们的算法达到最佳的1/n$收敛速度下组的可实现性。
摘要：We prove the tightest-known upper bounds on the sample complexity of multi-group learning. Our algorithm extends the one-inclusion graph prediction strategy using a generalization of bipartite $b$-matching. In the group-realizable setting, we provide a lower bound confirming that our algorithm's $\log n / n$ convergence rate is optimal in general. If one relaxes the learning objective such that the group on which we are evaluated is chosen obliviously of the sample, then our algorithm achieves the optimal $1/n$ convergence rate under group-realizability.
【3】HGNet: Scalable Foundation Model for Automated Knowledge Graph Generation from Scientific Literature
标题：HGNet：从科学文献自动生成知识图的可扩展基础模型
链接：https://arxiv.org/abs/2603.23136
摘要：自动化知识图（KG）构建对于导航快速增长的科学文献至关重要。然而，现有的方法很难识别长的多词实体，往往无法跨域概括，通常忽视了科学知识的层次性。虽然通用大型语言模型（LLM）提供了适应性，但它们在计算上是昂贵的，并且在专门的任务上产生不一致的准确性。因此，目前的KGS是浅的和不一致的，限制了它们的勘探和合成的效用。我们提出了一个两阶段的框架，可扩展的，zero-shot科学KG建设。第一阶段，Z-NERD，介绍（i）正交语义分解（OSD），它通过隔离文本中的语义“转折”来促进域无关实体识别，以及（ii）多尺度TCQK注意机制，通过n-gram感知注意头捕获连贯的多词实体。第二阶段，HGNet，执行关系提取与层次意识的消息传递，明确建模父，子和对等关系。为了加强全局一致性，我们引入了两个互补的目标：一个可微层次损失，以阻止循环和捷径边缘，和一个连续体抽象场（CAF）损失，嵌入抽象水平沿一个可学习的轴在欧几里得空间。这是第一种将层次抽象形式化为标准欧几里得嵌入中的连续属性的方法，为双曲方法提供了一种更简单的替代方法。我们发布了SPHERE（https：//github.com/basilalab/SPHERE），一个用于层次关系提取的多域基准测试。我们的框架在SciERC，SciER和SPHERE上建立了一个新的最先进的状态，在分布外测试中将NER提高了8.08%，RE提高了5.99%。在zero-shot设置中，NER和RE的增益分别达到10.76%和26.2%。
摘要：Automated knowledge graph (KG) construction is essential for navigating the rapidly expanding body of scientific literature. However, existing approaches struggle to recognize long multi-word entities, often fail to generalize across domains, and typically overlook the hierarchical nature of scientific knowledge. While general-purpose large language models (LLMs) offer adaptability, they are computationally expensive and yield inconsistent accuracy on specialized tasks. As a result, current KGs are shallow and inconsistent, limiting their utility for exploration and synthesis. We propose a two-stage framework for scalable, zero-shot scientific KG construction. The first stage, Z-NERD, introduces (i) Orthogonal Semantic Decomposition (OSD), which promotes domain-agnostic entity recognition by isolating semantic "turns" in text, and (ii) a Multi-Scale TCQK attention mechanism that captures coherent multi-word entities through n-gram-aware attention heads. The second stage, HGNet, performs relation extraction with hierarchy-aware message passing, explicitly modeling parent, child, and peer relations. To enforce global consistency, we introduce two complementary objectives: a Differentiable Hierarchy Loss to discourage cycles and shortcut edges, and a Continuum Abstraction Field (CAF) Loss that embeds abstraction levels along a learnable axis in Euclidean space. This is the first approach to formalize hierarchical abstraction as a continuous property within standard Euclidean embeddings, offering a simpler alternative to hyperbolic methods. We release SPHERE (https://github.com/basiralab/SPHERE), a multi-domain benchmark for hierarchical relation extraction. Our framework establishes a new state of the art on SciERC, SciER, and SPHERE, improving NER by 8.08% and RE by 5.99% on out-of-distribution tests. In zero-shot settings, gains reach 10.76% for NER and 26.2% for RE.
【4】Can Graph Foundation Models Generalize Over Architecture?
标题：图基础模型可以在架构上推广吗？
链接：https://arxiv.org/abs/2603.22984
备注：9 pages main text + 18 pages references and appendix (27 pages total), 5 figures. Accepted to GRaM Workshop @ ICLR 2026: Workshop on Geometry-grounded Representation Learning and Generative Modeling (to appear in PMLR)
摘要：图基础模型（GFM）最近引起了人们的兴趣，这是由于图神经网络（GNN）架构的前景，该架构在任意尺度、特征维度和域的图上推广了zero-shot。虽然现有的工作已经在不同的现实世界基准测试中实证地证明了这种能力，但这些任务都有一个关键的隐藏限制：它们只允许一组狭窄的有效GNN架构。特别是，目前的领域无关的GFM依赖于固定的架构骨干，隐含地假设一个单一的消息传递制度足以跨任务。在本文中，我们认为，架构自适应性是一个必要的要求，真正的GFM。我们表明，现有的方法是不强大的任务相关的建筑属性，并作为一个案例研究，使用范围作为一个最小的和可测量的轴，沿此限制变得明确。通过理论分析和受控合成实验，我们证明了固定骨干GFM可证明在其架构要求与训练时所见不同的任务上存在不足。为了解决这个问题，我们引入了一个框架，通过发现和混合特定于任务的线性图运算符，在推理时适应有效的GNN架构，使zero-shot泛化跨任务异构的架构要求，而无需再训练。我们验证了我们的方法在任意范围的合成任务和一套真实世界的基准测试，表现出改进的性能和鲁棒性比现有的域不可知的GFM。
摘要：Graph foundation models (GFMs) have recently attracted interest due to the promise of graph neural network (GNN) architectures that generalize zero-shot across graphs of arbitrary scales, feature dimensions, and domains. While existing work has demonstrated this ability empirically across diverse real-world benchmarks, these tasks share a crucial hidden limitation: they admit a narrow set of effective GNN architectures. In particular, current domain-agnostic GFMs rely on fixed architectural backbones, implicitly assuming that a single message-passing regime suffices across tasks. In this paper, we argue that architecture adaptivity is a necessary requirement for true GFMs. We show that existing approaches are non-robust to task-dependent architectural attributes and, as a case study, use range as a minimal and measurable axis along which this limitation becomes explicit. With theoretical analysis and controlled synthetic experiments, we demonstrate that fixed-backbone GFMs provably under-reach on tasks whose architectural requirements differ from those seen at training time. To address this issue, we introduce a framework that adapts effective GNN architecture at inference time by discovering and mixing task-specific linear graph operators, enabling zero-shot generalization across tasks with heterogeneous architectural requirements, without retraining. We validate our approach on arbitrary-range synthetic tasks and a suite of real-world benchmarks, demonstrating improved performance and robustness over existing domain-agnostic GFMs.
【5】Universal and efficient graph neural networks with dynamic attention for machine learning interatomic potentials
标题：通用且高效的图神经网络，具有机器学习原子间势的动态关注
链接：https://arxiv.org/abs/2603.22810
备注：10 pages, 6 figures, 6 tables
摘要：分子动力学模拟的核心是原子间相互作用势。传统的经验势缺乏准确性，而第一性原理方法在计算上是禁止的。机器学习原子间相互作用势（MLIPs）承诺以线性成本实现近量子精度，但现有模型在效率和稳定性方面仍面临挑战。我们提出了机器学习先进神经网络（MLANet），一个高效和强大的图神经网络框架。MLANet引入了一种双路径动态注意机制，用于几何感知的消息传递和多视角池化策略，以构建全面的系统表示。这种设计能够对原子环境进行高度精确的建模，同时实现卓越的计算效率，使高保真模拟更容易实现。在跨越不同系统的广泛数据集上进行测试，包括有机分子（例如，QM7、MD17）、周期性无机材料（例如，含Li晶体），二维材料（例如，双层石墨烯，黑磷），表面催化反应（例如，甲酸分解）和带电系统，MLANet保持了具有竞争力的预测精度，同时其计算成本明显低于主流的等变模型，并且能够进行稳定的长时间分子动力学模拟。MLANet为大规模、高精度的原子模拟提供了一个高效实用的工具。
摘要 ：The core of molecular dynamics simulation fundamentally lies in the interatomic potential. Traditional empirical potentials lack accuracy, while first-principles methods are computationally prohibitive. Machine learning interatomic potentials (MLIPs) promise near-quantum accuracy at linear cost, but existing models still face challenges in efficiency and stability. We presents Machine Learning Advances Neural Network (MLANet), an efficient and robust graph neural network framework. MLANet introduces a dual-path dynamic attention mechanism for geometry-aware message passing and a multi-perspective pooling strategy to construct comprehensive system representations. This design enables highly accurate modeling of atomic environments while achieving exceptional computational efficiency, making high-fidelity simulations more accessible. Tested across a wide range of datasets spanning diverse systems, including organic molecules (e.g., QM7, MD17), periodic inorganic materials (e.g., Li-containing crystals), two-dimensional materials (e.g., bilayer graphene, black phosphorus), surface catalytic reactions (e.g., formate decomposition), and charged systems, MLANet maintains competitive prediction accuracy while its computational cost is markedly lower than mainstream equivariant models, and it enables stable long-time molecular dynamics simulations. MLANet provides an efficient and practical tool for large-scale, high-accuracy atomic simulations.
【6】Symbolic Graph Networks for Robust PDE Discovery from Noisy Sparse Data
标题：符号图网络用于从有噪的稀疏数据中发现稳健的DTE
链接：https://arxiv.org/abs/2603.22380
备注：31 pages, 5 figures, 7 tables
摘要：数据驱动的偏微分方程（PDE）的发现提供了一个很有前途的范式，从观测数据揭示管理的物理规律。然而，在实际情况下，测量往往被噪声污染，并受到稀疏采样的限制，这对基于数值微分或积分公式的现有方法构成了重大挑战。在这项工作中，我们提出了一个符号图网络（SGN）框架的PDE发现噪声和稀疏条件下。SGN不依赖于局部差分近似，而是利用图形消息传递来建模空间交互，提供对高频噪声不太敏感的非局部表示。基于这种表示，学习的潜在特征由符号回归模块进一步处理，以提取可解释的数学表达式。我们评估所提出的方法在几个基准系统，包括波动方程，对流扩散方程，不可压缩Navier-Stokes方程。实验结果表明，SGN可以恢复有意义的控制关系或解决方案的形式在不同的噪声水平下，并表现出更好的鲁棒性相比，基线方法在稀疏和噪声设置。这些结果表明，将基于图形的表示与符号回归相结合，为从不完美的观测中发现强大的数据驱动的物理定律提供了一个可行的方向。该代码可从https://github.com/CXY0112/SGN获得
摘要：Data-driven discovery of partial differential equations (PDEs) offers a promising paradigm for uncovering governing physical laws from observational data. However, in practical scenarios, measurements are often contaminated by noise and limited by sparse sampling, which poses significant challenges to existing approaches based on numerical differentiation or integral formulations. In this work, we propose a Symbolic Graph Network (SGN) framework for PDE discovery under noisy and sparse conditions. Instead of relying on local differential approximations, SGN leverages graph message passing to model spatial interactions, providing a non-local representation that is less sensitive to high frequency noise. Based on this representation, the learned latent features are further processed by a symbolic regression module to extract interpretable mathematical expressions. We evaluate the proposed method on several benchmark systems, including the wave equation, convection-diffusion equation, and incompressible Navier-Stokes equations. Experimental results show that SGN can recover meaningful governing relations or solution forms under varying noise levels, and demonstrates improved robustness compared to baseline methods in sparse and noisy settings. These results suggest that combining graph-based representations with symbolic regression provides a viable direction for robust data-driven discovery of physical laws from imperfect observations. The code is available at https://github.com/CXY0112/SGN
【7】Q-AGNN: Quantum-Enhanced Attentive Graph Neural Network for Intrusion Detection
标题：Q-AGNN：用于入侵检测的量子增强型注意图神经网络
链接：https://arxiv.org/abs/2603.22365
摘要：随着互联设备的快速增长，准确检测网络流量中的恶意活动变得越来越具有挑战性。大多数现有的基于深度学习的入侵检测系统将网络流视为独立的实例，从而无法利用网络通信中固有的关系依赖性。为了解决这个问题，我们提出了Q-AGNN，一个用于入侵检测的量子增强注意图神经网络，其中网络流被建模为节点和边缘表示相似性关系。Q-AGNN利用参数化量子电路（PQC）将多跳邻域信息编码到高维潜在空间中，从而产生有界量子特征映射，该映射在量子诱导希尔伯特空间中实现二阶多项式图滤波器。随后应用注意力机制来自适应地对量子增强嵌入进行加权，使模型能够专注于导致异常行为的最有影响力的节点。在四个基准入侵检测数据集上进行的大量实验表明，与最先进的基于图的方法相比，Q-AGNN实现了具有竞争力或优越的检测性能，同时在硬件校准的噪声条件下始终保持低误报率。此外，我们还在实际的IBM量子硬件上执行了Q-AGNN框架，以证明所提出的流水线在真实NISQ条件下的实际可操作性。这些结果突出了将量子增强表示与基于图的入侵检测的注意力机制相结合的有效性，并强调了混合量子经典学习框架在网络安全应用中的潜力。
摘要：With the rapid growth of interconnected devices, accurately detecting malicious activities in network traffic has become increasingly challenging. Most existing deep learning-based intrusion detection systems treat network flows as independent instances, thereby failing to exploit the relational dependencies inherent in network communications. To address this limitation, we propose Q-AGNN, a Quantum-Enhanced Attentive Graph Neural Network for intrusion detection, where network flows are modeled as nodes and edges represent similarity relationships. Q-AGNN leverages parameterized quantum circuits (PQCs) to encode multi-hop neighborhood information into a high-dimensional latent space, inducing a bounded quantum feature map that implements a second-order polynomial graph filter in a quantum-induced Hilbert space. An attention mechanism is subsequently applied to adaptively weight the quantum-enhanced embeddings, allowing the model to focus on the most influential nodes contributing to anomalous behavior. Extensive experiments conducted on four benchmark intrusion detection datasets demonstrate that Q-AGNN achieves competitive or superior detection performance compared to state-of-the-art graph-based methods, while consistently maintaining low false positive rates under hardware-calibrated noise conditions. Moreover, we also executed the Q-AGNN framework on actual IBM quantum hardware to demonstrate the practical operability of the proposed pipeline under real NISQ conditions. These results highlight the effectiveness of integrating quantum-enhanced representations with attention mechanisms for graph-based intrusion detection and underscore the potential of hybrid quantum-classical learning frameworks in cybersecurity applications.
【8】Graph Signal Processing Meets Mamba2: Adaptive Filter Bank via Delta Modulation
标题：图形信号处理满足Mamba 2：通过Delta调制的自适应过滤器组
链接：https://arxiv.org/abs/2603.22333
备注：The Fourteenth International Conference on Learning Representations (ICLR 2026)
摘要：状态空间模型（SSM）提供了有效的替代注意力与线性时间递归。Mamba2是一个最近的基于SSM的语言模型，它使用选择性输入门控和多头结构，实现了并行计算和强大的基准性能。然而，它的多头递归独立运行，没有结构化的利用或分析。在这项工作中，我们提出了一种新的方法，称为高效SSM（HADES）的分层自适应滤波器组，这是一个图形信号处理（GSP）启发的框架，将Mamba2重新解释为线图上的自适应滤波器组。我们的分层架构引入了两种滤波器类型：用于全局低通行为的共享滤波器和用于局部高通行为的专家滤波器，通过对参数Δ的结构化偏置来实现。HADES在语言建模、常识推理和长上下文检索的各种基准测试中实现了与包括Mamba2在内的基线模型相当的性能，同时仅使用了58.9%的原始参数。在这方面，HADES桥接了GSP和神经序列建模，在状态空间模型中实现了高效、分层和可解释的过滤。
摘要：State-space models (SSMs) offer efficient alternatives to attention with linear-time recurrence. Mamba2, a recent SSM-based language model, uses selective input gating and a multi-head structure, enabling parallel computation and strong benchmark performance. However, its multi-head recurrence operates independently without structured utilization or analysis. In this work, we propose a novel method called Hierarchical ADaptive filter bank for Efficient SSMs (HADES), a Graph Signal Processing (GSP)-inspired framework that reinterprets Mamba2 as an adaptive filter bank on a line graph. Our hierarchical architecture introduces two filter types: shared filters for global low-pass behavior and expert filters for local high-pass behavior, achieved through structured bias on the parameter Δ. HADES achieves comparable performance to baseline models including Mamba2 across various benchmarks in language modeling, commonsense reasoning, and long-context retrieval, while using only 58.9% of the original parameters. In this regard, HADES bridges GSP and neural sequence modeling, enabling efficient, hierarchical, and interpretable filtering within state-space models.
【9】Conformal Risk Control for Safety-Critical Wildfire Evacuation Mapping: A Comparative Study of Tabular, Spatial, and Graph-Based Models
标题：安全关键野火疏散地图的保形风险控制：表格、空间和基于图形的模型的比较研究
链接：https://arxiv.org/abs/2603.22331
摘要：今天部署的每一个野火预测模型都有一个共同的危险属性：这些方法都没有提供关于错过多少火势蔓延的正式保证。尽管使用深度学习在野火蔓延预测方面做了大量工作，但之前的研究还没有将无分布安全保证应用于该领域，这使得疏散规划者依赖于概率阈值，而没有正式的保证。据我们所知，我们通过提出共形风险控制（CRC）在野火蔓延预测中的首次应用来解决这一差距，提供了假阴性率（FNR <= 0.05）的有限样本保证。我们揭露了一个明显的失败：在三个复杂度不断增加的模型系列中（表格：LightGBM，AUROC 0.854;卷积：Tiny U-Net，AUROC 0.969;基于图形：Hybrid ResGNN-UNet，AUROC 0.964），标准阈值仅捕获真实火灾蔓延的7-72%。CRC统一消除了这种故障。我们的主要发现是，模型架构决定疏散效率，而CRC决定安全性：两个具有CRC的空间模型都实现了约95%的火灾覆盖率，而标记的像素仅占总像素的约15%，这使得它们的效率比LightGBM高4.2倍，而图形模型在简单的U-Net上的额外复杂性并没有产生有意义的效率增益。我们提出了一个转移感知的三向CRC框架，该框架为操作分诊分配安全/监控/疏散区域，并描述了极端类别不平衡（火灾发生率约为5%）下患病率加权界限的根本限制。所有模型、校准代码和评估管道均已发布，以实现再现性。
摘要：Every wildfire prediction model deployed today shares a dangerous property: none of these methods provides formal guarantees on how much fire spread is missed. Despite extensive work on wildfire spread prediction using deep learning, no prior study has applied distribution-free safety guarantees to this domain, leaving evacuation planners reliant on probability thresholds with no formal assurance. We address this gap by presenting, to our knowledge, the first application of conformal risk control (CRC) to wildfire spread prediction, providing finite-sample guarantees on false negative rate (FNR <= 0.05). We expose a stark failure: across three model families of increasing complexity (tabular: LightGBM, AUROC 0.854; convolutional: Tiny U-Net, AUROC 0.969; and graph-based: Hybrid ResGNN-UNet, AUROC 0.964), standard thresholds capture only 7-72% of true fire spread. CRC eliminates this failure uniformly. Our central finding is that model architecture determines evacuation efficiency, while CRC determines safety: both spatial models with CRC achieve approximately 95% fire coverage while flagging only approximately 15% of total pixels, making them 4.2x more efficient than LightGBM, while the graph model's additional complexity over a simple U-Net yields no meaningful efficiency gain. We propose a shift-aware three-way CRC framework that assigns SAFE/MONITOR/EVACUATE zones for operational triage, and characterize a fundamental limitation of prevalence-weighted bounds under extreme class imbalance (approximately 5% fire prevalence). All models, calibration code, and evaluation pipelines are released for reproducibility.
【10】A graph neural network based chemical mechanism reduction method for combustion applications
标题：基于图神经网络的燃烧化学机制简化方法
链接：https://arxiv.org/abs/2603.22318
摘要：直接数值模拟湍流反应流涉及数百万个网格点和详细的化学机制与数百个物种和数千个反应是计算上禁止的。为了解决这一挑战，我们提出了两个数据驱动的化学机理简化配方的基础上图神经网络（GNNs）与消息传递Transformer层，学习物种和反应之间的非线性依赖关系。第一个公式GNN-SM采用预先训练的代理模型来指导在广泛的反应堆条件下的还原。第二种配方GNN-AE使用自动编码器配方来获得高度紧凑的机制，这些机制在训练期间使用的热化学制度内保持准确。该方法被证明对甲烷（53种，325个反应），乙烯（96种，1054个反应），和异辛烷（1034种，8453个反应）的详细机制。GNN-SM实现了与已建立的基于图形的方法DRGEP相当的减少，同时在广泛的热化学状态下保持准确性。相比之下，GNN-AE的物种和反应减少了高达95%，并且在其目标条件下优于DRGEP。总的来说，拟议的框架提供了一个自动化的，基于机器学习的化学机制减少的途径，可以补充传统的专家指导的分析方法。
摘要：Direct numerical simulations of turbulent reacting flows involving millions of grid points and detailed chemical mechanisms with hundreds of species and thousands of reactions are computationally prohibitive. To address this challenge, we present two data-driven chemical mechanism reduction formulations based on graph neural networks (GNNs) with message-passing transformer layers that learn nonlinear dependencies among species and reactions. The first formulation, GNN-SM, employs a pre-trained surrogate model to guide reduction across a broad range of reactor conditions. The second formulation, GNN-AE, uses an autoencoder formulation to obtain highly compact mechanisms that remain accurate within the thermochemical regimes used during training. The approaches are demonstrated on detailed mechanisms for methane (53 species, 325 reactions), ethylene (96 species, 1054 reactions), and iso-octane (1034 species, 8453 reactions). GNN-SM achieves reductions comparable to the established graph-based method DRGEP while maintaining accuracy across a wide range of thermochemical states. In contrast, GNN-AE achieves up to 95% reduction in species and reactions and outperforms DRGEP within its target conditions. Overall, the proposed framework provides an automated, machine-learning-based pathway for chemical mechanism reduction that can complement traditional expert-guided analytical approaches.
【11】Geometric Mixture-of-Experts with Curvature-Guided Adaptive Routing for Graph Representation Learning
标题：具有曲线引导自适应路由的几何专家混合用于图表示学习
链接：https://arxiv.org/abs/2603.22317
摘要：图结构数据通常表现出复杂的拓扑异构性，使得难以在单个黎曼流形内精确建模。虽然新兴的混合曲率方法试图捕捉这种多样性，但它们通常依赖于缺乏基本几何基础的隐式任务驱动路由。为了应对这一挑战，我们提出了一个几何混合专家框架（GeoMoE），自适应融合节点表示在不同的黎曼空间，以更好地适应多尺度拓扑结构。GeoMoE的核心是利用Ollivier-Ricci Curvature（ORC）作为内在几何，然后协调专业专家的协作。具体来说，我们设计了一个图形感知的门控网络，分配节点特定的融合权重，通过曲率引导的对齐损失进行正则化，以确保可解释和几何一致的路由。此外，我们引入了一个曲率感知的对比目标，通过根据曲率一致性构建正负对来提高几何可辨别性。在六个基准数据集上进行的大量实验表明，GeoMoE在不同图形类型上的性能优于最先进的基线。
摘要：Graph-structured data typically exhibits complex topological heterogeneity, making it difficult to model accurately within a single Riemannian manifold. While emerging mixed-curvature methods attempt to capture such diversity, they often rely on implicit, task-driven routing that lacks fundamental geometric grounding. To address this challenge, we propose a Geometric Mixture-of-Experts framework (GeoMoE) that adaptively fuses node representations across diverse Riemannian spaces to better accommodate multi-scale topological structures. At its core, GeoMoE leverages Ollivier-Ricci Curvature (ORC) as an intrinsic geometric prior to orchestrate the collaboration of specialized experts. Specifically, we design a graph-aware gating network that assigns node-specific fusion weights, regularized by a curvature-guided alignment loss to ensure interpretable and geometry-consistent routing. Additionally, we introduce a curvature-aware contrastive objective that promotes geometric discriminability by constructing positive and negative pairs according to curvature consistency. Extensive experiments on six benchmark datasets demonstrate that GeoMoE outperforms state-of-the-art baselines across diverse graph types.
【12】Contextual Graph Matching with Correlated Gaussian Features
标题：具有相关高斯特征的上下文图匹配
链接：https://arxiv.org/abs/2603.23305
摘要：我们研究了高斯设置中的上下文图匹配，其中边缘权重和节点特征在两个网络中都是相关的。我们推导出精确的信息理论阈值的精确恢复，并确定条件下，几乎精确的恢复是可能的或不可能的，在图形和功能的相关强度，节点的数量，和功能的尺寸。有趣的是，尽管在标准图匹配场景中观察到了全有或全无的相变，但额外的上下文信息引入了更丰富的结构：精确和几乎精确恢复的阈值不再一致。我们的研究结果提供了第一个严格的表征结构和上下文信息如何在图匹配中相互作用，并建立了一个基准设计高效的算法。
摘要 ：We investigate contextual graph matching in the Gaussian setting, where both edge weights and node features are correlated across two networks. We derive precise information-theoretic thresholds for exact recovery, and identify conditions under which almost exact recovery is possible or impossible, in terms of graph and feature correlation strengths, the number of nodes, and feature dimension. Interestingly, whereas an all-or-nothing phase transition is observed in the standard graph-matching scenario, the additional contextual information introduces a richer structure: thresholds for exact and almost exact recovery no longer coincide. Our results provide the first rigorous characterization of how structural and contextual information interact in graph matching, and establish a benchmark for designing efficient algorithms.
Transformer(4篇)
【1】Central Dogma Transformer III: Interpretable AI Across DNA, RNA, and Protein
标题：中心教条Transformer III：跨DNA、RNA和蛋白质的可解释人工智能
链接：https://arxiv.org/abs/2603.23361
备注：20 pages, 8 figures
摘要：生物AI模型越来越多地预测复杂的细胞反应，但它们的学习表征仍然与它们旨在捕获的分子过程脱节。我们提出了CDT-III，它将面向机制的AI扩展到整个中心法则：DNA，RNA和蛋白质。它的两阶段虚拟细胞嵌入器架构反映了细胞的空间划分：VCE-N模型在细胞核中转录，VCE-C模型在细胞质中翻译。在5个保留基因上，CDT-III实现了每个基因RNA的r=0.843和蛋白质的r=0.969。添加蛋白质预测提高了RNA性能（r=0.804至0.843），表明下游任务规范化了上游表示。蛋白质监督提高了DNA水平的可解释性，使CTCF富集增加了30%。应用于计算机模拟的CD 52敲低（近似于Alemtuzumab），该模型正确预测了29/29的蛋白质变化，并重新发现了7种已知临床副作用中的5种，但没有临床数据。基于副作用的分析仅需要未受干扰的基线数据（r=0.939），从而能够在不进行新实验的情况下筛选所有2，361个基因。
摘要：Biological AI models increasingly predict complex cellular responses, yet their learned representations remain disconnected from the molecular processes they aim to capture. We present CDT-III, which extends mechanism-oriented AI across the full central dogma: DNA, RNA, and protein. Its two-stage Virtual Cell Embedder architecture mirrors the spatial compartmentalization of the cell: VCE-N models transcription in the nucleus and VCE-C models translation in the cytosol. On five held-out genes, CDT-III achieves per-gene RNA r=0.843 and protein r=0.969. Adding protein prediction improves RNA performance (r=0.804 to 0.843), demonstrating that downstream tasks regularize upstream representations. Protein supervision sharpens DNA-level interpretability, increasing CTCF enrichment by 30%. Applied to in silico CD52 knockdown approximating Alemtuzumab, the model predicts 29/29 protein changes correctly and rediscovers 5 of 7 known clinical side effects without clinical data. Gradient-based side effect profiling requires only unperturbed baseline data (r=0.939), enabling screening of all 2,361 genes without new experiments.
【2】Transformers Trained via Gradient Descent Can Provably Learn a Class of Teacher Models
标题：通过梯度下降训练的Transformer可以证明学习一类教师模型
链接：https://arxiv.org/abs/2603.22801
备注：64 pages, 9 figures
摘要：Transformers在广泛的应用中取得了巨大的成功，但其成功背后的理论基础仍然在很大程度上未被探索。为了揭开Transformers应用于多用途场景和任务的强大能力的神秘面纱，我们从理论上研究利用Transformers作为学生向一类教师模型学习。具体来说，我们的分析中涵盖的教师模型包括具有平均池化的卷积层、图卷积层和各种经典统计学习模型，包括稀疏令牌选择模型的变体[Sanford等人，2023，Wang等人，2024]和组稀疏线性预测器[Zhang等人，2025年]。当从这类教师模型中学习时，我们证明了具有简化的“仅位置”注意力的单层Transformers可以成功地恢复教师模型的所有参数块，从而实现最优的种群损失。基于对教师模型的训练Transformers的有效模仿，我们进一步证明了它们可以在温和的假设下很好地推广到广泛的分布外数据。在我们的分析中，关键是要确定一个基本的双线性结构共享的各种学习任务，这使我们能够建立统一的学习保证这些任务时，把他们作为教师的Transformers器。
摘要：Transformers have achieved great success across a wide range of applications, yet the theoretical foundations underlying their success remain largely unexplored. To demystify the strong capacities of transformers applied to versatile scenarios and tasks, we theoretically investigate utilizing transformers as students to learn from a class of teacher models. Specifically, the teacher models covered in our analysis include convolution layers with average pooling, graph convolution layers, and various classic statistical learning models, including a variant of sparse token selection models [Sanford et al., 2023, Wang et al., 2024] and group-sparse linear predictors [Zhang et al., 2025]. When learning from this class of teacher models, we prove that one-layer transformers with simplified "position-only'' attention can successfully recover all parameter blocks of the teacher models, thus achieving the optimal population loss. Building upon the efficient mimicry of trained transformers towards teacher models, we further demonstrate that they can generalize well to a broad class of out-of-distribution data under mild assumptions. The key in our analysis is to identify a fundamental bilinear structure shared by various learning tasks, which enables us to establish unified learning guarantees for these tasks when treating them as teachers for transformers.
【3】Emergency Preemption Without Online Exploration: A Decision Transformer Approach
标题：无需在线探索的紧急先发制人：决策Transformer方法
链接：https://arxiv.org/abs/2603.22315
摘要：应急车辆（EV）响应时间是生存结果的关键决定因素，但部署的信号抢占策略仍然是反应性和不可控制的。提出了一种基于决策Transformer（DT）的应急通道优化返回条件框架。通过将走廊优化转换为离线的返回条件序列建模，我们的方法（1）消除了策略学习期间的在线环境交互，（2）通过单个目标返回标量实现调度级紧急控制，以及（3）通过多代理决策Transformer（MADT）扩展到多代理设置，并具有用于空间协调的图形注意力。在LightSim模拟器上，相对于4x 4电网上的固定定时抢占，DT将平均电动汽车行驶时间减少了37.7%（88.6秒与142.3秒），实现了最低的民用延误（11.3秒/辆）和最少的电动汽车停车（1.2）在所有方法中，包括需要环境交互的在线RL基线。MADT在更大的网格上进一步改进，通过图形注意力协调，在8x8上减少了45.2%，超过了DT。返回调节产生平滑的调度界面：将目标返回从100到-400个交易EV旅行时间（72.4-138.2 s）变化到民用延迟（16.8-5.4 s/veh），不需要再培训。受约束的DT扩展增加了明确的民用中断预算作为第二个控制旋钮。
摘要：Emergency vehicle (EV) response time is a critical determinant of survival outcomes, yet deployed signal preemption strategies remain reactive and uncontrollable. We propose a return-conditioned framework for emergency corridor optimization based on the Decision Transformer (DT). By casting corridor optimization as offline, return-conditioned sequence modeling, our approach (1) eliminates online environment interaction during policy learning, (2) enables dispatch-level urgency control through a single target-return scalar, and (3) extends to multi-agent settings via a Multi-Agent Decision Transformer (MADT) with graph attention for spatial coordination. On the LightSim simulator, DT reduces average EV travel time by 37.7% relative to fixed-timing preemption on a 4x4 grid (88.6 s vs. 142.3 s), achieving the lowest civilian delay (11.3 s/veh) and fewest EV stops (1.2) among all methods, including online RL baselines that require environment interaction. MADT further improves on larger grids, overtaking DT with 45.2% reduction on 8x8 via graph-attention coordination. Return conditioning produces a smooth dispatch interface: varying the target return from 100 to -400 trades EV travel time (72.4-138.2 s) against civilian delay (16.8-5.4 s/veh), requiring no retraining. A Constrained DT extension adds explicit civilian disruption budgets as a second control knob.
【4】Neutrino Oscillation Parameter Estimation Using Structured Hierarchical Transformers
标题：使用结构化分层Transformer的中微子振荡参数估计
链接
：https://arxiv.org/abs/2603.22342
备注：Paper accepted to appear in the IEEE International Joint Conference on Neural Networks 2026
摘要：中微子振荡编码了关于中微子质量和混合参数的基本信息，提供了一个超越标准模型的独特窗口。然而，从振荡概率图估计这些参数在计算上是具有挑战性的，因为这些图的高维性和对基础物理的非线性依赖性。传统的推理方法，如基于似然或蒙特卡罗抽样方法，需要大量的模拟来探索参数空间，这为大规模分析带来了主要瓶颈。在这项工作中，我们引入了一个数据驱动的框架，重新制定了大气中微子振荡参数推断作为一个监督回归任务结构化的振荡地图。我们提出了一个层次化的Transformer架构，明确地模拟这些地图的二维结构，捕获在固定的能量和全局相关性的能量谱的角度依赖性。为了提高物理一致性，使用替代模拟约束来训练模型，该替代模拟约束强制预测参数与重建的振荡模式之间的一致性。此外，我们引入了一个基于神经网络的不确定性量化机制，该机制产生具有正式覆盖保证的无分布预测区间。在地球物质条件下的模拟振荡图上的实验表明，所提出的方法在估计精度上与马尔可夫链蒙特卡罗基线相当，在计算成本上有实质性的改进（大约240$\times$更少的FLOPs和33$\times$平均处理时间更快）。此外，保形校准的预测区间仍然很窄，同时实现了90%的目标标称覆盖率，证实了我们的方法的可靠性和效率。
摘要：Neutrino oscillations encode fundamental information about neutrino masses and mixing parameters, offering a unique window into physics beyond the Standard Model. Estimating these parameters from oscillation probability maps is, however, computationally challenging due to the maps' high dimensionality and nonlinear dependence on the underlying physics. Traditional inference methods, such as likelihood-based or Monte Carlo sampling approaches, require extensive simulations to explore the parameter space, creating major bottlenecks for large-scale analyses. In this work, we introduce a data-driven framework that reformulates atmospheric neutrino oscillation parameter inference as a supervised regression task over structured oscillation maps. We propose a hierarchical transformer architecture that explicitly models the two-dimensional structure of these maps, capturing angular dependencies at fixed energies and global correlations across the energy spectrum. To improve physical consistency, the model is trained using a surrogate simulation constraint that enforces agreement between the predicted parameters and the reconstructed oscillation patterns. Furthermore, we introduce a neural network-based uncertainty quantification mechanism that produces distribution-free prediction intervals with formal coverage guarantees. Experiments on simulated oscillation maps under Earth-matter conditions demonstrate that the proposed method is comparable to a Markov Chain Monte Carlo baseline in estimation accuracy, with substantial improvements in computational cost (around 240$\times$ fewer FLOPs and 33$\times$ faster in average processing time). Moreover, the conformally calibrated prediction intervals remain narrow while achieving the target nominal coverage of 90%, confirming both the reliability and efficiency of our approach.
GAN|对抗|攻击|生成相关(10篇)
【1】Estimating Flow Velocity and Vehicle Angle-of-Attack from Non-invasive Piezoelectric Structural Measurements Using Deep Learning
标题：使用深度学习根据无创式压电结构测量估计流速和车辆迎角
链接：https://arxiv.org/abs/2603.23496
摘要：气动状态变量的准确估计对于气动载荷预测、飞行控制和模型验证都具有重要意义。这项工作提出了一种非侵入性的方法来估计飞行器的速度和AoA的结构振动测量，而不是直接流量仪表，如皮托管。安装在空气动力学壳体的内部蒙皮上的密集的压电传感器阵列捕获由湍流边界层压力波动引起的振动，并且训练卷积神经网络（CNN）来反转这些结构响应以恢复速度和AoA。 通过在Sandia高超音速风洞中进行的跨越零和非零攻角构型、马赫数~5和马赫数~8条件以及恒定和连续变化风洞操作的受控实验，验证了概念验证。CNN使用来自16个风洞运行的数据进行训练和评估，每个运行内的时间中心保持间隔用于形成训练、验证和测试数据集，并评估运行内的时间泛化。原始CNN预测在连续变化的条件下表现出增加的方差;短窗口移动中值后处理步骤抑制了这种方差并提高了鲁棒性。经过后处理后，该方法获得了相对于低通滤波参考速度的平均速度误差低于2.27 m/s（0.21%），来自相同实验活动的保持测试数据的平均AoA误差为0.44 ^{\circ}（8.25%）$，证明了在受控实验室环境中基于振动的速度和AoA估计的可行性。
摘要：Accurate estimation of aerodynamic state variables such as freestream velocity and angle of attack (AoA) is important for aerodynamic load prediction, flight control, and model validation. This work presents a non-intrusive method for estimating vehicle velocity and AoA from structural vibration measurements rather than direct flow instrumentation such as pitot tubes. A dense array of piezoelectric sensors mounted on the interior skin of an aeroshell capture vibrations induced by turbulent boundary layer pressure fluctuations, and a convolutional neural network (CNN) is trained to invert these structural responses to recover velocity and AoA. Proof-of-concept is demonstrated through controlled experiments in Sandia's hypersonic wind tunnel spanning zero and nonzero AoA configurations, Mach~5 and Mach~8 conditions, and both constant and continuously varying tunnel operations. The CNN is trained and evaluated using data from 16 wind tunnel runs, with a temporally centered held-out interval within each run used to form training, validation, and test datasets and assess intra-run temporal generalization. Raw CNN predictions exhibit increased variance during continuously varying conditions; a short-window moving-median post-processing step suppresses this variance and improves robustness. After post-processing, the method achieves a mean velocity error relative to the low-pass filtered reference velocity below 2.27~m/s (0.21\%) and a mean AoA error of $0.44^{\circ} (8.25\%)$ on held-out test data from the same experimental campaign, demonstrating feasibility of vibration-based velocity and AoA estimation in a controlled laboratory environment.
【2】Permutation-Symmetrized Diffusion for Unconditional Molecular Generation
标题：无条件分子生成的排列对称扩散
链接：https://arxiv.org/abs/2603.23255
摘要：置换不变性是分子点云生成的基础，然而大多数扩散模型通过有序空间上的置换等变网络间接地实施它。我们建议直接在商流形$\tilde{\calX}=\sR^{d\times N}/S_N$上模拟扩散，其中所有原子排列都被识别。我们表明，热内核$\tilde{\calX}$承认一个明确的表达作为一个欧几里德热内核的总和排列，这澄清了如何扩散的商不同于有序粒子扩散。训练需要一个置换对称的分数，涉及一个棘手的总和超过$S_N$，我们推导出一个期望形式的后验置换和近似使用MCMC在置换空间。我们评估了在EQGAT-Diff协议下在QM 9上的无条件3D分子生成，使用SemlaFlow风格的骨架并连续处理所有变量。结果表明，基于矩阵的置换对称化是可行的，并产生具有竞争力的生成质量，提高了效率。
摘要：Permutation invariance is fundamental in molecular point-cloud generation, yet most diffusion models enforce it indirectly via permutation-equivariant networks on an ordered space. We propose to model diffusion directly on the quotient manifold $\tilde{\calX}=\sR^{d\times N}/S_N$, where all atom permutations are identified. We show that the heat kernel on $\tilde{\calX}$ admits an explicit expression as a sum of Euclidean heat kernels over permutations, which clarifies how diffusion on the quotient differs from ordered-particle diffusion. Training requires a permutation-symmetrized score involving an intractable sum over $S_N$; we derive an expectation form over a posterior on permutations and approximate it using MCMC in permutation space. We evaluate on unconditional 3D molecule generation on QM9 under the EQGAT-Diff protocol, using SemlaFlow-style backbone and treating all variables continuously. The results demonstrate that quotient-based permutation symmetrization is practical and yields competitive generation quality with improved efficiency.
【3】A Learning Method with Gap-Aware Generation for Heterogeneous DAG Scheduling
标题：一种用于异类DAQ调度的具有间隙感知生成的学习方法
链接：https://arxiv.org/abs/2603.23249
备注：30pages, 8 figures
摘要：异构环境中有向无环图的高效调度是一个挑战，由于资源容量和依赖性。在实践中，需要跨具有不同资源池和任务类型的环境进行适应性，以及快速生成计划，使这些挑战变得复杂。我们提出了WeCAN，一个端到端的强化学习框架异构DAG调度，解决任务池的兼容性系数和代诱导的最优性差距。它采用了两阶段单遍设计：一个单一的向前传递产生任务池分数和全局参数，其次是一个生成图，构建没有重复的网络调用的时间表。它的加权交叉注意编码器模型的任务-池的相互作用由兼容性系数门控，是大小不可知的环境波动。此外，广泛使用的列表调度映射可能会因受限的可达性而导致生成诱导的最优性缺口。我们引入了一个序空间分析，其特征在于通过可行的时间表订单的可达集的生成地图，解释了背后的机制产生的差距，并产生了消除差距的充分条件。在这些条件的指导下，我们设计了一个跳跃扩展实现与解析参数化递减跳跃规则，扩大可达的顺序集，同时保持单遍效率。在计算图和真实世界的TPC-H DAG上的实验表明，在强基线上提高了完工时间，推理时间与经典的推理机相当，比多轮神经推理机更快。
摘要：Efficient scheduling of directed acyclic graphs (DAGs) in heterogeneous environments is challenging due to resource capacities and dependencies. In practice, the need for adaptability across environments with varying resource pools and task types, alongside rapid schedule generation, complicates these challenges. We propose WeCAN, an end-to-end reinforcement learning framework for heterogeneous DAG scheduling that addresses task--pool compatibility coefficients and generation-induced optimality gaps. It adopts a two-stage single-pass design: a single forward pass produces task--pool scores and global parameters, followed by a generation map that constructs schedules without repeated network calls. Its weighted cross-attention encoder models task--pool interactions gated by compatibility coefficients, and is size-agnostic to environment fluctuations. Moreover, widely used list-scheduling maps can incur generation-induced optimality gaps from restricted reachability. We introduce an order-space analysis that characterizes the reachable set of generation maps via feasible schedule orders, explains the mechanism behind generation-induced gaps, and yields sufficient conditions for gap elimination. Guided by these conditions, we design a skip-extended realization with an analytically parameterized decreasing skip rule, which enlarges the reachable order set while preserving single-pass efficiency. Experiments on computation graphs and real-world TPC-H DAGs demonstrate improved makespan over strong baselines, with inference time comparable to classical heuristics and faster than multi-round neural schedulers.
【4】A Sobering Look at Tabular Data Generation via Probabilistic Circuits
标题：清醒地审视通过概率电路生成表格数据
链接：https://arxiv.org/abs/2603.23016
摘要：表格数据的生成比文本和图像更具挑战性，因为它具有异构的特征和更小的样本量。在这项任务中，基于扩散的模型是当前最先进的（SotA）模型类，在常用的基准测试中实现了几乎完美的性能。在本文中，我们质疑的看法，表格数据生成的进展。首先，我们强调了当前协议的局限性，以评估生成的数据的保真度，并提倡替代方案。接下来，我们重新审视一个简单的基线-以深度概率电路（PC）形式的分层混合模型-它以一小部分成本提供了与SotA模型竞争或优越的性能。PC是决策森林的生成对应物，因此可以原生地处理异构数据，并提供易于处理的概率生成和推理。最后，在一个严格的实证分析，我们表明，明显饱和的进展SotA模型在很大程度上是由于使用不充分的指标。因此，我们强调，仍然有很多工作要做，以产生现实的表格数据。代码可在https://github.com/april-tools/tabpc上获得。
摘要：Tabular data is more challenging to generate than text and images, due to its heterogeneous features and much lower sample sizes. On this task, diffusion-based models are the current state-of-the-art (SotA) model class, achieving almost perfect performance on commonly used benchmarks. In this paper, we question the perception of progress for tabular data generation. First, we highlight the limitations of current protocols to evaluate the fidelity of generated data, and advocate for alternative ones. Next, we revisit a simple baseline -- hierarchical mixture models in the form of deep probabilistic circuits (PCs) -- which delivers competitive or superior performance to SotA models for a fraction of the cost. PCs are the generative counterpart of decision forests, and as such can natively handle heterogeneous data as well as deliver tractable probabilistic generation and inference. Finally, in a rigorous empirical analysis we show that the apparent saturation of progress for SotA models is largely due to the use of inadequate metrics. As such, we highlight that there is still much to be done to generate realistic tabular data. Code available at https://github.com/april-tools/tabpc.
【5】A Critical Review on the Effectiveness and Privacy Threats of Membership Inference Attacks
标题：对会员推断攻击的有效性和隐私威胁的批判性评论
链接：https://arxiv.org/abs/2603.22987
备注：To appear in ESORICS 2026
摘要：隶属推理攻击（MIA）旨在确定数据样本是否包含在机器学习（ML）模型的训练集中，并已成为衡量ML中隐私泄露的事实标准。我们提出了一个评估框架，定义的条件下，MIA构成一个真正的隐私威胁，并审查代表MIA对it.We发现，在我们的框架中定义的现实条件下，MIA代表弱隐私威胁。因此，依赖它们作为ML中的隐私度量可能会导致高估风险，并由于采用太强的防御措施而导致模型效用的不必要牺牲。
摘要：Membership inference attacks (MIAs) aim to determine whether a data sample was included in a machine learning (ML) model's training set and have become the de facto standard for measuring privacy leakages in ML. We propose an evaluation framework that defines the conditions under which MIAs constitute a genuine privacy threat, and review representative MIAs against it. We find that, under the realistic conditions defined in our framework, MIAs represent weak privacy threats. Thus, relying on them as a privacy metric in ML can lead to an overestimation of risk and to unnecessary sacrifices in model utility as a consequence of employing too strong defenses.
【6】Non-Adversarial Imitation Learning Provably Free of Compounding Errors: The Role of Bellman Constraints
标题：非对抗性模仿学习可以证明没有复合错误：贝尔曼约束的作用
链接：https://arxiv.org/abs/2603.22713
摘要：对抗性模仿学习（AIL）通过减少行为克隆（BC）中的复合错误来实现高质量的模仿，但由于对抗性优化，经常表现出训练不稳定性。为了避免这个问题，以IQ-Learn为代表的一类非对抗性基于Q的模仿学习（IL）方法已经出现，并且被广泛认为通过利用在线环境交互来超越BC。然而，本文重新审视了IQ-Learn，并证明了它可证明地减少到BC，并且具有对水平二次依赖的模仿差距下限，因此仍然存在复合错误。理论分析表明，尽管使用在线交互，但IQ-Learn一致地抑制了示威所发现的状态的所有动作的Q值，从而无法推广。为了解决这个问题，我们引入了一个原始-对偶框架的分布匹配，产生一个新的基于Q的IL方法，双Q-DM。Dual Q-DM的关键机制是引入Bellman约束，将高Q值从访问状态传播到未访问状态，从而实现超越演示的推广。我们证明，双Q-DM是等同于AIL，可以恢复专家的行动超出示范，从而减轻复合错误。据我们所知，Dual Q-DM是第一个理论上保证消除复合误差的非对抗IL方法。实验结果进一步证实了我们的理论结果。
摘要 ：Adversarial imitation learning (AIL) achieves high-quality imitation by mitigating compounding errors in behavioral cloning (BC), but often exhibits training instability due to adversarial optimization. To avoid this issue, a class of non-adversarial Q-based imitation learning (IL) methods, represented by IQ-Learn, has emerged and is widely believed to outperform BC by leveraging online environment interactions. However, this paper revisits IQ-Learn and demonstrates that it provably reduces to BC and suffers from an imitation gap lower bound with quadratic dependence on horizon, therefore still suffering from compounding errors. Theoretical analysis reveals that, despite using online interactions, IQ-Learn uniformly suppresses the Q-values for all actions on states uncovered by demonstrations, thereby failing to generalize. To address this limitation, we introduce a primal-dual framework for distribution matching, yielding a new Q-based IL method, Dual Q-DM. The key mechanism in Dual Q-DM is incorporating Bellman constraints to propagate high Q-values from visited states to unvisited ones, thereby achieving generalization beyond demonstrations. We prove that Dual Q-DM is equivalent to AIL and can recover expert actions beyond demonstrations, thereby mitigating compounding errors. To the best of our knowledge, Dual Q-DM is the first non-adversarial IL method that is theoretically guaranteed to eliminate compounding errors. Experimental results further corroborate our theoretical results.
【7】Precision-Varying Prediction (PVP): Robustifying ASR systems against adversarial attacks
标题：精确变化预测（VP）：增强ASB系统抵御对抗攻击
链接：https://arxiv.org/abs/2603.22590
摘要：随着自动化和代理系统的部署越来越多，确保自动语音识别（ASR）模型的对抗鲁棒性变得至关重要。我们观察到，在推理过程中改变ASR模型的精度可以降低对抗性攻击成功的可能性。我们利用这一事实，通过在预测期间对精度进行简单随机抽样，使模型更加鲁棒。此外，通过比较不同精度产生的输出并利用简单的高斯分类器，可以将洞察力转化为对抗性示例检测策略。实验分析表明，在各种ASR模型和攻击类型的鲁棒性和竞争力的检测性能显着增加。
摘要：With the increasing deployment of automated and agentic systems, ensuring the adversarial robustness of automatic speech recognition (ASR) models has become critical. We observe that changing the precision of an ASR model during inference reduces the likelihood of adversarial attacks succeeding. We take advantage of this fact to make the models more robust by simple random sampling of the precision during prediction. Moreover, the insight can be turned into an adversarial example detection strategy by comparing outputs resulting from different precisions and leveraging a simple Gaussian classifier. An experimental analysis demonstrates a significant increase in robustness and competitive detection performance for various ASR models and attack types.
【8】Adversarial Vulnerabilities in Neural Operator Digital Twins: Gradient-Free Attacks on Nuclear Thermal-Hydraulic Surrogates
标题：神经操作员数字双胞胎中的对抗漏洞：对核热力水力代理的无攻击
链接：https://arxiv.org/abs/2603.22525
备注：56 pages, 14 figures, 22 tables
摘要：操作员学习模型正在迅速成为核和能源系统数字孪生的预测核心，有望从稀疏传感器测量中进行实时现场重建。然而，它们对对抗性扰动的鲁棒性仍然没有得到表征，这是在安全关键系统中部署的关键差距。在这里，我们表明，神经运营商是非常脆弱的极其稀疏（小于1%的输入），物理上合理的扰动，利用其对边界条件的敏感性。在四个运营商架构使用无梯度差分进化，我们证明，最小的修改触发灾难性的预测失败，增加相对L_2 $错误从1.5%（验证的准确性）到37-63%，同时保持完全检测不到的标准验证指标。值得注意的是，100%成功的单点攻击都通过了z分数异常检测。我们引入了有效扰动维数d_{eff}$，这是一个基于雅可比的诊断，与敏感性大小一起，产生了一个双因素脆弱性模型，解释了为什么具有极端敏感性集中的架构（POD-DeepONet，$d_{\text{eff}} \approx 1$）不一定是最可利用的，因为低秩输出预测限制了最大误差，而具有足够扩增的中等浓度（S-DeepONet，$d_{\text{eff}} \approx 4$）产生最高的攻击成功率。在具有梯度病理的架构上，无干扰搜索的性能优于基于梯度的替代方案（PGD），而同等大小的随机扰动的成功率接近于零，这证实了所发现的漏洞是结构性的。我们的研究结果揭示了以前被忽视的攻击面在运营商学习模型，并建立这些模型需要超出标准验证部署前的鲁棒性保证。
摘要：Operator learning models are rapidly emerging as the predictive core of digital twins for nuclear and energy systems, promising real-time field reconstruction from sparse sensor measurements. Yet their robustness to adversarial perturbations remains uncharacterized, a critical gap for deployment in safety-critical systems. Here we show that neural operators are acutely vulnerable to extremely sparse (fewer than 1% of inputs), physically plausible perturbations that exploit their sensitivity to boundary conditions. Using gradient-free differential evolution across four operator architectures, we demonstrate that minimal modifications trigger catastrophic prediction failures, increasing relative $L_2$ error from $\sim$1.5% (validated accuracy) to 37-63% while remaining completely undetectable by standard validation metrics. Notably, 100% of successful single-point attacks pass z-score anomaly detection. We introduce the effective perturbation dimension $d_{\text{eff}}$, a Jacobian-based diagnostic that, together with sensitivity magnitude, yields a two-factor vulnerability model explaining why architectures with extreme sensitivity concentration (POD-DeepONet, $d_{\text{eff}} \approx 1$) are not necessarily the most exploitable, since low-rank output projections cap maximum error, while moderate concentration with sufficient amplification (S-DeepONet, $d_{\text{eff}} \approx 4$) produces the highest attack success. Gradient-free search outperforms gradient-based alternatives (PGD) on architectures with gradient pathologies, while random perturbations of equal magnitude achieve near-zero success rates, confirming that the discovered vulnerabilities are structural. Our findings expose a previously overlooked attack surface in operator learning models and establish that these models require robustness guarantees beyond standard validation before deployment.
【9】Efficient Embedding-based Synthetic Data Generation for Complex Reasoning Tasks
标题：针对复杂推理任务的高效基于嵌入的合成数据生成
链接：https://arxiv.org/abs/2603.22294
摘要：综合数据生成（SDG），利用大型语言模型（LLM），最近已被公认并广泛采用，作为一种有效的方法，以提高性能的更小，但更多的资源和计算效率的LLM通过微调。SDG的一个关键挑战是确保生成数据的质量和多样性。在本文中，我们分析了嵌入空间中生成的数据的多样性和分布，并证明了特定邻域内的示例密度与从该区域提取的示例的预测精度之间的强相关性。基于这一见解，我们提出了一个有针对性的管道，用于基于嵌入的采样，增强了数据的多样性，并在多个基准测试中不断提高性能。
摘要：Synthetic Data Generation (SDG), leveraging Large Language Models (LLMs), has recently been recognized and broadly adopted as an effective approach to improve the performance of smaller but more resource and compute efficient LLMs through fine-tuning. A key challenge in SDG is ensuring the quality and diversity of the generated data. In this paper, we analyze the diversity and distribution of generated data in the embedding space, and demonstrate a strong correlation between the density of examples within a specific neighborhood and the accuracy of predictions on examples drawn from that region. Building on this insight, we present a targeted pipeline for embedding-based sampling that enhances data diversity and consistently improves performance across several benchmarks.
【10】Latent Style-based Quantum Wasserstein GAN for Drug Design
标题：基于潜在风格的量子沃瑟斯坦GAN用于药物设计
链接：https://arxiv.org/abs/2603.22399
备注：Main part: 22 pages, 11 figures, 6 tables. Supplementary material: 16 pages, 15 figures, 14 tables
摘要 ：新药的开发是一个繁琐、耗时和昂贵的过程，平均成本估计高达25亿美元左右。在这个漫长的过程中，第一步是新药的设计，在人工智能的帮助下，从头药物设计近年来蓬勃发展，并彻底改变了该领域。特别是，生成式人工智能在药物发现和开发方面取得了可喜的成果，降低了成本，缩短了解决问题的时间。然而，经典的生成模型，如生成对抗网络（GANs），由于贫瘠的高原而难以训练，并且容易发生模式崩溃。量子计算可能是克服这些问题并提供具有更少参数的模型的一种途径，从而增强了GANs的通用性。我们提出了一种新的基于风格的量子GAN（QGAN）架构用于药物设计，该架构在电路的每个旋转门处实现噪声编码，并在损失函数中引入梯度惩罚以减轻模式崩溃。我们的管道采用变分自动编码器来表示潜在空间中的分子结构，然后将其用作QGAN的输入。我们的基线模型运行在多达15个量子位上，以验证我们在量子模拟器上的架构，并使用五个量子位设置中的156个量子位IBM Heron量子计算机进行推理，以研究使用真实量子硬件对分析的影响。我们基准测试我们的结果对经典模型所提供的MOSES基准套件。
摘要：The development of new drugs is a tedious, time-consuming, and expensive process, for which the average costs are estimated to be up to around $2.5 billion. The first step in this long process is the design of the new drug, for which de novo drug design, assisted by artificial intelligence, has blossomed in recent years and revolutionized the field. In particular, generative artificial intelligence has delivered promising results in drug discovery and development, reducing costs and the time to solution. However, classical generative models, such as generative adversarial networks (GANs), are difficult to train due to barren plateaus and prone to mode collapse. Quantum computing may be an avenue to overcome these issues and provide models with fewer parameters, thereby enhancing the generalizability of GANs. We propose a new style-based quantum GAN (QGAN) architecture for drug design that implements noise encoding at every rotational gate of the circuit and a gradient penalty in the loss function to mitigate mode collapse. Our pipeline employs a variational autoencoder to represent the molecular structure in a latent space, which is then used as input to our QGAN. Our baseline model runs on up to 15 qubits to validate our architecture on quantum simulators, and a 156-qubit IBM Heron quantum computer in the five-qubit setup is used for inference to investigate the effects of using real quantum hardware on the analysis. We benchmark our results against classical models as provided by the MOSES benchmark suite.
半/弱/无/有监督|不确定性|主动学习(5篇)
【1】PhysSkin: Real-Time and Generalizable Physics-Based Animation via Self-Supervised Neural Skinning
标题：PhysSkin：通过自我监督神经剥皮实现实时且可推广的基于物理的动画
链接：https://arxiv.org/abs/2603.23194
备注：Accepted by CVPR 2026. Project Page: https://zju3dv.github.io/PhysSkin/
摘要：实现实时的基于物理的动画，概括不同的3D形状和离散化仍然是一个根本的挑战。我们介绍PhysSkin，一个物理信息框架，解决了这一挑战。本着线性混合蒙皮的精神，我们学习连续蒙皮场作为基函数，将运动子空间坐标提升到全空间变形，子空间由手柄变换定义。为了生成无网格、离散化不可知和物理上一致的蒙皮场，这些蒙皮场可以很好地概括各种3D形状，PhysSkin采用了一种新的神经蒙皮场自动编码器，该编码器由基于变换的编码器和交叉注意解码器组成。此外，我们还开发了一种新的物理信息自监督学习策略，该策略结合了动态皮肤场归一化和冲突感知梯度校正，从而实现了能量最小化，空间平滑性和正交性约束的有效平衡。PhysSkin在泛化神经皮肤方面表现出色，并支持基于物理的实时动画。
摘要：Achieving real-time physics-based animation that generalizes across diverse 3D shapes and discretizations remains a fundamental challenge. We introduce PhysSkin, a physics-informed framework that addresses this challenge. In the spirit of Linear Blend Skinning, we learn continuous skinning fields as basis functions lifting motion subspace coordinates to full-space deformation, with subspace defined by handle transformations. To generate mesh-free, discretization-agnostic, and physically consistent skinning fields that generalize well across diverse 3D shapes, PhysSkin employs a new neural skinning fields autoencoder which consists of a transformer-based encoder and a cross-attention decoder. Furthermore, we also develop a novel physics-informed self-supervised learning strategy that incorporates on-the-fly skinning-field normalization and conflict-aware gradient correction, enabling effective balancing of energy minimization, spatial smoothness, and orthogonality constraints. PhysSkin shows outstanding performance on generalizable neural skinning and enables real-time physics-based animation.
【2】Conformal Cross-Modal Active Learning
标题：共形跨模式主动学习
链接：https://arxiv.org/abs/2603.23159
备注：20 pages, 14 figures
摘要：视觉的基础模型已经通过强大的预训练表示和强大的zero-shot能力改变了视觉识别，但它们在数据高效学习方面的潜力在很大程度上尚未开发。主动学习（AL）的目标是通过策略性地选择最具信息量的样本进行标注来最大限度地降低标注成本，但现有方法在很大程度上忽略了嵌入在现代视觉语言模型（VLM）中的丰富的多模态知识。我们介绍了共形跨模态习得（CCMA），一个新的AL框架，桥梁视觉和语言模态，通过教师和学生的架构。CCMA采用预先训练的VLM作为教师，以提供基于语义的不确定性估计，共形校准以指导仅视觉学生模型的样本选择。通过将多模态共形评分与多样性感知选择策略相结合，CCMA在多个基准测试中实现了卓越的数据效率。我们的方法始终优于最先进的AL基线，与仅依赖不确定性或多样性度量的方法相比具有明显的优势。
摘要：Foundation models for vision have transformed visual recognition with powerful pretrained representations and strong zero-shot capabilities, yet their potential for data-efficient learning remains largely untapped. Active Learning (AL) aims to minimize annotation costs by strategically selecting the most informative samples for labeling, but existing methods largely overlook the rich multimodal knowledge embedded in modern vision-language models (VLMs). We introduce Conformal Cross-Modal Acquisition (CCMA), a novel AL framework that bridges vision and language modalities through a teacher-student architecture. CCMA employs a pretrained VLM as a teacher to provide semantically grounded uncertainty estimates, conformally calibrated to guide sample selection for a vision-only student model. By integrating multimodal conformal scoring with diversity-aware selection strategies, CCMA achieves superior data efficiency across multiple benchmarks. Our approach consistently outperforms state-of-the-art AL baselines, demonstrating clear advantages over methods relying solely on uncertainty or diversity metrics.
【3】Robustness Quantification and Uncertainty Quantification: Comparing Two Methods for Assessing the Reliability of Classifier Predictions
标题：稳健性量化和不确定性量化：比较评估分类器预测可靠性的两种方法
链接：https://arxiv.org/abs/2603.22988
摘要：我们考虑两种方法来评估分类器的个体预测的可靠性：鲁棒性量化（RQ）和不确定性量化（UQ）。我们解释了这两种方法之间的概念差异，比较两种方法在一些基准数据集，并表明RQ是能够优于UQ，无论是在标准设置和分布的变化。除了显示RQ可以与UQ竞争，我们还展示了RQ和UQ的互补性，这两种方法的组合可以导致更好的可靠性评估。
摘要：We consider two approaches for assessing the reliability of the individual predictions of a classifier: Robustness Quantification (RQ) and Uncertainty Quantification (UQ). We explain the conceptual differences between the two approaches, compare both approaches on a number of benchmark datasets and show that RQ is capable of outperforming UQ, both in a standard setting and in the presence of distribution shift. Beside showing that RQ can be competitive with UQ, we also demonstrate the complementarity of RQ and UQ by showing that a combination of both approaches can lead to even better reliability assessments.
【4】Sparsely-Supervised Data Assimilation via Physics-Informed Schrödinger Bridge
标题：基于物理信息薛定谔桥的稀疏监督数据同化
链接：https://arxiv.org/abs/2603.22319
备注：39 pages
摘要：偏微分方程（PDE）系统的数据同化（DA）旨在从稀疏的高保真（HF）观测值重建完整的时空场，同时尊重物理约束。虽然全网格低保真（LF）模拟提供了丰富的先验信息，在多保真度设置，恢复HF场与稀疏观测和管理PDE一致，通常需要每个实例的测试时间优化，这成为一个主要的瓶颈，在时间关键型应用程序。为了缓解这一问题，最近提出了使用生成模型的摊销重建;然而，这种方法依赖于训练期间的全场HF监督，这在现实世界中通常是不切实际的。从一个更现实的角度来看，我们提出了物理信息条件薛定谔桥（PICSB），它传输一个信息的LF前向观察条件HF后，没有任何额外的推理时间的指导。为了在没有HF端点的情况下进行学习，PICSB采用迭代代理端点刷新方案，并直接将PDE残差纳入训练目标，同时在整个采样过程中通过硬调节来强制观察。流体PDE基准上的实验表明，PICSB能够非常快速的时空场重建，同时保持稀疏HF监督下的竞争力的准确性。
摘要：Data assimilation (DA) for systems governed by partial differential equations (PDE) aims to reconstruct full spatiotemporal fields from sparse high-fidelity (HF) observations while respecting physical constraints. While full-grid low-fidelity (LF) simulations provide informative priors in multi-fidelity settings, recovering an HF field consistent with both sparse observations and the governing PDE typically requires per-instance test-time optimization, which becomes a major bottleneck in time-critical applications. To alleviate this, amortized reconstruction using generative models has recently been proposed; however, such approaches rely on full-field HF supervision during training, which is often impractical in real-world settings. From a more realistic perspective, we propose the Physics-Informed Conditional Schrödinger Bridge (PICSB), which transports an informative LF prior toward an observation-conditioned HF posterior without any additional inference-time guidance. To enable learning without HF endpoints, PICSB employs an iterative surrogate-endpoint refresh scheme, and directly incorporates PDE residuals into the training objective while enforcing observations via hard conditioning throughout sampling. Experiments on fluid PDE benchmarks demonstrate that PICSB enables extremely fast spatiotemporal field reconstruction while maintaining competitive accuracy under sparse HF supervision.
【5】REALITrees: Rashomon Ensemble Active Learning for Interpretable Trees
标题：REALITrees：罗生门鼓励可解释树的主动学习
链接：https://arxiv.org/abs/2603.22750
摘要：主动学习通过选择最大化信息增益的样本来降低标记成本。一个占主导地位的框架，按委员会查询（QBC），通常依赖于扰动为基础的多样性，通过随机特征子集或数据盲诱导模型的分歧。虽然这近似于认知不确定性的一个概念，但它牺牲了对合理假设空间的直接表征。我们提出了一种补充方法：罗生门集成主动学习（REAL），它通过穷举所有接近最优模型的罗生门集来构建一个委员会。为了解决功能冗余在这个集合中，我们采用了PAC贝叶斯框架使用吉布斯后加权委员会成员的经验风险。利用最近的算法进步，我们确切地枚举这一组稀疏决策树类。在合成和已建立的主动学习基线中，REAL的性能优于随机集合，特别是在中等噪声环境中，它战略性地利用扩展的模型多样性来实现更快的收敛。
摘要：Active learning reduces labeling costs by selecting samples that maximize information gain. A dominant framework, Query-by-Committee (QBC), typically relies on perturbation-based diversity by inducing model disagreement through random feature subsetting or data blinding. While this approximates one notion of epistemic uncertainty, it sacrifices direct characterization of the plausible hypothesis space. We propose the complementary approach: Rashomon Ensembled Active Learning (REAL) which constructs a committee by exhaustively enumerating the Rashomon Set of all near-optimal models. To address functional redundancy within this set, we adopt a PAC-Bayesian framework using a Gibbs posterior to weight committee members by their empirical risk. Leveraging recent algorithmic advances, we exactly enumerate this set for the class of sparse decision trees. Across synthetic and established active learning baselines, REAL outperforms randomized ensembles, particularly in moderately noisy environments where it strategically leverages expanded model multiplicity to achieve faster convergence.
迁移|Zero/Few/One-Shot|自适应(7篇)
【1】ARGENT: Adaptive Hierarchical Image-Text Representations
标题：ARGENT：自适应分层图像-文本表示
链接：https://arxiv.org/abs/2603.23311
摘要：像CLIP这样的大规模视觉语言模型（VLM）学习强大的语义表示，但在欧几里得空间中操作，无法捕捉视觉和语言概念的固有层次结构。双曲几何，其指数体积增长，提供了一个原则性的替代嵌入这样的层次结构与低失真。然而，现有的双曲VLM使用不稳定的蕴涵损失：当父嵌入向原点收缩时，它们的蕴涵锥向半空间扩展，导致灾难性的锥崩溃，破坏预期的层次结构。此外，这些模型的分层评估仍然不可靠，主要是基于检索和基于相关性的指标，并倾向于分类依赖和模糊的负面影响。为了解决这些限制，我们提出了一个自适应蕴涵损失配对规范正则化，防止锥崩溃没有启发式孔径裁剪。我们进一步介绍了一个基于角度的概率蕴涵协议（PEP），用于评估层次的理解，评分与AUC-ROC和平均精度。本文介绍了一种更强的双曲线VLM基线ARGENT，自适应分层图像-tExt表示。ARGENT在图像分类、文本到图像检索和提出的分层度量上分别将SOTA双曲VLM提高了0.7、1.1和0.8个绝对点。
摘要：Large-scale Vision-Language Models (VLMs) such as CLIP learn powerful semantic representations but operate in Euclidean space, which fails to capture the inherent hierarchical structure of visual and linguistic concepts. Hyperbolic geometry, with its exponential volume growth, offers a principled alternative for embedding such hierarchies with low distortion. However, existing hyperbolic VLMs use entailment losses that are unstable: as parent embeddings contract toward the origin, their entailment cones widen toward a half-space, causing catastrophic cone collapse that destroys the intended hierarchy. Additionally, hierarchical evaluation of these models remains unreliable, being largely retrieval-based and correlation-based metrics and prone to taxonomy dependence and ambiguous negatives. To address these limitations, we propose an adaptive entailment loss paired with a norm regularizer that prevents cone collapse without heuristic aperture clipping. We further introduce an angle-based probabilistic entailment protocol (PEP) for evaluating hierarchical understanding, scored with AUC-ROC and Average Precision. This paper introduces a stronger hyperbolic VLM baseline ARGENT, Adaptive hieRarchical imaGe-tExt represeNTation. ARGENT improves the SOTA hyperbolic VLM by 0.7, 1.1, and 0.8 absolute points on image classification, text-to-image retrieval, and proposed hierarchical metrics, respectively.
【2】Neural ODE and SDE Models for Adaptation and Planning in Model-Based Reinforcement Learning
标题：用于基于模型的强化学习中适应和规划的神经ODE和SYS模型
链接：https://arxiv.org/abs/2603.23245
摘要：我们研究神经常微分方程和随机微分方程（神经常微分方程和随机微分方程），在基于模型的强化学习（RL）框架内对完全和部分观察环境中的随机动力学进行建模。通过一系列的模拟，我们表明，神经SDES更有效地捕捉过渡动态的固有随机性，使高性能的政策，提高样本效率在具有挑战性的情况下。我们利用神经ODE和SDES通过逆模型对环境动态变化进行有效的政策适应，只需要与新环境进行有限的交互。为了解决部分可观测性问题，我们引入了一个潜在的可观测性模型，该模型将ODE与潜在空间中的GAN训练的随机分量相结合。从这个模型中得出的政策提供了一个强大的基线，优于或匹配一般的基于模型和无模型的方法在随机连续控制基准。这项工作证明了行动条件的潜在的SDES RL规划的随机过渡的环境中的适用性。我们的代码可在：https://github.com/ChaoHan-UoS/NeuralRL
摘要 ：We investigate neural ordinary and stochastic differential equations (neural ODEs and SDEs) to model stochastic dynamics in fully and partially observed environments within a model-based reinforcement learning (RL) framework. Through a sequence of simulations, we show that neural SDEs more effectively capture the inherent stochasticity of transition dynamics, enabling high-performing policies with improved sample efficiency in challenging scenarios. We leverage neural ODEs and SDEs for efficient policy adaptation to changes in environment dynamics via inverse models, requiring only limited interactions with the new environment. To address partial observability, we introduce a latent SDE model that combines an ODE with a GAN-trained stochastic component in latent space. Policies derived from this model provide a strong baseline, outperforming or matching general model-based and model-free approaches across stochastic continuous-control benchmarks. This work demonstrates the applicability of action-conditional latent SDEs for RL planning in environments with stochastic transitions. Our code is available at: https://github.com/ChaoHan-UoS/NeuralRL
【3】Dual-Teacher Distillation with Subnetwork Rectification for Black-Box Domain Adaptation
标题：双教师蒸馏和子网络纠正以实现黑匣子域适应
链接：https://arxiv.org/abs/2603.22908
备注：This manuscript is under review at IEEE Transactions on Multimedia
摘要：假设源数据和源模型都不可访问，黑盒域自适应表示高度实用但极具挑战性的设置，因为可传递的信息仅限于黑盒源模型的预测，只能使用目标样本进行查询。现有的方法试图通过伪标签细化或利用外部视觉语言模型（ViLs）来提取可转移的知识，但它们经常受到噪声监督或ViLs提供的语义先验利用不足的影响，这最终会阻碍自适应性能。为了克服这些限制，我们提出了一个双教师蒸馏与子网络整流（DDSR）模型，共同利用嵌入在黑盒源模型和一般的语义信息的ViL的特定知识。DDSR自适应地整合它们的互补预测，为目标域生成可靠的伪标签，并引入子网络驱动的正则化策略，以减轻由噪声监督引起的过拟合。此外，改进的目标预测迭代地增强伪标签和ViL提示，从而实现更准确和语义一致的自适应。最后，目标模型进一步优化，通过自我训练与类的原型。在多个基准数据集上进行的大量实验验证了我们方法的有效性，证明了与最先进的方法（包括使用源数据或模型的方法）相比的一致改进。
摘要：Assuming that neither source data nor the source model is accessible, black box domain adaptation represents a highly practical yet extremely challenging setting, as transferable information is restricted to the predictions of the black box source model, which can only be queried using target samples. Existing approaches attempt to extract transferable knowledge through pseudo label refinement or by leveraging external vision language models (ViLs), but they often suffer from noisy supervision or insufficient utilization of the semantic priors provided by ViLs, which ultimately hinder adaptation performance. To overcome these limitations, we propose a dual teacher distillation with subnetwork rectification (DDSR) model that jointly exploits the specific knowledge embedded in black box source models and the general semantic information of a ViL. DDSR adaptively integrates their complementary predictions to generate reliable pseudo labels for the target domain and introduces a subnetwork driven regularization strategy to mitigate overfitting caused by noisy supervision. Furthermore, the refined target predictions iteratively enhance both the pseudo labels and ViL prompts, enabling more accurate and semantically consistent adaptation. Finally, the target model is further optimized through self training with classwise prototypes. Extensive experiments on multiple benchmark datasets validate the effectiveness of our approach, demonstrating consistent improvements over state of the art methods, including those using source data or models.
【4】Transfer learning via interpolating structures
标题：通过插值结构的迁移学习
链接：https://arxiv.org/abs/2603.22621
备注：preprint submitted to Mechanical Systems and Signal Processing
摘要：尽管最近在基于人群的结构健康监测（PBSHM）方面取得了进展，但高度不同的结构（即，（不同的人群）仍然是一个挑战。目前的工作提出，异构传输可以通过中间结构，桥接感兴趣的结构之间的信息差距来完成。该技术的一个关键方面是通过改变材料特性和几何形状等参数，一种结构可以连续变形为另一种结构。该方法是通过一个案例研究，涉及参数化（和之间的传输）模拟异构桥梁设计（案例1）。然后在案例2中通过一系列有限元模型演示了“桥梁”和“飞机”的简化物理表示之间的转换。一个有趣的问题是“什么时候桥不是飞机？”“以前在预测基于结构相似性的正迁移的背景下被问到。虽然这个问题的明显答案是“总是”，但本文中的结果表明，在某些情况下，确实可以在高度不同的系统之间实现正迁移。
摘要：Despite recent advances in population-based structural health monitoring (PBSHM), knowledge transfer between highly-disparate structures (i.e., heterogeneous populations) remains a challenge. The current work proposes that heterogeneous transfer may be accomplished via intermediate structures that bridge the gap in information between the structures of interest. A key aspect of the technique is the idea that by varying parameters such as material properties and geometry, one structure can be continuously morphed into another. The approach is demonstrated via a case study involving the parameterisation of (and transfer between) simulated heterogeneous bridge designs (Case 1). Transfer between simplified physical representations of a 'bridge' and 'aeroplane' is then demonstrated in Case 2, via a chain of finite-element models. The facetious question 'When is a bridge not an aeroplane?' has been previously asked in the context of predicting positive transfer based on structural similarity. While the obvious answer to this question is 'Always,' the results presented in the current paper show that, in some cases, positive transfer can indeed be achieved between highly-disparate systems.
【5】FAAR: Format-Aware Adaptive Rounding for NVFP4
标题：FAAR：NVFP 4的感知自适应舍入
链接：https://arxiv.org/abs/2603.22370
摘要：在边缘设备上部署大型语言模型（LLM）需要极低的比特量化。NVFP 4等超低精度格式为减少内存占用和加速计算提供了一种有前途的解决方案。然而，现有的量化方法通常依赖于传统的舍入策略，并且不能考虑NVFP 4数值网格的非均匀性，从而导致次优舍入决策和放大的量化误差。为了解决这个问题，我们提出了一种可学习的自适应舍入策略（FAAR），这是一种为NVFP 4格式量身定制的舍入策略。与传统的量化范式不同，FAAR明确地将非均匀NVFP 4网格纳入优化过程。通过自适应调整舍入决策的损失梯度的指导下，我们的方法有效地逼近理论上的最佳量化。为了补充FAAR，我们引入了一个2阶段格式对齐（2FA）微调方案，将LLM参数逐层对齐到NVFP 4数值空间，进一步缩小了性能差距。值得注意的是，这种可学习的优化在Llama 3 -1B上只产生了4个GPU小时的最小训练开销。大量的实验证明了我们的方法的有效性。与Round-to-Nearest（RTN）算法相比，该算法在Llama 3 -1B和Qwen 3 -1.7B上的复杂度分别从14.28和23.06降低到12.60和21.27。此外，我们的方法在各种zero-shot下游任务中始终优于最先进的方法。
摘要：Deploying large language models (LLMs) on edge devices requires extremely low-bit quantization. Ultra-low precision formats such as NVFP4 offer a promising solution for reducing memory footprint and accelerating computation. However, existing quantization methods typically rely on conventional rounding strategies and fail to account for the non-uniformity of the NVFP4 numerical grid, resulting in suboptimal rounding decisions and amplified quantization errors. To address this, we propose Format-Aware Adaptive Rounding (FAAR), a learnable rounding strategy tailored for the NVFP4 format. Unlike conventional quantization paradigms, FAAR explicitly incorporates the non-uniform NVFP4 grid into the optimization process. By adaptively adjusting rounding decisions guided by loss gradients, our method effectively approximates the theoretically optimal quantization. To complement FAAR, we introduce a 2-stages Format Alignment (2FA) fine-tuning scheme that aligns LLM parameters layer-by-layer to the NVFP4 numerical space, further narrowing the performance gap. Remarkably, this learnable optimization incurs a minimal training overhead of only 4 GPU hours on Llama3-1B. Extensive experiments demonstrate the effectiveness of our approach. Compared with Round-to-Nearest (RTN), our method reduces perplexity on WikiText-2 from 14.28 to 12.60 on Llama3-1B and from 23.06 to 21.27 on Qwen3-1.7B. Additionally, our method consistently outperforms state-of-the-art approaches across various zero-shot downstream tasks.
【6】AEGIS: An Operational Infrastructure for Post-Market Governance of Adaptive Medical AI Under US and EU Regulations
标题：AEGIS：根据美国和欧盟法规进行自适应医疗人工智能市场后治理的运营基础设施
链接：https://arxiv.org/abs/2603.22322
摘要：部署在医疗设备中的机器学习系统需要治理框架，以确保安全性，同时实现持续改进。包括FDA和欧盟在内的监管机构已经引入了预定变更控制计划（PCCP）和上市后监督（PMS）等机制，以管理迭代模型更新，而无需重复提交。本文介绍了AI/ML安全评估和治理基础设施（AEGIS），这是一个适用于任何医疗AI系统的治理框架。AEGIS包括三个模块，即，数据集同化和再培训、模型监测和有条件决策，使FDA PCCP和欧盟AI法案第43（4）条的规定得以实施。我们使用独立的PMS报警信号实现了四类部署决策分类（批准、有条件批准、临床审查、审查），从而能够检测到临界状态，即不存在可部署模型，而发布的模型同时处于风险中。 为了说明AEGIS如何在异构临床环境中实例化，我们提供了两个示例：来自电子健康记录的脓毒症预测和来自医学成像的脑肿瘤分割。这两种情况都使用相同的治理架构，不同之处仅在于配置。在脓毒症示例的11次模拟迭代中，AEGIS产生了8次批准、1次有条件批准、1次临床审查和1次转诊决策，执行了所有四个类别。在第8次和第10次迭代中同时发出了警报信号，包括不存在可部署模型且发布的模型同时失败的临界状态。AEGIS在可观察到的性能下降之前检测到漂移。这些结果表明，AEGIS将监管变更控制概念转化为可执行的治理程序，支持在各种临床应用中安全地持续学习自适应医疗AI。
摘要：Machine learning systems deployed in medical devices require governance frameworks that ensure safety while enabling continuous improvement. Regulatory bodies including the FDA and European Union have introduced mechanisms such as the Predetermined Change Control Plan (PCCP) and Post-Market Surveillance (PMS) to manage iterative model updates without repeated submissions. This paper presents AI/ML Evaluation and Governance Infrastructure for Safety (AEGIS), a governance framework applicable to any healthcare AI system. AEGIS comprises three modules, i.e., dataset assimilation and retraining, model monitoring, and conditional decision, that operationalize FDA PCCP and EU AI Act Article 43(4) provisions. We implement a four-category deployment decision taxonomy (APPROVE, CONDITIONAL APPROVAL, CLINICAL REVIEW, REJECT) with an independent PMS ALARM signal, enabling detection of the critical state in which no deployable model exists while the released model is simultaneously at risk. To illustrate how AEGIS can be instantiated across heterogeneous clinical contexts, we provide two examples: sepsis prediction from electronic health records and brain tumor segmentation from medical imaging. Both cases use identical governance architecture, differing only in configuration. Across 11 simulated iterations on the sepsis example, AEGIS yielded 8 APPROVE, 1 CONDITIONAL APPROVAL, 1 CLINICAL REVIEW, and 1 REJECT decision, exercising all four categories. ALARM signals were co-issued at iterations 8 and 10, including the critical state where no deployable model exists and the released model is simultaneously failing. AEGIS detected drift before observable performance degradation. These results demonstrate that AEGIS translates regulatory change-control concepts into executable governance procedures, supporting safe continuous learning for adaptive medical AI across diverse clinical applications.
【7】Research on Individual Trait Clustering and Development Pathway Adaptation Based on the K-means Algorithm
标题：基于K-means算法的个人特质集群与发展路径适应研究
链接：https://arxiv.org/abs/2603.22302
摘要：随着信息技术的发展，人工智能和机器学习在教育领域的应用显示出巨大的潜力。本研究旨在探讨如何利用K-means聚类算法为大学生提供准确的就业指导。现有的方法大多集中在职业生涯路径的预测上，而对不同特征组合的学生在特定职业方向上的适应性研究较少。本研究通过对3000多名大学生的CET-4成绩、GPA、人格特质和学生干部经历的数据分析，运用K-means聚类算法将学生分为四个主要群体。K-means聚类算法通过最小化聚类内的平方误差将具有相似特征的学生分组到一个组中，确保同一聚类内的学生在特征上高度相似，并且最大化不同聚类之间的差异。根据聚类结果，为每个群体提供了有针对性的职业指导建议。研究结果表明，不同特征组合的学生适合不同的职业方向，为个性化的职业指导提供了科学依据，有效提升了学生的就业成功率。未来的研究可以通过扩大样本量、增加特征变量和考虑外部因素等方法进一步提高聚类精度和指导效果。
摘要：With the development of information technology, the application of artificial intelligence and machine learning in the field of education shows great potential. This study aims to explore how to utilize K-means clustering algorithm to provide accurate career guidance for college students. Existing methods mostly focus on the prediction of career paths, but there are fewer studies on the fitness of students with different combinations of characteristics in specific career directions. In this study, we analyze the data of more than 3000 students on their CET-4 scores, GPA, personality traits and student cadre experiences, and use the K-means clustering algorithm to classify the students into four main groups. The K-means clustering algorithm groups students with similar characteristics into one group by minimizing the intra-cluster squared error, ensuring that the students within the same cluster are highly similar in their characteristics, and that differences between different clusters are maximized. Based on the clustering results, targeted career guidance suggestions are provided for each group. The results of the study show that students with different combinations of characteristics are suitable for different career directions, which provides a scientific basis for personalized career guidance and effectively enhances students' employment success rate. Future research can further improve the precision of clustering and the guidance effect by expanding the sample size, increasing the feature variables and considering external factors.
强化学习(5篇)
【1】VLGOR: Visual-Language Knowledge Guided Offline Reinforcement Learning for Generalizable Agents
标题：VLGR：面向可概括代理的视觉语言知识引导的离线强化学习
链接：https://arxiv.org/abs/2603.22892
摘要：将大型语言模型（LLM）与强化学习（RL）相结合，使智能体能够更有效地解释语言指令以执行任务。然而，LLM通常缺乏对物理环境的直接感知，这限制了他们对环境动态的理解以及概括看不见的任务的能力。为了解决这一限制，我们提出了视觉语言知识引导的离线强化学习（VLGOR），这是一个整合视觉和语言知识以生成虚拟展示的框架，从而丰富了交互数据。VLGOR的核心前提是微调视觉语言模型，以预测未来的状态和动作，并以初始视觉观察和高级指令为条件，确保生成的卷展保持时间上的连贯性和空间上的合理性。此外，我们采用反事实提示来为离线RL训练产生更多样化的推出，使代理能够获得有助于遵循语言指令的知识，同时在基于视觉提示的环境中接地。在机器人操作基准上的实验表明，VLGOR显著提高了需要新的最优策略的不可见任务的性能，成功率比基线方法高出24%以上。
摘要：Combining Large Language Models (LLMs) with Reinforcement Learning (RL) enables agents to interpret language instructions more effectively for task execution. However, LLMs typically lack direct perception of the physical environment, which limits their understanding of environmental dynamics and their ability to generalize to unseen tasks. To address this limitation, we propose Visual-Language Knowledge-Guided Offline Reinforcement Learning (VLGOR), a framework that integrates visual and language knowledge to generate imaginary rollouts, thereby enriching the interaction data. The core premise of VLGOR is to fine-tune a vision-language model to predict future states and actions conditioned on an initial visual observation and high-level instructions, ensuring that the generated rollouts remain temporally coherent and spatially plausible. Furthermore, we employ counterfactual prompts to produce more diverse rollouts for offline RL training, enabling the agent to acquire knowledge that facilitates following language instructions while grounding in environments based on visual cues. Experiments on robotic manipulation benchmarks demonstrate that VLGOR significantly improves performance on unseen tasks requiring novel optimal policies, achieving a success rate over 24% higher than the baseline methods.
【2】Model Predictive Control with Differentiable World Models for Offline Reinforcement Learning
标题：用于离线强化学习的具有可区分世界模型的模型预测控制
链接：https://arxiv.org/abs/2603.22430
摘要：离线强化学习（RL）旨在从固定的离线数据集中学习最优策略，而无需与环境进行进一步的交互。这样的方法训练离线策略（或值函数），并且在推理时应用它而无需进一步细化。我们引入了一个受模型预测控制（MPC）启发的推理时间自适应框架，该框架利用了预训练的策略以及状态转换和奖励的学习世界模型。虽然现有的世界模型和扩散规划方法在训练期间使用学习的动态来生成想象的轨迹，或者在推理时间对候选计划进行采样，但它们不使用推理时间信息来动态优化策略参数。相比之下，我们的设计是一个可微世界模型（DWM）管道，它可以通过基于MPC的推理时间策略优化的想象展开来实现端到端的梯度计算。我们在D4RL连续控制基准测试（MuJoCo运动任务和AntMaze）上评估了我们的算法，并表明利用推理时间信息优化策略参数可以在强大的离线RL基线上获得一致的收益。
摘要：Offline Reinforcement Learning (RL) aims to learn optimal policies from fixed offline datasets, without further interactions with the environment. Such methods train an offline policy (or value function), and apply it at inference time without further refinement. We introduce an inference time adaptation framework inspired by model predictive control (MPC) that utilizes a pretrained policy along with a learned world model of state transitions and rewards. While existing world model and diffusion-planning methods use learned dynamics to generate imagined trajectories during training, or to sample candidate plans at inference time, they do not use inference-time information to optimize the policy parameters on the fly. In contrast, our design is a Differentiable World Model (DWM) pipeline that enables endto-end gradient computation through imagined rollouts for policy optimization at inference time based on MPC. We evaluate our algorithm on D4RL continuous-control benchmarks (MuJoCo locomotion tasks and AntMaze), and show that exploiting inference-time information to optimize the policy parameters yields consistent gains over strong offline RL baselines.
【3】Learning When to Act: Interval-Aware Reinforcement Learning with Predictive Temporal Structure
标题：学习何时采取行动：具有预测时间结构的区间感知强化学习
链接：https://arxiv.org/abs/2603.22384
摘要：在连续环境中运行的自治代理不仅要决定做什么，还要决定何时行动。我们介绍了一个轻量级的自适应时间控制系统，学习的最佳间隔之间的认知滴答从经验，取代特设生物启发定时器与原则性的学习政策。政策状态被一个预测性的双曲线价差信号（一种“曲率信号”的简写）增强，该信号来自双曲线几何：嵌入庞加莱球中的n个采样期货之间的平均成对庞加莱距离。高价差表示一个分支，不确定的未来，并驱使代理人更快地采取行动;低价差表示可预测性，并允许更长的休息时间。我们进一步提出了一个区间感知的奖励，明确惩罚相对于所选择的等待时间的低效率，纠正系统的信贷分配失败的天真的基于结果的奖励在时间问题。我们还介绍了一个联合时空嵌入（ATCPG-ST），连接独立的标准化状态和位置投影的庞加莱球;空间轨迹发散提供了一个独立的定时信号不可用的状态只有变体（ATCPG-SO）。这一扩展将平均双曲线价差（kappa）从1.88提高到3.37，并在仅限州的基线上进一步提高了5.8%的效率。五个随机种子的消融实验表明：（i）学习是主要的效率因素（比无学习高54.8%），（ii）双曲线扩展提供了显著的互补增益（比无几何形状控制高26.2%），（iii）组合系统比固定间隔基线高22.8%的效率，以及（iv）将空间位置信息添加到扩展嵌入产生额外的5.8%。
摘要：Autonomous agents operating in continuous environments must decide not only what to do, but when to act. We introduce a lightweight adaptive temporal control system that learns the optimal interval between cognitive ticks from experience, replacing ad hoc biologically inspired timers with a principled learned policy. The policy state is augmented with a predictive hyperbolic spread signal (a "curvature signal" shorthand) derived from hyperbolic geometry: the mean pairwise Poincare distance among n sampled futures embedded in the Poincare ball. High spread indicates a branching, uncertain future and drives the agent to act sooner; low spread signals predictability and permits longer rest intervals. We further propose an interval-aware reward that explicitly penalises inefficiency relative to the chosen wait time, correcting a systematic credit-assignment failure of naive outcome-based rewards in timing problems. We additionally introduce a joint spatio-temporal embedding (ATCPG-ST) that concatenates independently normalised state and position projections in the Poincare ball; spatial trajectory divergence provides an independent timing signal unavailable to the state-only variant (ATCPG-SO). This extension raises mean hyperbolic spread (kappa) from 1.88 to 3.37 and yields a further 5.8 percent efficiency gain over the state-only baseline. Ablation experiments across five random seeds demonstrate that (i) learning is the dominant efficiency factor (54.8 percent over no-learning), (ii) hyperbolic spread provides significant complementary gain (26.2 percent over geometry-free control), (iii) the combined system achieves 22.8 percent efficiency over the fixed-interval baseline, and (iv) adding spatial position information to the spread embedding yields an additional 5.8 percent.
【4】Beyond Hard Constraints: Budget-Conditioned Reachability For Safe Offline Reinforcement Learning
标题：超越硬约束：安全离线强化学习的预算条件可达性
链接：https://arxiv.org/abs/2603.22292
备注：Accepted to the 36th International Conference on Automated Planning and Scheduling (ICAPS 2026)
摘要：使用马尔可夫决策过程的顺序决策支持许多现实世界的应用。基于模型和无模型的方法在这些环境中都取得了很好的效果。然而，现实世界的任务必须平衡奖励最大化与安全约束，这往往是相互冲突的目标，可能导致不稳定的最小/最大，对抗性优化。一个有前途的替代方案是安全可达性分析，它预先计算一个前向不变的安全状态，动作集，确保在这个集合内启动的代理无限期地保持安全。然而，大多数基于可达性的方法只解决硬安全约束，很少有工作扩展到累积成本约束的可达性。为了解决这个问题，首先，我们定义了一个safetyconditioned的可达性集，从累积安全成本的约束条件下的奖励最大化。其次，我们展示了这个集合如何在没有不稳定的最小值/最大值或拉格朗日优化的情况下实施安全约束，从而产生一种新的离线安全RL算法，该算法可以从固定的数据集学习安全策略，而无需环境交互。最后，在标准离线安全RL基准测试和现实世界的海上导航任务上的实验表明，我们的方法在保持安全的同时匹配或优于最先进的基线。
摘要：Sequential decision making using Markov Decision Process underpins many realworld applications. Both model-based and model free methods have achieved strong results in these settings. However, real-world tasks must balance reward maximization with safety constraints, often conflicting objectives, that can lead to unstable min/max, adversarial optimization. A promising alternative is safety reachability analysis, which precomputes a forward-invariant safe state, action set, ensuring that an agent starting inside this set remains safe indefinitely. Yet, most reachability based methods address only hard safety constraints, and little work extends reachability to cumulative cost constraints. To address this, first, we define a safetyconditioned reachability set that decouples reward maximization from cumulative safety cost constraints. Second, we show how this set enforces safety constraints without unstable min/max or Lagrangian optimization, yielding a novel offline safe RL algorithm that learns a safe policy from a fixed dataset without environment interaction. Finally, experiments on standard offline safe RL benchmarks, and a real world maritime navigation task demonstrate that our method matches or outperforms state of the art baselines while maintaining safety.
【5】Privacy-Preserving Reinforcement Learning from Human Feedback via Decoupled Reward Modeling
标题：通过脱钩奖励模型从人类反馈中进行隐私保护强化学习
链接：https://arxiv.org/abs/2603.22563
摘要 ：基于偏好的微调已经成为训练大型语言模型的重要组成部分，在这个阶段使用的数据可能包含敏感的用户信息。一个核心问题是如何设计一个差异化的私有管道，非常适合来自人类反馈的强化学习的独特结构。我们提出了一个隐私保护框架，只对奖励学习施加差分隐私，并从由此产生的私人奖励模型中得出最终的政策。从理论上讲，我们研究了次优差距，并表明隐私贡献了一个额外的附加项超出通常的非私人统计误差。我们还建立了一个极大极小的下限，并表明，占主导地位的长期的样本量和隐私水平的变化，这反过来又表征制度的上限是率最优的对数因子。从经验上讲，合成实验证实了理论预测的缩放，并且使用Gemma-2B-IT模型在Anthropic HH-RLHF数据集上的实验显示出比现有的隐私预算差异隐私基线方法更强的隐私对齐性能。
摘要：Preference-based fine-tuning has become an important component in training large language models, and the data used at this stage may contain sensitive user information. A central question is how to design a differentially private pipeline that is well suited to the distinct structure of reinforcement learning from human feedback. We propose a privacy-preserving framework that imposes differential privacy only on reward learning and derives the final policy from the resulting private reward model. Theoretically, we study the suboptimality gap and show that privacy contributes an additional additive term beyond the usual non-private statistical error. We also establish a minimax lower bound and show that the dominant term changes with sample size and privacy level, which in turn characterizes regimes in which the upper bound is rate-optimal up to logarithmic factors. Empirically, synthetic experiments confirm the scaling predicted by the theory, and experiments on the Anthropic HH-RLHF dataset using the Gemma-2B-IT model show stronger private alignment performance than existing differentially private baseline methods across privacy budgets.
符号|符号学习(2篇)
【1】Weak-PDE-Net: Discovering Open-Form PDEs via Differentiable Symbolic Networks and Weak Formulation
标题：Weak-PDE-Net：通过可微符号网络和弱公式发现开放形式的PDEs
链接：https://arxiv.org/abs/2603.22951
摘要：从稀疏和噪声数据中发现支配偏微分方程是数据驱动科学计算中的一个具有挑战性的问题。传统的稀疏回归方法通常受到两个主要限制：（i）在稀疏和噪声数据下数值微分的不稳定性，以及（ii）预定义候选库的受限灵活性。我们提出了Weak-PDE-Net，这是一个端到端的可微分框架，可以鲁棒地识别开放形式的PDE。Weak-PDE-Net由两个相互连接的模块组成：前向响应学习器和弱形式PDE生成器。学习器将可学习的高斯内核嵌入到轻量级MLP中，作为代理模型，自适应地从稀疏观测中捕获系统动态。同时，该生成器集成了一个符号网络与积分模块来构造弱形式的偏微分方程，避免了显式数值微分，提高了对噪声的鲁棒性。为了放松预定义库的约束，我们在训练过程中利用可区分神经架构搜索策略来探索函数空间，这使得能够有效地发现开放形式的PDE。通过引入伽利略不变性约束和对称等变性假设，保证了系统的物理一致性，进一步增强了弱PDE网络发现多变量系统的能力。在几个具有挑战性的PDE基准测试上的实验表明，即使在高度稀疏和噪声的观测下，弱PDE网络也能准确地恢复控制方程。
摘要：Discovering governing Partial Differential Equations (PDEs) from sparse and noisy data is a challenging issue in data-driven scientific computing. Conventional sparse regression methods often suffer from two major limitations: (i) the instability of numerical differentiation under sparse and noisy data, and (ii) the restricted flexibility of a pre-defined candidate library. We propose Weak-PDE-Net, an end-to-end differentiable framework that can robustly identify open-form PDEs. Weak-PDE-Net consists of two interconnected modules: a forward response learner and a weak-form PDE generator. The learner embeds learnable Gaussian kernels within a lightweight MLP, serving as a surrogate model that adaptively captures system dynamics from sparse observations. Meanwhile, the generator integrates a symbolic network with an integral module to construct weak-form PDEs, avoiding explicit numerical differentiation and improving robustness to noise. To relax the constraints of the pre-defined library, we leverage Differentiable Neural Architecture Search strategy during training to explore the functional space, which enables the efficient discovery of open-form PDEs. The capability of Weak-PDE-Net in multivariable systems discovery is further enhanced by incorporating Galilean Invariance constraints and symmetry equivariance hypotheses to ensure physical consistency. Experiments on several challenging PDE benchmarks demonstrate that Weak-PDE-Net accurately recovers governing equations, even under highly sparse and noisy observations.
【2】Neural Structure Embedding for Symbolic Regression via Continuous Structure Search and Coefficient Optimization
标题：通过连续结构搜索和系数优化实现符号回归的神经结构嵌入
链接：https://arxiv.org/abs/2603.22429
备注：9 pages, 7 figures
摘要：符号回归旨在发现人类可解释的方程来解释观测数据。然而，现有方法严重依赖于离散结构搜索（例如，遗传编程），这通常导致高计算成本、不稳定的性能以及对大方程空间的有限可扩展性。为了解决这些挑战，我们提出了SRCO，一个统一的嵌入驱动的框架，符号回归，符号结构转换成一个连续的，可优化的表示空间。该框架由三个关键组件组成：（1）结构嵌入：我们首先使用传统的符号回归算法生成大量探索方程，并训练Transformer模型，以将符号结构压缩到连续的嵌入空间中;（2）连续结构搜索：嵌入空间使得能够使用基于梯度或基于采样的优化进行有效探索，显著降低了导航组合结构空间的成本;以及（3）系数优化：对于每个发现的结构，我们将符号系数视为可学习的参数，并应用梯度优化来获得准确的数值。在合成数据集和真实数据集上的实验表明，我们的方法在方程精度、鲁棒性和搜索效率方面始终优于最先进的方法。这项工作引入了一个新的范式符号回归的桥梁符号方程发现与连续嵌入学习和优化。
摘要：Symbolic regression aims to discover human-interpretable equations that explain observational data. However, existing approaches rely heavily on discrete structure search (e.g., genetic programming), which often leads to high computational cost, unstable performance, and limited scalability to large equation spaces. To address these challenges, we propose SRCO, a unified embedding-driven framework for symbolic regression that transforms symbolic structures into a continuous, optimizable representation space. The framework consists of three key components: (1) structure embedding: we first generate a large pool of exploratory equations using traditional symbolic regression algorithms and train a Transformer model to compress symbolic structures into a continuous embedding space; (2) continuous structure search: the embedding space enables efficient exploration using gradient-based or sampling-based optimization, significantly reducing the cost of navigating the combinatorial structure space; and (3) coefficient optimization: for each discovered structure, we treat symbolic coefficients as learnable parameters and apply gradient optimization to obtain accurate numerical values. Experiments on synthetic and real-world datasets show that our approach consistently outperforms state-of-the-art methods in equation accuracy, robustness, and search efficiency. This work introduces a new paradigm for symbolic regression by bridging symbolic equation discovery with continuous embedding learning and optimization.
医学相关(4篇)
【1】FixationFormer: Direct Utilization of Expert Gaze Trajectories for Chest X-Ray Classification
标题：FixationFormer：直接利用专家凝视轨迹进行胸部X射线分类
链接：https://arxiv.org/abs/2603.22939
摘要：专家的眼球运动提供了丰富的，被动的放射学领域知识的来源，提供了一个强大的线索，诊断推理集成到计算机辅助分析。然而，直接集成到基于CNN的系统中是具有挑战性的，该系统在历史上主导了医学图像分析领域：凝视记录是连续的，时间密集但空间稀疏，有噪声，并且在专家之间是可变的。因此，大多数现有的基于图像的模型利用简化的表示，例如热图。相比之下，凝视自然与Transformer架构一致，因为两者本质上都是顺序的，并且依赖于注意力来突出相关的输入区域。在这项工作中，我们介绍了FixationFormer，一个基于变换器的架构，表示专家的视线轨迹作为令牌序列，从而保留其时间和空间结构。通过与图像特征一起对凝视序列进行建模，我们的方法解决了凝视数据的稀疏性和可变性，同时通过图像和凝视令牌序列之间的明确交叉关注实现了专家诊断线索的更直接和细粒度的整合。我们在三个公开的基准胸部X射线数据集上评估了我们的方法，并证明它达到了最先进的分类性能，突出了在基于变换的医学图像分析中将凝视表示为序列的价值。
摘要 ：Expert eye movements provide a rich, passive source of domain knowledge in radiology, offering a powerful cue for integrating diagnostic reasoning into computer-aided analysis. However, direct integration into CNN-based systems, which historically have dominated the medical image analysis domain, is challenging: gaze recordings are sequential, temporally dense yet spatially sparse, noisy, and variable across experts. As a consequence, most existing image-based models utilize reduced representations such as heatmaps. In contrast, gaze naturally aligns with transformer architectures, as both are sequential in nature and rely on attention to highlight relevant input regions. In this work, we introduce FixationFormer, a transformer-based architecture that represents expert gaze trajectories as sequences of tokens, thereby preserving their temporal and spatial structure. By modeling gaze sequences jointly with image features, our approach addresses sparsity and variability in gaze data while enabling a more direct and fine-grained integration of expert diagnostic cues through explicit cross-attention between the image and gaze token sequences. We evaluate our method on three publicly available benchmark chest X-ray datasets and demonstrate that it achieves state-of-the-art classification performance, highlighting the value of representing gaze as a sequence in transformer-based medical image analysis.
【2】Vision-based Deep Learning Analysis of Unordered Biomedical Tabular Datasets via Optimal Spatial Cartography
标题：通过最佳空间制图对无序生物医学表格数据集进行基于视觉的深度学习分析
链接：https://arxiv.org/abs/2603.22675
备注：54 Pages, 8 main figures, 26 supplementary figures
摘要：表格数据是生物医学研究的核心，从液体活检和批量和单细胞转录组学到电子健康记录和表型分析。然而，与图像或序列不同，表格数据集缺乏内在的空间组织：特征被视为无序的维度，它们的关系必须由模型隐式推断。这限制了视觉架构利用非空间生物医学数据中的局部结构和高阶特征交互的能力。在这里，我们介绍动态特征映射（Dynomap），这是一个端到端的深度学习框架，可以直接从数据中学习任务优化的特征空间拓扑。Dynomap通过完全可区分的渲染机制联合优化特征放置和预测，而不依赖于几何学、预定义分组或外部先验。通过将高维表格向量转换为学习的特征图，Dynomap使基于视觉的模型能够有效地对无序的生物医学输入进行操作。在多个临床和生物数据集上，Dynomap始终优于经典机器学习、现代深度表格模型和现有的矢量到图像方法。在液体活检数据中，Dynomap将临床相关的基因特征组织成连贯的空间模式，并将多类癌症亚型预测准确性提高了18%。在帕金森病语音数据集中，它聚类了与疾病相关的声学描述符，并将准确率提高了8%。在其他生物医学数据集中观察到类似的收益和可解释的特征组织。这些结果使Dynomap成为一种通用策略，用于桥接表格和基于视觉的深度学习，并用于发现高维生物医学数据中的结构化临床相关模式。
摘要：Tabular data are central to biomedical research, from liquid biopsy and bulk and single-cell transcriptomics to electronic health records and phenotypic profiling. Unlike images or sequences, however, tabular datasets lack intrinsic spatial organization: features are treated as unordered dimensions, and their relationships must be inferred implicitly by the model. This limits the ability of vision architectures to exploit local structure and higher-order feature interactions in non-spatial biomedical data. Here we introduce Dynamic Feature Mapping (Dynomap), an end-to-end deep learning framework that learns a task-optimized spatial topology of features directly from data. Dynomap jointly optimizes feature placement and prediction through a fully differentiable rendering mechanism, without relying on heuristics, predefined groupings, or external priors. By transforming high-dimensional tabular vectors into learned feature maps, Dynomap enables vision-based models to operate effectively on unordered biomedical inputs. Across multiple clinical and biological datasets, Dynomap consistently outperformed classical machine learning, modern deep tabular models, and existing vector-to-image approaches. In liquid biopsy data, Dynomap organized clinically relevant gene signatures into coherent spatial patterns and improved multiclass cancer subtype prediction accuracy by up to 18%. In a Parkinson disease voice dataset, it clustered disease-associated acoustic descriptors and improved accuracy by up to 8%. Similar gains and interpretable feature organization were observed in additional biomedical datasets. These results establish Dynomap as a general strategy for bridging tabular and vision-based deep learning and for uncovering structured, clinically relevant patterns in high-dimensional biomedical data.
【3】Errors in AI-Assisted Retrieval of Medical Literature: A Comparative Study
标题：人工智能辅助医学文献检索中的错误：比较研究
链接：https://arxiv.org/abs/2603.22344
摘要：大型语言模型（LLM）辅助文献检索可能会导致错误的参考文献，但这些错误尚未被严格量化。因此，我们定量评估广泛使用的免费版本LLM平台的参考检索错误，并确定与检索错误相关的因素。我们评估了5个LLM（Grok-2，ChatGPT GPT-4.1，Google Gemini Flash 2.5，Perplexity AI和DeepSeek GPT-4）检索到的2，000篇参考文献，其中40篇随机选择的原始文章（每种期刊10篇）发表于2024年1月至2025年7月，来自英国医学杂志（BMJ），美国医学协会杂志和新英格兰医学杂志（NEJM）。主要结局是结合数字对象标识符、PubMed ID、Google-Scholar链接和相关性的有效性的多指标评分比;以及完全缺失率（未通过所有适用指标的参考文献比例）。多变量回归被用来检查独立的协会。LLM平台在47.8%的时间内完全无法检索到正确的参考数据。5个LLM平台的平均得分比为0.29（标准差为0.35;范围为0-1.25），得分比越高，表明检索相关参考文献和正确书目数据的准确性越高。最高和最低的准确度分别由Grok（0.57）和Genimi（0.11）实现。与《英国医学杂志》相比，《新英格兰医学杂志》的文章得分率较低，完全缺失率较高。多变量分析显示，LLM平台和期刊分别与得分率和完全缺失率独立相关。我们显示出LLM的总体性能适中，并且在平台和期刊之间的检索准确性存在显著差异。LLM平台和期刊与LLM在检索医学文献方面的表现有关。使用法学硕士辅助文献检索时，应仔细审查书目数据。
摘要：Large language models (LLMs) assisted literature retrieval may lead to erroneous references, but these errors have not been rigorously quantified. Therefore, we quantitatively assess errors in reference retrieval of widely used free-version LLM platforms and identify the factors associated with retrieval errors. We evaluated 2,000 references retrieved by 5 LLMs (Grok-2, ChatGPT GPT-4.1, Google Gemini Flash 2.5, Perplexity AI, and DeepSeek GPT-4) for 40 randomly-selected original articles (10 per journal) published Jan. 2024 to July 2025 from British Medical Journal (BMJ), Journal of the American Medical Association, and The New England Journal of Medicine (NEJM). Primary outcomes were a multimetric score ratio combining validity of digital object identifier, PubMed ID, Google-Scholar link, and relevance; and complete miss rate (proportion of references failing all applicable metrics). Multivariable regression was used to examine independent associations. LLM platforms completely failed to retrieve correct reference data 47.8% of the time. The average score ratio of the 5 LLM platforms was 0.29 (standard deviation, 0.35; range, 0-1.25), with a higher score ratio indicating a higher accuracy in retrieving relevant references and correct bibliographic data. The highest and lowest accuracies were achieved by Grok (0.57) and Genimi (0.11), respectively. Compared with BMJ, NEJM articles had lower score ratios and higher complete miss rates. Multivariable analysis shows LLM platforms and journals were independently associated with score ratios and complete miss rate, respectively. We show modest overall performance of LLMs and significant variability in retrieval accuracy across platforms and journals. LLM platforms and journals are associated with LLM's performance in retrieving medical literature. Bibliographic data should be carefully reviewed when using LLM-assisted literature retrieval.
【4】SynLeaF: A Dual-Stage Multimodal Fusion Framework for Synthetic Lethality Prediction Across Pan- and Single-Cancer Contexts
标题：SynLeaF：一个用于跨泛癌症和单一癌症背景合成死亡率预测的双阶段多模式融合框架
链接：https://arxiv.org/abs/2603.22369
备注：29 pages, 5 figures, 3 tables
摘要 ：准确预测综合致死率（SL）对于指导癌症药物和治疗方法的开发具有重要意义。SL预测面临着异构多源数据有效融合的重大挑战。现有的多模态方法往往遭受“模态懒惰”，由于不同的收敛速度，这阻碍了互补信息的开发。这也是为什么大多数现有的SL预测模型不能在泛癌症和单癌症SL对预测上表现良好的一个原因。在这项研究中，我们提出了SynLeaF，一个跨泛和单一癌症背景的SL预测的双阶段多模态融合框架。该框架采用基于VAE的交叉编码器和专家机制的产品来融合四种组学数据类型（基因表达，突变，甲基化和CNV），同时利用关系图卷积网络从生物医学知识图中捕获结构化的基因表示。为了减轻模态懒惰，SynLeaF引入了一个双阶段的训练机制，采用特征级知识蒸馏与自适应单峰教师和合奏策略。在八种特定癌症类型和一个泛癌数据集的广泛实验中，SynLeaF在19种情况中的17种情况下表现出色。烧蚀研究和梯度分析进一步验证了所提出的融合和蒸馏机制对模型鲁棒性和泛化的关键贡献。为了方便社区使用，我们在https://synleaf.bioinformatics-lilab.cn上提供了一个网络服务器。
摘要：Accurate prediction of synthetic lethality (SL) is important for guiding the development of cancer drugs and therapies. SL prediction faces significant challenges in the effective fusion of heterogeneous multi-source data. Existing multimodal methods often suffer from "modality laziness" due to disparate convergence speeds, which hinders the exploitation of complementary information. This is also one reason why most existing SL prediction models cannot perform well on both pan-cancer and single-cancer SL pair prediction. In this study, we propose SynLeaF, a dual-stage multimodal fusion framework for SL prediction across pan- and single-cancer contexts. The framework employs a VAE-based cross-encoder with a product of experts mechanism to fuse four omics data types (gene expression, mutation, methylation, and CNV), while simultaneously utilizing a relational graph convolutional network to capture structured gene representations from biomedical knowledge graphs. To mitigate modality laziness, SynLeaF introduces a dual-stage training mechanism employing featurelevel knowledge distillation with adaptive uni-modal teacher and ensemble strategies. In extensive experiments across eight specific cancer types and a pancancer dataset, SynLeaF achieves superior performance in 17 out of 19 scenarios. Ablation studies and gradient analyses further validate the critical contributions of the proposed fusion and distillation mechanisms to model robustness and generalization. To facilitate community use, a web server is available at https://synleaf.bioinformatics-lilab.cn.
蒸馏|知识提取(2篇)
【1】MemCollab: Cross-Agent Memory Collaboration via Contrastive Trajectory Distillation
标题：MemCollab：通过对比轨迹蒸馏的跨主体记忆协作
链接：https://arxiv.org/abs/2603.23234
摘要：基于大型语言模型（LLM）的代理依赖于记忆机制，以重用过去解决问题的经验知识。现有的方法通常以每个代理的方式构建内存，将存储的知识紧密耦合到单个模型的推理风格。在使用异构代理的现代部署中，一个自然的问题出现了：可以在不同的模型之间共享单个内存系统吗？我们发现，在代理之间天真地转移记忆通常会降低性能，因为这样的记忆会使任务相关的知识与代理特定的偏见纠缠在一起。为了解决这一挑战，我们提出MemCollab，一个协作的内存框架，通过对比不同的代理在同一任务上产生的推理轨迹，构建代理不可知的内存。这种对比过程提取抽象推理约束，捕获共享的任务级不变式，同时抑制代理特定的工件。我们进一步引入了一个任务感知的检索机制，条件的任务类别的内存访问，确保只有相关的约束条件在推理时使用。数学推理和代码生成基准测试的实验表明，MemCollab在不同的代理（包括跨模态系列设置）中始终提高了准确性和推理时间效率。我们的研究结果表明，协作构建的内存可以作为一个共享的推理资源，为不同的LLM为基础的代理。
摘要：Large language model (LLM)-based agents rely on memory mechanisms to reuse knowledge from past problem-solving experiences. Existing approaches typically construct memory in a per-agent manner, tightly coupling stored knowledge to a single model's reasoning style. In modern deployments with heterogeneous agents, a natural question arises: can a single memory system be shared across different models? We found that naively transferring memory between agents often degrades performance, as such memory entangles task-relevant knowledge with agent-specific biases. To address this challenge, we propose MemCollab, a collaborative memory framework that constructs agent-agnostic memory by contrasting reasoning trajectories generated by different agents on the same task. This contrastive process distills abstract reasoning constraints that capture shared task-level invariants while suppressing agent-specific artifacts. We further introduce a task-aware retrieval mechanism that conditions memory access on task category, ensuring that only relevant constraints are used at inference time. Experiments on mathematical reasoning and code generation benchmarks demonstrate that MemCollab consistently improves both accuracy and inference-time efficiency across diverse agents, including cross-modal-family settings. Our results show that the collaboratively constructed memory can function as a shared reasoning resource for diverse LLM-based agents.
【2】Balancing Safety and Efficiency in Aircraft Health Diagnosis: A Task Decomposition Framework with Heterogeneous Long-Micro Scale Cascading and Knowledge Distillation-based Interpretability
标题：飞机健康诊断中的安全性与效率平衡：具有异类长微尺度级联和基于知识蒸馏的可解释性的任务分解框架
链接：https://arxiv.org/abs/2603.22885
备注：Submitted to Reliability Engineering & System Safety (RESS)
摘要：通用航空全机诊断面临三重挑战：数据不确定性、任务异质性和计算效率低下。现有的端到端方法统一模型的健康歧视和故障特征，忽略了内在的感受野全局上下文建模和局部特征提取之间的冲突，同时在严重的类不平衡下产生高昂的训练成本。为了解决这些问题，本研究提出了诊断分解框架（DDF），明确解耦诊断到异常检测（AD）和故障分类（FC）子任务通过长微尺度诊断（LMSD）。LMSD采用“远程全局筛选和微尺度局部精确诊断”的策略，利用具有多头自注意力的卷积标记器（ConvTokMHSA）进行全局操作模式识别，利用多微核网络（MMK Net）进行局部故障特征提取。解耦训练将“大样本轻量级”和“小样本复杂”优化路径分开，显著降低了计算开销。同时，关键提取层（KEL）通过知识蒸馏，为两阶段决策提供了物理上可追溯的解释，实现了设计的可解释性。在NGAFID真实世界航空数据集上的实验表明，多类加权惩罚度量（MCWPM）比基线提高了约4-8%，大大减少了训练时间，验证了任务适应性，可解释性和效率的综合优势。这为通用航空健康管理提供了一种可部署的方法。
摘要：Whole-aircraft diagnosis for general aviation faces threefold challenges: data uncertainty, task heterogeneity, and computational inefficiency. Existing end-to-end approaches uniformly model health discrimination and fault characterization, overlooking intrinsic receptive field conflicts between global context modeling and local feature extraction, while incurring prohibitive training costs under severe class imbalance. To address these, this study proposes the Diagnosis Decomposition Framework (DDF), explicitly decoupling diagnosis into Anomaly Detection (AD) and Fault Classification (FC) subtasks via the Long-Micro Scale Diagnostician (LMSD). Employing a "long-range global screening and micro-scale local precise diagnosis" strategy, LMSD utilizes Convolutional Tokenizer with Multi-Head Self-Attention (ConvTokMHSA) for global operational pattern discrimination and Multi-Micro Kernel Network (MMK Net) for local fault feature extraction. Decoupled training separates "large-sample lightweight" and "small-sample complex" optimization pathways, significantly reducing computational overhead. Concurrently, Keyness Extraction Layer (KEL) via knowledge distillation furnishes physically traceable explanations for two-stage decisions, materializing interpretability-by-design. Experiments on the NGAFID real-world aviation dataset demonstrate approximately 4-8% improvement in Multi-Class Weighted Penalty Metric (MCWPM) over baselines with substantially reduced training time, validating comprehensive advantages in task adaptability, interpretability, and efficiency. This provides a deployable methodology for general aviation health management.
点云|SLAM|雷达|激光|深度RGBD相关(1篇)
【1】Contrastive Metric Learning for Point Cloud Segmentation in Highly Granular Detectors
标题：用于高粒度检测器中点云分割的对比度量学习
链接：https://arxiv.org/abs/2603.23356
摘要 ：我们提出了一种新的基于监督对比度量学习（CML）的点云分割聚类方法。该方法不是预测聚类分配或以对象为中心的变量，而是学习一种潜在的表示，其中属于同一对象的点被嵌入附近，而不相关的点被分离。然后，在学习的度量空间中使用基于密度的读出来重建集群，将表示学习与集群形成解耦，并实现灵活的推理。该方法是评估模拟数据从一个高度颗粒量热计，其中的任务是分离高度重叠的粒子阵雨表示为套量热计命中。使用相同的图神经网络骨干和相等的潜在维度进行与对象凝聚（OC）的直接比较，隔离学习目标的影响。CML方法产生一个更稳定和可分离的嵌入几何的电磁和强子粒子簇射，导致改善当地的邻里一致性，更可靠的分离重叠簇射，更好的推广时，外推到看不见的多重性和能量。这直接转化为更高的重建效率和纯度，特别是在高多重性的制度，以及改善的能量分辨率。在混合粒子的环境中，CML保持强大的性能，这表明强大的学习的淋浴拓扑结构，而OC表现出显着的退化。这些结果表明，基于相似性的表示学习结合基于密度的聚合是一个很有前途的替代对象为中心的方法在高粒度检测器的点云分割。
摘要：We propose a novel clustering approach for point-cloud segmentation based on supervised contrastive metric learning (CML). Rather than predicting cluster assignments or object-centric variables, the method learns a latent representation in which points belonging to the same object are embedded nearby while unrelated points are separated. Clusters are then reconstructed using a density-based readout in the learned metric space, decoupling representation learning from cluster formation and enabling flexible inference. The approach is evaluated on simulated data from a highly granular calorimeter, where the task is to separate highly overlapping particle showers represented as sets of calorimeter hits. A direct comparison with object condensation (OC) is performed using identical graph neural network backbones and equal latent dimensionality, isolating the effect of the learning objective. The CML method produces a more stable and separable embedding geometry for both electromagnetic and hadronic particle showers, leading to improved local neighbourhood consistency, a more reliable separation of overlapping showers, and better generalization when extrapolating to unseen multiplicities and energies. This translates directly into higher reconstruction efficiency and purity, particularly in high-multiplicity regimes, as well as improved energy resolution. In mixed-particle environments, CML maintains strong performance, suggesting robust learning of the shower topology, while OC exhibits significant degradation. These results demonstrate that similarity-based representation learning combined with density-based aggregation is a promising alternative to object-centric approaches for point cloud segmentation in highly granular detectors.
联邦学习|隐私保护|加密(2篇)
【1】Spiking Personalized Federated Learning for Brain-Computer Interface-Enabled Immersive Communication
标题：峰值个性化联邦学习，实现脑机接口的沉浸式通信
链接：https://arxiv.org/abs/2603.22727
备注：6 pages, 3 figures
摘要：这项工作提出了一种新的沉浸式通信框架，该框架利用脑机接口（BCI）来获取大脑信号，以推断以用户为中心的状态（例如，意图和感知相关的不适），从而在强个体差异性下实现更个性化和鲁棒的沉浸式适应。具体来说，我们开发了一个个性化的联邦学习（PFL）模型来分析和处理收集到的大脑信号，它不仅可以容纳神经多样性的大脑信号数据，而且还可以防止敏感的大脑信号信息的泄漏。为了解决能量受限的沉浸式终端（例如，头戴式显示器），我们进一步嵌入尖峰神经网络（SNN）到PFL。通过利用稀疏的、事件驱动的尖峰计算，支持SNN的PFL降低了训练和推理的计算和能量成本，同时保持了有竞争力的个性化性能。在真实脑信号数据集上的实验表明，与传统的基于人工神经网络的个性化基线相比，该方法在降低推理能量6.46倍的同时，实现了最佳的整体识别准确率。
摘要：This work proposes a novel immersive communication framework that leverages brain-computer interface (BCI) to acquire brain signals for inferring user-centric states (e.g., intention and perception-related discomfort), thereby enabling more personalized and robust immersive adaptation under strong individual variability. Specifically, we develop a personalized federated learning (PFL) model to analyze and process the collected brain signals, which not only accommodates neurodiverse brain-signal data but also prevents the leakage of sensitive brain-signal information. To address the energy bottleneck of continual on-device learning and inference on energy-limited immersive terminals (e.g., head-mounted display), we further embed spiking neural networks (SNNs) into the PFL. By exploiting sparse, event-driven spike computation, the SNN-enabled PFL reduces the computation and energy cost of training and inference while maintaining competitive personalization performance. Experiments on real brain-signal dataset demonstrate that our method achieves the best overall identification accuracy while reducing inference energy by 6.46$\times$ compared with conventional artificial neural network-based personalized baselines.
【2】A Theoretical Framework for Energy-Aware Gradient Pruning in Federated Learning
标题：联邦学习中能量感知梯度修剪的理论框架
链接：https://arxiv.org/abs/2603.22465
备注：8 pages, 2 figures. This work has been submitted to the IEEE for possible publication
摘要：联合学习（FL）受到分散边缘设备的通信和能量限制的约束。虽然通过Top-K幅度修剪的梯度稀疏化有效地减少了通信有效载荷，但它本质上仍然是能量不可知的。它假设所有参数更新都会导致相同的下行传输和内存更新成本，而忽略了硬件现实。我们正式的修剪过程作为一个能量约束的投影问题，占内存密集型和计算效率的操作在后反向传播阶段的硬件级的差距。我们提出了成本加权幅度修剪（CWMP），一个选择规则，优先参数更新的基础上，其幅度相对于其物理成本。我们证明了CWMP是最佳的贪婪解决方案，这个约束投影，并提供了一个概率分析其全球能源效率。非IID CIFAR-10基准的数值结果表明，CWMP一贯建立一个优越的性能-能量帕累托前沿相比，Top-K基线。
摘要：Federated Learning (FL) is constrained by the communication and energy limitations of decentralized edge devices. While gradient sparsification via Top-K magnitude pruning effectively reduces the communication payload, it remains inherently energy-agnostic. It assumes all parameter updates incur identical downstream transmission and memory-update costs, ignoring hardware realities. We formalize the pruning process as an energy-constrained projection problem that accounts for the hardware-level disparities between memory-intensive and compute-efficient operations during the post-backpropagation phase. We propose Cost-Weighted Magnitude Pruning (CWMP), a selection rule that prioritizes parameter updates based on their magnitude relative to their physical cost. We demonstrate that CWMP is the optimal greedy solution to this constrained projection and provide a probabilistic analysis of its global energy efficiency. Numerical results on a non-IID CIFAR-10 benchmark show that CWMP consistently establishes a superior performance-energy Pareto frontier compared to the Top-K baseline.
推理|分析|理解|解释(6篇)
【1】Dynamical Systems Theory Behind a Hierarchical Reasoning Model
标题：分层推理模型背后的动态系统理论
链接：https://arxiv.org/abs/2603.22871
摘要：当前的大型语言模型（LLM）主要依赖于线性序列生成和大量参数计数，但它们严重地与复杂的算法推理作斗争。虽然最近的推理架构，如分层推理模型（HRM）和微型递归模型（TRM），证明紧凑的递归网络可以解决这些任务，但它们的训练动态往往缺乏严格的数学保证，导致不稳定和代表性崩溃。我们提出了收缩映射模型（CMM），一种新的架构，重新制定离散递归推理到连续的神经普通和随机微分方程（NODE/NSDES）。通过明确地强制潜在相位点收敛到稳定的平衡状态，并减轻超球面排斥损失的特征崩溃，CMM提供了一个数学上接地和高度稳定的推理引擎。在Sudoku-Extreme基准测试中，5 M参数CMM达到了最先进的93.7%的精度，优于27 M参数HRM（55.0%）和5 M参数TRM（87.4%）。值得注意的是，即使被压缩到只有0.26M参数的超小尺寸，CMM仍然保持了强大的预测能力，在Sudoku-Extreme上达到85.4%，在Maze基准上达到82.2%。这些结果为极端参数效率建立了一个新的前沿，证明了数学上严格的潜在动力学可以有效地取代人工推理中的蛮力缩放。
摘要 ：Current large language models (LLMs) primarily rely on linear sequence generation and massive parameter counts, yet they severely struggle with complex algorithmic reasoning. While recent reasoning architectures, such as the Hierarchical Reasoning Model (HRM) and Tiny Recursive Model (TRM), demonstrate that compact recursive networks can tackle these tasks, their training dynamics often lack rigorous mathematical guarantees, leading to instability and representational collapse. We propose the Contraction Mapping Model (CMM), a novel architecture that reformulates discrete recursive reasoning into continuous Neural Ordinary and Stochastic Differential Equations (NODEs/NSDEs). By explicitly enforcing the convergence of the latent phase point to a stable equilibrium state and mitigating feature collapse with a hyperspherical repulsion loss, the CMM provides a mathematically grounded and highly stable reasoning engine. On the Sudoku-Extreme benchmark, a 5M-parameter CMM achieves a state-of-the-art accuracy of 93.7 %, outperforming the 27M-parameter HRM (55.0 %) and 5M-parameter TRM (87.4 %). Remarkably, even when aggressively compressed to an ultra-tiny footprint of just 0.26M parameters, the CMM retains robust predictive power, achieving 85.4 % on Sudoku-Extreme and 82.2 % on the Maze benchmark. These results establish a new frontier for extreme parameter efficiency, proving that mathematically rigorous latent dynamics can effectively replace brute-force scaling in artificial reasoning.
【2】TorR: Towards Brain-Inspired Task-Oriented Reasoning via Cache-Oriented Algorithm-Architecture Co-design
标题：TorR：通过面向缓存的架构协同设计实现脑启发的面向任务的推理
链接：https://arxiv.org/abs/2603.22855
备注：Accepted to DAC 2026
摘要：CLIP上的面向任务的对象检测（TOOD）提供了开放的词汇表，语义驱动，但密集的每窗口计算和繁重的内存流量阻碍了实时，功率受限的边缘部署。我们提出了一个大脑启发的算法-架构协同设计，它用超维（HDC）关联推理器取代CLIP风格的密集对齐，并将时间一致性转化为重用。在算法方面，TorR将对齐重新定义为HDC相似性和图形组合，通过（i）查询缓存与每个类的分数累积，（ii）当只有一小部分超向量位发生变化时精确的$δ$更新，以及（iii）在高系统负载下的相似性/负载门控旁路，引入部分相似性重用。在架构方面，TorR实例化了一个通道可扩展的位切片项存储器，具有库/精度门控和一个轻量级控制器，该控制器可以在对象计数变化时调度旁路/$δ$/完整路径以满足RT-30/RT-60目标。TorR在TSMC 28 nm工艺中合成，并使用周期精确模拟器进行训练，每个窗口的能量为毫焦耳级，可维持实时吞吐量（60\，FPS时约50\，mJ; 30\，FPS时约为113\，mJ）和低延迟抖动，同时在五个任务提示中提供具有竞争力的AP@0.5（平均44.27%）在一个有界的利润率强VLM基线，但在数量级较低的能量。该设计公开了部署时的可配置性（有效维度$D '$、阈值、精度），以权衡边缘预算的准确性、延迟和能源。
摘要：Task-oriented object detection (TOOD) atop CLIP offers open-vocabulary, prompt-driven semantics, yet dense per-window computation and heavy memory traffic hinder real-time, power-limited edge deployment. We present \emph{TorR}, a brain-inspired \textbf{algorithm--architecture co-design} that \textbf{replaces CLIP-style dense alignment with a hyperdimensional (HDC) associative reasoner} and turns temporal coherence into reuse. On the \emph{algorithm} side, TorR reformulates alignment as HDC similarity and graph composition, introducing \emph{partial-similarity reuse} via (i) query caching with per-class score accumulation, (ii) exact $δ$-updates when only a small set of hypervector bits change, and (iii) similarity/load-gated bypass under high system load. On the \emph{architecture} side, TorR instantiates a lane-scalable, bit-sliced item memory with bank/precision gating and a lightweight controller that schedules bypass/$δ$/full paths to meet RT-30/RT-60 targets as object counts vary. Synthesized in a TSMC 28\,nm process and exercised with a cycle-accurate simulator, TorR sustains real-time throughput with millijoule-scale energy per window ($\approx$50\,mJ at 60\,FPS; $\approx$113\,mJ at 30\,FPS) and low latency jitter, while delivering competitive AP@0.5 across five task prompts (mean 44.27\%) within a bounded margin to strong VLM baselines, but at orders-of-magnitude lower energy. The design exposes deployment-time configurability (effective dimension $D'$, thresholds, precision) to trade accuracy, latency, and energy for edge budgets.
【3】Explainable Threat Attribution for IoT Networks Using Conditional SHAP and Flow Behavior Modelling
标题：使用条件SHAP和流行为建模的物联网网络可解释威胁归因
链接：https://arxiv.org/abs/2603.22771
摘要：随着物联网（IoT）在关键基础设施、智能环境和消费设备中的不断扩展，保护它们免受网络威胁变得越来越重要。传统的入侵检测模型通常将物联网威胁视为二进制分类问题或依赖于不透明模型，从而限制了信任。这项工作使用CICIoT2023数据集研究了物联网环境中的多类威胁归因，将30多种攻击变体分为8个语义上有意义的类别。我们利用梯度提升模型和SHAP（SHapley Additive exPlanations）的组合来提供全局和特定于类的解释，从而详细了解驱动每个攻击分类的特征。结果表明，该模型区分不同的行为特征的攻击流定时，数据包大小的均匀性，TCP标志动态，统计方差。额外的分析揭示了每个类的特征属性和决策轨迹，进一步验证了这些观察到的模式。我们的研究结果有助于开发更准确和更易于解释的入侵检测系统，弥合高性能机器学习与物联网环境中人工智能驱动的网络安全的信任和问责制需求之间的差距。
摘要：As the Internet of Things (IoT) continues to expand across critical infrastructure, smart environments, and consumer devices, securing them against cyber threats has become increasingly vital. Traditional intrusion detection models often treat IoT threats as binary classification problems or rely on opaque models, thereby limiting trust. This work studies multiclass threat attribution in IoT environments using the CICIoT2023 dataset, grouping over 30 attack variants into 8 semantically meaningful classes. We utilize a combination of a gradient boosting model and SHAP (SHapley Additive exPlanations) to deliver both global and class-specific explanations, enabling detailed insight into the features driving each attack classification. The results show that the model distinguishes distinct behavioral signatures of the attacks using flow timing, packet size uniformity, TCP flag dynamics, and statistical variance. Additional analysis that exposes both feature attribution and the decision trajectory per class further validates these observed patterns. Our findings contribute to the development of more accurate and explainable intrusion detection systems, bridging the gap between high-performance machine learning and the need for trust and accountability in AI-driven cybersecurity for IoT environments.
【4】WIST: Web-Grounded Iterative Self-Play Tree for Domain-Targeted Reasoning Improvement
标题：WIST：基于Web的迭代自玩树，用于改进领域目标推理
链接：https://arxiv.org/abs/2603.22352
备注：23 pages, 4 figures. Submitted to ACL2026
摘要：Recent progress in reinforcement learning with verifiable rewards (RLVR) offers a practical path to self-improvement of language models, but existing methods face a key trade-off: endogenous self-play can drift over iterations, while corpus-grounded approaches rely on curated data environments. We present \textbf{WIST}, a \textbf{W}eb-grounded \textbf{I}terative \textbf{S}elf-play \textbf{T}ree framework for domain-targeted reasoning improvement that learns directly from the open web without requiring any pre-arranged domain corpus. WIST incrementally expands a domain tree for exploration, and retrieves and cleans path-consistent web corpus to construct a controllable training environment. It then performs Challenger--Solver self-play with verifiable rewards, and feeds learnability signals back to update node posteriors and guide subsequent exploration through an adaptive curriculum. Across four backbones, WIST consistently improves over the base models and typically outperforms both purely endogenous self-evolution and corpus-grounded self-play baselines, with the Overall gains reaching \textbf{+9.8} (\textit{Qwen3-4B-Base}) and \textbf{+9.7} (\textit{OctoThinker-8B}). WIST is also domain-steerable, improving \textit{Qwen3-8B-Base} by \textbf{+14.79} in medicine and \textit{Qwen3-4B-Base} by \textbf{+5.28} on PhyBench. Ablations further confirm the importance of WIST's key components for stable open-web learning. Our Code is available at https://github.com/lfy-123/WIST.
摘要 ：Recent progress in reinforcement learning with verifiable rewards (RLVR) offers a practical path to self-improvement of language models, but existing methods face a key trade-off: endogenous self-play can drift over iterations, while corpus-grounded approaches rely on curated data environments. We present \textbf{WIST}, a \textbf{W}eb-grounded \textbf{I}terative \textbf{S}elf-play \textbf{T}ree framework for domain-targeted reasoning improvement that learns directly from the open web without requiring any pre-arranged domain corpus. WIST incrementally expands a domain tree for exploration, and retrieves and cleans path-consistent web corpus to construct a controllable training environment. It then performs Challenger--Solver self-play with verifiable rewards, and feeds learnability signals back to update node posteriors and guide subsequent exploration through an adaptive curriculum. Across four backbones, WIST consistently improves over the base models and typically outperforms both purely endogenous self-evolution and corpus-grounded self-play baselines, with the Overall gains reaching \textbf{+9.8} (\textit{Qwen3-4B-Base}) and \textbf{+9.7} (\textit{OctoThinker-8B}). WIST is also domain-steerable, improving \textit{Qwen3-8B-Base} by \textbf{+14.79} in medicine and \textit{Qwen3-4B-Base} by \textbf{+5.28} on PhyBench. Ablations further confirm the importance of WIST's key components for stable open-web learning. Our Code is available at https://github.com/lfy-123/WIST.
【5】First-Mover Bias in Gradient Boosting Explanations: Mechanism, Detection, and Resolution
标题：梯度提升补偿中的先行者偏见：机制、检测和解决方案
链接：https://arxiv.org/abs/2603.22346
备注：28 pages, 8 figures, 12 tables; code at https://github.com/DrakeCaraker/dash-shap
摘要：我们隔离和经验性地表征先行者偏差-梯度提升中顺序残差拟合引起的特征重要性的路径依赖性浓度-作为多重共线性下基于SHAP的特征排名的众所周知的不稳定性的具体机制原因。当相关特征竞争早期分裂时，梯度提升为首先选择的任何特征创造了一个自我增强的优势：随后的树继承了有利于现任者的修改后的残差，将SHAP重要性集中在任意特征上，而不是将其分布在相关组中。放大单个模型会放大这种效果--与我们的方法具有相同总树数的大型单个模型会产生任何测试方法的最差解释。我们证明了模型独立性足以解决线性区域中的先行者偏差，并且在非线性数据生成过程中仍然是最有效的缓解方法。我们提出的方法DASH（SHAP的多样化聚合）和简单的种子平均（随机再训练）都通过打破顺序依赖链来恢复稳定性，从而证实了操作机制在所解释的模型之间是独立的。在rho=0.9时，两者都达到稳定性=0.977，而单一最佳工作流降低到0.958，大型单一模型降低到0.938。在乳腺癌数据集上，DASH相对于树计数匹配的基线将稳定性从0.32提高到0.93（+0.61）。DASH还提供了两个诊断工具--特征稳定性指数（FSI）和重要性稳定性（IS）图--它们可以在没有真实数据的情况下检测先行者偏差，使从业者能够在对特征排名采取行动之前审计解释的可靠性。可在https://github.com/DrakeCaraker/dash-shap上获得软件和可复制的基准测试。
摘要：We isolate and empirically characterize first-mover bias -- a path-dependent concentration of feature importance caused by sequential residual fitting in gradient boosting -- as a specific mechanistic cause of the well-known instability of SHAP-based feature rankings under multicollinearity. When correlated features compete for early splits, gradient boosting creates a self-reinforcing advantage for whichever feature is selected first: subsequent trees inherit modified residuals that favor the incumbent, concentrating SHAP importance on an arbitrary feature rather than distributing it across the correlated group. Scaling up a single model amplifies this effect -- a Large Single Model with the same total tree count as our method produces the worst explanations of any approach tested. We demonstrate that model independence is sufficient to resolve first-mover bias in the linear regime, and remains the most effective mitigation under nonlinear data-generating processes. Both our proposed method, DASH (Diversified Aggregation of SHAP), and simple seed-averaging (Stochastic Retrain) restore stability by breaking the sequential dependency chain, confirming that the operative mechanism is independence between explained models. At rho=0.9, both achieve stability=0.977, while the single-best workflow degrades to 0.958 and the Large Single Model to 0.938. On the Breast Cancer dataset, DASH improves stability from 0.32 to 0.93 (+0.61) against a tree-count-matched baseline. DASH additionally provides two diagnostic tools -- the Feature Stability Index (FSI) and Importance-Stability (IS) Plot -- that detect first-mover bias without ground truth, enabling practitioners to audit explanation reliability before acting on feature rankings. Software and reproducible benchmarks are available at https://github.com/DrakeCaraker/dash-shap.
【6】Stepwise Variational Inference with Vine Copulas
标题：Vine Copulas的逐步变分推理
链接：https://arxiv.org/abs/2603.22959
摘要：我们提出了逐步变分推断（VI）与藤copulas：一个通用的VI程序，结合藤copulas与一个新的逐步估计过程的变分参数。藤蔓连接函数由连接函数构建的嵌套树序列组成，其中可以随着树的数量增加来建模更复杂的潜在依赖。我们建议估计葡萄系的copula近似后逐步的方式，树树沿藤结构。进一步，我们证明了通常的向后Kullback-Leibler发散不能恢复vine copula模型中的正确参数，从而基于Rényi发散定义了证据下界.最后，一个直观的停止标准，增加更多的树木藤消除了需要预先定义的复杂性参数的变化分布，所需的大多数其他方法。因此，我们的方法在平均场VI（MFVI）和完全潜在依赖之间插值。在许多应用中，特别是稀疏高斯过程，我们的方法是吝啬的参数，而优于MFVI。
摘要：We propose stepwise variational inference (VI) with vine copulas: a universal VI procedure that combines vine copulas with a novel stepwise estimation procedure of the variational parameters. Vine copulas consist of a nested sequence of trees built from copulas, where more complex latent dependence can be modeled with increasing number of trees. We propose to estimate the vine copula approximate posterior in a stepwise fashion, tree by tree along the vine structure. Further, we show that the usual backward Kullback-Leibler divergence cannot recover the correct parameters in the vine copula model, thus the evidence lower bound is defined based on the Rényi divergence. Finally, an intuitive stopping criterion for adding further trees to the vine eliminates the need to pre-define a complexity parameter of the variational distribution, as required for most other approaches. Thus, our method interpolates between mean-field VI (MFVI) and full latent dependence. In many applications, in particular sparse Gaussian processes, our method is parsimonious with parameters, while outperforming MFVI.
检测相关(7篇)
【1】CSTS: A Canonical Security Telemetry Substrate for AI-Native Cyber Detection
标题：CSTS：用于人工智能原生网络检测的规范安全远程通信基础
链接：https://arxiv.org/abs/2603.23459
备注：21 pages including 1 appendix
摘要：人工智能驱动的网络安全系统通常会在跨环境部署下失败，这是由于分散的、以事件为中心的遥测表示。我们介绍了规范的安全Telemetry基板（CSTS），一个实体关系的抽象，强制身份持久性，类型化的关系，和时间状态不变。在异构环境中，CSTS提高了跨拓扑结构的传输，以身份为中心的检测，并防止模式扰动下的崩溃。对于零日检测，CSTS将语义方向不稳定性隔离为建模而非模式现象，从而阐明分层的可移植性需求。
摘要：AI-driven cybersecurity systems often fail under cross-environment deployment due to fragmented, event-centric telemetry representations. We introduce the Canonical Security Telemetry Substrate (CSTS), an entity-relational abstraction that enforces identity persistence, typed relationships, and temporal state invariants. Across heterogeneous environments, CSTS improves cross-topology transfer for identity-centric detection and prevents collapse under schema perturbation. For zero-day detection, CSTS isolates semantic orientation instability as a modeling, not schema, phenomenon, clarifying layered portability requirements.
【2】Machine Learning Models for the Early Detection of Burnout in Software Engineering: a Systematic Literature Review
标题：用于软件工程倦怠早期检测的机器学习模型：系统性文献综述
链接：https://arxiv.org/abs/2603.23063
备注：This paper is under review
摘要：职业倦怠是一种职业综合症，像许多其他职业一样，影响着大多数软件工程师。过去的研究显示了重要的趋势，包括越来越多地使用机器学习技术来早期检测倦怠。 本文是对提出机器学习（ML）方法的研究论文的系统性文献综述（SLR），重点是检测软件开发人员和IT专业人员的倦怠。我们的目标是审查拟议的ML技术的准确性和精确性，并为未来有兴趣复制或扩展这些研究的研究人员提出建议。 从我们的SLR中，我们观察到大多数主要研究都集中在检测情绪或利用情绪维度来检测或预测倦怠的存在。我们还进行了一项横截面研究，以检测哪种ML方法在检测情绪方面表现更好;以及哪种数据集在捕捉情绪方面具有更大的潜力和表现力。 我们相信，通过识别哪些机器学习工具和数据集在检测情绪方面表现更好，并间接识别倦怠，我们的论文可以成为在这一重要研究方向取得进展的宝贵资产。
摘要：Burnout is an occupational syndrome that, like many other professions, affects the majority of software engineers. Past research studies showed important trends, including an increasing use of machine learning techniques to allow for an early detection of burnout. This paper is a systematic literature review (SLR) of the research papers that proposed machine learning (ML) approaches, and focused on detecting burnout in software developers and IT professionals. Our objective is to review the accuracy and precision of the proposed ML techniques, and to formulate recommendations for future researchers interested to replicate or extend those studies. From our SLR we observed that a majority of primary studies focuses on detecting emotions or utilise emotional dimensions to detect or predict the presence of burnout. We also performed a cross-sectional study to detect which ML approach shows a better performance at detecting emotions; and which dataset has more potential and expressivity to capture emotions. We believe that, by identifying which ML tools and datasets show a better performance at detecting emotions, and indirectly at identifying burnout, our paper can be a valuable asset to progress in this important research direction.
【3】DariMis: Harm-Aware Modeling for Dari Misinformation Detection on YouTube
标题：DariMis：YouTube上Dari错误信息检测的危害意识建模
链接：https://arxiv.org/abs/2603.22977
备注：9 pages, 8 figures. Accepted for submission; dataset and code will be released upon publication
摘要：达里语是阿富汗的主要语言，有数千万人使用，但在错误信息检测文献中基本上没有出现。我们使用DariMis来解决这个问题，DariMis是第一个手动注释的9，224个Dari语言YouTube视频数据集，分为两个维度：信息类型（错误信息，部分真实，真实）和伤害水平（低，中，高）。一个重要的经验发现是，这些维度在结构上是相互耦合的，而不是独立的：55.9%的错误信息至少具有中等危害潜力，而真正的内容只有1.0%。这使得信息类型分类器能够用作内容审核管道中的隐式损害分类过滤器。 我们进一步提出了一种对输入编码策略，将视频标题和描述表示为单独的BERT片段输入，明确建模标题声明和主体内容之间的语义关系，这是误导信息的关键信号。一项针对单场级联的消融研究表明，尽管总体宏观F1差异不大（0.09个百分点），但对输入编码在错误信息回忆（60.1%至67.1%）中产生了7.0个百分点的增益，这是安全关键的少数类别。我们将Dari/Farsi专用模型（ParsBERT）与XLM-RoBERTa-base进行基准测试; ParsBERT实现了最佳测试性能，准确率为76.60%，宏F1为72.77%。Bootstrap 95%的置信区间报告的所有指标，我们讨论了结果的实际意义和统计局限性。
摘要：Dari, the primary language of Afghanistan, is spoken by tens of millions of people yet remains largely absent from the misinformation detection literature. We address this gap with DariMis, the first manually annotated dataset of 9,224 Dari-language YouTube videos, labeled across two dimensions: Information Type (Misinformation, Partly True, True) and Harm Level (Low, Medium, High). A central empirical finding is that these dimensions are structurally coupled, not independent: 55.9 percent of Misinformation carries at least Medium harm potential, compared with only 1.0 percent of True content. This enables Information Type classifiers to function as implicit harm-triage filters in content moderation pipelines. We further propose a pair-input encoding strategy that represents the video title and description as separate BERT segment inputs, explicitly modeling the semantic relationship between headline claims and body content, a key signal of misleading information. An ablation study against single-field concatenation shows that pair-input encoding yields a 7.0 percentage point gain in Misinformation recall (60.1 percent to 67.1 percent), the safety-critical minority class, despite modest overall macro F1 differences (0.09 percentage points). We benchmark a Dari/Farsi-specialized model (ParsBERT) against XLM-RoBERTa-base; ParsBERT achieves the best test performance with accuracy of 76.60 percent and macro F1 of 72.77 percent. Bootstrap 95 percent confidence intervals are reported for all metrics, and we discuss both the practical significance and statistical limitations of the results.
【4】Bounding Box Anomaly Scoring for simple and efficient Out-of-Distribution detection
标题：边界盒异常评分，用于简单有效的分布外检测
链接：https://arxiv.org/abs/2603.22660
备注：45 pages, 4 figures, 17 tables
摘要：分布外（OOD）检测旨在识别与训练分布不同的输入，以减少深度神经网络的不可靠预测。在事后特征空间方法中，OOD检测通常通过近似预训练网络的表示空间中的分布支持来执行。现有的方法通常反映了紧凑的参数模型之间的权衡，如马氏分数，更灵活，但基于参考的方法，如k-最近邻。边界框抽象提供了一个有吸引力的中间视角，通过紧凑的轴对齐的隐藏激活摘要来表示分布中的支持。在本文中，我们介绍了边界框异常评分（BBAS），一个事后OOD检测方法，利用边界框抽象。BBAS结合了基于区间一致性的分级异常分数、适应卷积层的监控变量以及解耦聚类和框构造，以实现更丰富的多层表示。图像分类基准测试的实验表明，BBAS提供了强大的分布和分布样本之间的分离，同时保持简单，紧凑性和可更新的边界框方法。
摘要：Out-of-distribution (OOD) detection aims to identify inputs that differ from the training distribution in order to reduce unreliable predictions by deep neural networks. Among post-hoc feature-space approaches, OOD detection is commonly performed by approximating the in-distribution support in the representation space of a pretrained network. Existing methods often reflect a trade-off between compact parametric models, such as Mahalanobis-based scores, and more flexible but reference-based methods, such as k-nearest neighbors. Bounding-box abstraction provides an attractive intermediate perspective by representing in-distribution support through compact axis-aligned summaries of hidden activations. In this paper, we introduce Bounding Box Anomaly Scoring (BBAS), a post-hoc OOD detection method that leverages bounding-box abstraction. BBAS combines graded anomaly scores based on interval exceedances, monitoring variables adapted to convolutional layers, and decoupled clustering and box construction for richer and multi-layer representations. Experiments on image-classification benchmarks show that BBAS provides robust separation between in-distribution and out-of-distribution samples while preserving the simplicity, compactness, and updateability of the bounding-box approach.
【5】A Multi-Modal CNN-LSTM Framework with Multi-Head Attention and Focal Loss for Real-Time Elderly Fall Detection
标题：具有多头注意力和焦点丧失的多模式CNN-LSTM框架，用于实时老年人跌倒检测
链接：https://arxiv.org/abs/2603.22313
摘要：全球老龄化人口的不断增加加剧了对可靠的健康监测系统的需求，特别是那些能够检测老年人跌倒等关键事件的系统。传统的跌倒检测方法依赖于单模态加速度数据，误报率很高，而传统的机器学习方法需要大量的手工特征工程。本文提出了一种新型的多模态深度学习框架MultiModalFallDetector，旨在使用可穿戴传感器进行实时老年人跌倒检测。我们的方法集成了多项创新：基于多尺度CNN的特征提取器以不同的时间分辨率捕获运动动态;融合三轴加速度计，陀螺仪和四通道生理信号;将多头自注意机制用于动态时间加权;采用焦点损失来减轻严重的类别不平衡;引入辅助活动分类任务进行正则化;并实现了从UCI HAR到SisFall数据集的迁移学习。在SisFall数据集上进行的广泛实验，包括来自老年参与者（年龄在60-85岁之间）的真实模拟跌倒试验，表明我们的框架达到了98的F1分数。第98回回忆9，AUC-ROC为99。4、显著优于包括传统机器学习和标准深度学习方法在内的基线方法。该模型在边缘设备上保持低于50 ms的推理延迟，确认其适用于老年护理环境中的实时部署。
摘要：The increasing global aging population has intensified the demand for reliable health monitoring systems, particularly those capable of detecting critical events such as falls among elderly individuals. Traditional fall detection approaches relying on single-modality acceleration data suffer from high false alarm rates, while conventional machine learning methods require extensive hand-crafted feature engineering. This paper proposes a novel multi-modal deep learning framework, MultiModalFallDetector, designed for real-time elderly fall detection using wearable sensors. Our approach integrates multiple innovations: a multi-scale CNN-based feature extractor capturing motion dynamics at varying temporal resolutions; fusion of tri-axial accelerometer, gyroscope, and four-channel physiological signals; incorporation of a multi-head self-attention mechanism for dynamic temporal weighting; adoption of Focal Loss to mitigate severe class imbalance; introduction of an auxiliary activity classification task for regularization; and implementation of transfer learning from UCI HAR to SisFall dataset. Extensive experiments on the SisFall dataset, which includes real-world simulated fall trials from elderly participants (aged 60-85), demonstrate that our framework achieves an F1-score of 98. 7, Recall of 98. 9, and AUC-ROC of 99. 4, significantly outperforming baseline methods including traditional machine learning and standard deep learning approaches. The model maintains sub- 50ms inference latency on edge devices, confirming its suitability for real-time deployment in geriatric care settings.
【6】Between Resolution Collapse and Variance Inflation: Weighted Conformal Anomaly Detection in Low-Data Regimes
标题：分辨率崩溃和方差膨胀之间：低数据状态下的加权共形异常检测
链接：https://arxiv.org/abs/2603.23205
备注：18 pages, 2 figures, 7 tables
摘要：Standard conformal anomaly detection provides marginal finite-sample guarantees under the assumption of exchangeability . However, real-world data often exhibit distribution shifts, necessitating a weighted conformal approach to adapt to local non-stationarity. We show that this adaptation induces a critical trade-off between the minimum attainable p-value and its stability. As importance weights localize to relevant calibration instances, the effective sample size decreases. This can render standard conformal p-values overly conservative for effective error control, while the smoothing technique used to mitigate this issue introduces conditional variance, potentially masking anomalies. We propose a continuous inference relaxation that resolves this dilemma by decoupling local adaptation from tail resolution via continuous weighted kernel density estimation. While relaxing finite-sample exactness to asymptotic validity, our method eliminates Monte Carlo variability and recovers the statistical power lost to discretization. Empirical evaluations confirm that our approach not only restores detection capabilities where discrete baselines yield zero discoveries, but outperforms standard methods in statistical power while maintaining valid marginal error control in practice.
【7】Modeling Quantum Federated Autoencoder for Anomaly Detection in IoT Networks
标题：为物联网网络中的异常检测建模量子联邦自动编码器
链接：https://arxiv.org/abs/2603.22366
备注：This paper has been accepted at ICOIN 2026
摘要：We propose a Quantum Federated Autoencoder for Anomaly Detection, a framework that leverages quantum federated learning for efficient, secure, and distributed processing in IoT networks. By harnessing quantum autoencoders for high-dimensional feature representation and federated learning for decentralized model training, the approach transforms localized learning on edge devices without requiring transmission of raw data, thereby preserving privacy and minimizing communication overhead. The model leverages quantum advantage in pattern recognition to enhance detection sensitivity, particularly in complex and dynamic IoT network traffic. Experiments on a real-world IoT dataset show that the proposed method delivers anomaly detection accuracy and robustness comparable to centralized approaches, while ensuring data privacy.
分类|识别(4篇)
【1】Robustness Quantification for Discriminative Models: a New Robustness Metric and its Application to Dynamic Classifier Selection
标题：鉴别模型的鲁棒性量化：一种新的鲁棒性度量及其在动态分类器选择中的应用
链接：https://arxiv.org/abs/2603.23318
摘要：Among the different possible strategies for evaluating the reliability of individual predictions of classifiers, robustness quantification stands out as a method that evaluates how much uncertainty a classifier could cope with before changing its prediction. However, its applicability is more limited than some of its alternatives, since it requires the use of generative models and restricts the analyses either to specific model architectures or discrete features. In this work, we propose a new robustness metric applicable to any probabilistic discriminative classifier and any type of features. We demonstrate that this new metric is capable of distinguishing between reliable and unreliable predictions, and use this observation to develop new strategies for dynamic classifier selection.
【2】MCLR: Improving Conditional Modeling in Visual Generative Models via Inter-Class Likelihood-Ratio Maximization and Establishing the Equivalence between Classifier-Free Guidance and Alignment Objectives
标题：麦蒂：通过类间似然比最大化改进视觉生成模型中的条件建模并建立无分类器引导和对齐目标之间的等效性
链接：https://arxiv.org/abs/2603.22364
摘要 ：Diffusion models have achieved state-of-the-art performance in generative modeling, but their success often relies heavily on classifier-free guidance (CFG), an inference-time heuristic that modifies the sampling trajectory. From a theoretical perspective, diffusion models trained with standard denoising score matching (DSM) are expected to recover the target data distribution, raising the question of why inference-time guidance is necessary in practice. In this work, we ask whether the DSM training objective can be modified in a principled manner such that standard reverse-time sampling, without inference-time guidance, yields effects comparable to CFG. We identify insufficient inter-class separation as a key limitation of standard diffusion models. To address this, we propose MCLR, a principled alignment objective that explicitly maximizes inter-class likelihood-ratios during training. Models fine-tuned with MCLR exhibit CFG-like improvements under standard sampling, achieving comparable qualitative and quantitative gains without requiring inference-time guidance. Beyond empirical benefits, we provide a theoretical result showing that the CFG-guided score is exactly the optimal solution to a weighted MCLR objective. This establishes a formal equivalence between classifier-free guidance and alignment-based objectives, offering a mechanistic interpretation of CFG.
【3】A Direct Classification Approach for Reliable Wind Ramp Event Forecasting under Severe Class Imbalance
标题：严重等级不平衡下可靠风坡道事件预测的直接分类方法
链接：https://arxiv.org/abs/2603.22326
摘要：Decision support systems are essential for maintaining grid stability in low-carbon power systems, such as wind power plants, by providing real-time alerts to control room operators regarding potential events, including Wind Power Ramp Events (WPREs). These early warnings enable the timely initiation of more detailed system stability assessments and preventive actions. However, forecasting these events is challenging due to the inherent class imbalance in WPRE datasets, where ramp events are less frequent (typically less than 15\% of observed events) compared to normal conditions. Ignoring this characteristic undermines the performance of conventional machine learning models, which often favor the majority class. This paper introduces a novel methodology for WPRE forecasting as a multivariate time series classification task and proposes a data preprocessing strategy that extracts features from recent power observations and masks unavailable ramp information, making it integrable with traditional real-time ramp identification tools. Particularly, the proposed methodology combines majority-class undersampling and ensemble learning to enhance wind ramp event forecasting under class imbalance. Numerical simulations conducted on a real-world dataset demonstrate the superiority of our approach, achieving over 85% accuracy and 88% weighted F1 score, outperforming benchmark classifiers.
【4】Overfitting and Generalizing with (PAC) Bayesian Prediction in Noisy Binary Classification
标题：有噪二进制分类中的（PAC）Bayesian预测过拟和推广
链接：https://arxiv.org/abs/2603.22644
摘要：We consider a PAC-Bayes type learning rule for binary classification, balancing the training error of a randomized ''posterior'' predictor with its KL divergence to a pre-specified ''prior''. This can be seen as an extension of a modified two-part-code Minimum Description Length (MDL) learning rule, to continuous priors and randomized predictions. With a balancing parameter of $λ=1$ this learning rule recovers an (empirical) Bayes posterior and a modified variant recovers the profile posterior, linking with standard Bayesian prediction (up to the treatment of the single-parameter noise level). However, from a risk-minimization prediction perspective, this Bayesian predictor overfits and can lead to non-vanishing excess loss in the agnostic case. Instead a choice of $λ\gg 1$, which can be seen as using a sample-size-dependent-prior, ensures uniformly vanishing excess loss even in the agnostic case. We precisely characterize the effect of under-regularizing (and over-regularizing) as a function of the balance parameter $λ$, understanding the regimes in which this under-regularization is tempered or catastrophic. This work extends previous work by Zhu and Srebro [2025] that considered only discrete priors to PAC Bayes type learning rules and, through their rigorous Bayesian interpretation, to Bayesian prediction more generally.
表征(5篇)
【1】SynForceNet: A Force-Driven Global-Local Latent Representation Framework for Lithium-Ion Battery Fault Diagnosis
标题：SynForceNet：用于锂离子电池故障诊断的力驱动全局-局部潜在表示框架
链接：https://arxiv.org/abs/2603.23265
摘要：Online safety fault diagnosis is essential for lithium-ion batteries in electric vehicles(EVs), particularly under complex and rare safety-critical conditions in real-world operation. In this work, we develop an online battery fault diagnosis network based on a deep anomaly detection framework combining kernel one-class classification and minimum-volume estimation. Mechanical constraints and spike-timing-dependent plasticity(STDP)-based dynamic representations are introduced to improve complex fault characterization and enable a more compact normal-state boundary. The proposed method is validated using 8.6 million valid data points collected from 20 EVs. Compared with several advanced baseline methods, it achieves average improvements of 7.59% in TPR, 27.92% in PPV, 18.28% in F1 score, and 23.68% in AUC. In addition, we analyze the spatial separation of fault representations before and after modeling, and further enhance framework robustness by learning the manifold structure in the latent space. The results also suggest the possible presence of shared causal structures across different fault types, highlighting the promise of integrating deep learning with physical constraints and neural dynamics for battery safety diagnosis.
【2】Conditionally Identifiable Latent Representation for Multivariate Time Series with Structural Dynamics
标题：具有结构动力学的多元时间序列的连续可识别潜在表示
链接：https://arxiv.org/abs/2603.22886
备注：Accepted paper for 2026 ICLR FINAI workshop
摘要 ：We propose the Identifiable Variational Dynamic Factor Model (iVDFM), which learns latent factors from multivariate time series with identifiability guarantees. By applying iVAE-style conditioning to the innovation process driving the dynamics rather than to the latent states, we show that factors are identifiable up to permutation and component-wise affine (or monotone invertible) transformations. Linear diagonal dynamics preserve this identifiability and admit scalable computation via companion-matrix and Krylov methods. We demonstrate improved factor recovery on synthetic data, stable intervention accuracy on synthetic SCMs, and competitive probabilistic forecasting on real-world benchmarks.
【3】Behavioral Heterogeneity as Quantum-Inspired Representation
标题：作为量子启发表示的行为异因
链接：https://arxiv.org/abs/2603.22729
摘要：Driver heterogeneity is often reduced to labels or discrete regimes, compressing what is inherently dynamic into static categories. We introduce quantum-inspired representation that models each driver as an evolving latent state, presented as a density matrix with structured mathematical properties. Behavioral observations are embedded via non-linear Random Fourier Features, while state evolution blends temporal persistence of behavior with context-dependent profile activation. We evaluate our approach on empirical driving data, Third Generation Simulation Data (TGSIM), showing how driving profiles are extracted and analyzed.
【4】Generalizing Dynamics Modeling More Easily from Representation Perspective
标题：从表示的角度更轻松地概括动力学建模
链接：https://arxiv.org/abs/2603.22655
摘要：Learning system dynamics from observations is a critical problem in many applications over various real-world complex systems, e.g., climate, ecology, and fluid systems. Recently, neural dynamics modeling method have become a prevalent solution that embeds the object's observations into a latent space before learning dynamics using neural methods such as neural Ordinary Differential Equations (ODE). Existing dynamics modeling methods induce a specific model for each observation of different complex systems, resulting in poor generalization across systems. Inspired by the great success of pre-trained models, we conduct a generalized Pre-trained Dynamics EncoDER (PDEDER) which can embed the original state observations into a latent space where the dynamics can be captured more easily. To conduct the generalized PDEDER, we pre-train any Pre-trained Language Model (PLM) by minimizing the Lyapunov exponent objective, which constrains the chaotic behavior of governing dynamics learned in the latent space. By penalizing the divergence of embedded observations, our PDEDER promotes locally stable and well-structured latent dynamics, thereby facilitating more effective dynamics modeling than in the original observation space. In addition, we incorporate reconstruction and forecasting objectives to mitigate the risk of obtaining an over-smoothed latent space. Specifically, we collect 152 sets of real-world and synthetic observations from 23 complex systems as pre-training corpora and employ them to pre-train PDEDER. Given any future dynamic observation, we can fine-tune PDEDER with any specific dynamics modeling method. We evaluate PDEDER on 12 dynamic systems by short/long-term forecasting under both in-domain and cross-domain settings, and the empirical results indicate the effectiveness and generalizability of PDEDER.
【5】Unveiling the Mechanism of Continuous Representation Full-Waveform Inversion: A Wave Based Neural Tangent Kernel Framework
标题：揭开连续表示全波倒置的机制：基于波的神经切向核框架
链接：https://arxiv.org/abs/2603.22362
备注：Published as a conference paper at ICLR 2026 (poster)
摘要：Full-waveform inversion (FWI) estimates physical parameters in the wave equation from limited measurements and has been widely applied in geophysical exploration, medical imaging, and non-destructive testing. Conventional FWI methods are limited by their notorious sensitivity to the accuracy of the initial models. Recent progress in continuous representation FWI (CR-FWI) demonstrates that representing parameter models with a coordinate-based neural network, such as implicit neural representation (INR), can mitigate the dependence on initial models. However, its underlying mechanism remains unclear, and INR-based FWI shows slower high-frequency convergence. In this work, we investigate the general CR-FWI framework and develop a unified theoretical understanding by extending the neural tangent kernel (NTK) for FWI to establish a wave-based NTK framework. Unlike standard NTK, our analysis reveals that wave-based NTK is not constant, both at initialization and during training, due to the inherent nonlinearity of FWI. We further show that the eigenvalue decay behavior of the wave-based NTK can explain why CR-FWI alleviates the dependency on initial models and shows slower high-frequency convergence. Building on these insights, we propose several CR-FWI methods with tailored eigenvalue decay properties for FWI, including a novel hybrid representation combining INR and multi-resolution grid (termed IG-FWI) that achieves a more balanced trade-off between robustness and high-frequency convergence rate. Applications in geophysical exploration on Marmousi, 2D SEG/EAGE Salt and Overthrust, 2004 BP model, and the more realistic 2014 Chevron models show the superior performance of our proposed methods compared to conventional FWI and existing INR-based FWI methods.
3D|3D重建等相关(1篇)
【1】Drop-In Perceptual Optimization for 3D Gaussian Splatting
标题：3D高斯飞溅的直接感知优化
链接：https://arxiv.org/abs/2603.23297
备注：Project page: https://apple.github.io/ml-perceptual-3dgs
摘要 ：Despite their output being ultimately consumed by human viewers, 3D Gaussian Splatting (3DGS) methods often rely on ad-hoc combinations of pixel-level losses, resulting in blurry renderings. To address this, we systematically explore perceptual optimization strategies for 3DGS by searching over a diverse set of distortion losses. We conduct the first-of-its-kind large-scale human subjective study on 3DGS, involving 39,320 pairwise ratings across several datasets and 3DGS frameworks. A regularized version of Wasserstein Distortion, which we call WD-R, emerges as the clear winner, excelling at recovering fine textures without incurring a higher splat count. WD-R is preferred by raters more than $2.3\times$ over the original 3DGS loss, and $1.5\times$ over current best method Perceptual-GS. WD-R also consistently achieves state-of-the-art LPIPS, DISTS, and FID scores across various datasets, and generalizes across recent frameworks, such as Mip-Splatting and Scaffold-GS, where replacing the original loss with WD-R consistently enhances perceptual quality within a similar resource budget (number of splats for Mip-Splatting, model size for Scaffold-GS), and leads to reconstructions being preferred by human raters $1.8\times$ and $3.6\times$, respectively. We also find that this carries over to the task of 3DGS scene compression, with $\approx 50\%$ bitrate savings for comparable perceptual metric performance.
编码器(1篇)
【1】Coordinate Encoding on Linear Grids for Physics-Informed Neural Networks
标题：物理信息神经网络线性网格上的坐标编码
链接：https://arxiv.org/abs/2603.22700
备注：21 pages, 11 figures
摘要：In solving partial differential equations (PDEs), machine learning utilizing physical laws has received considerable attention owing to advantages such as mesh-free solutions, unsupervised learning, and feasibility for solving high-dimensional problems. An effective approach is based on physics-informed neural networks (PINNs), which are based on deep neural networks known for their excellent performance in various academic and industrial applications. However, PINNs struggled with model training owing to significantly slow convergence because of a spectral bias problem. In this study, we propose a PINN-based method equipped with a coordinate-encoding layer on linear grid cells. The proposed method improves the training convergence speed by separating the local domains using grid cells. Moreover, it reduces the overall computational cost by using axis-independent linear grid cells. The method also achieves efficient and stable model training by adequately interpolating the encoded coordinates between grid points using natural cubic splines, which guarantees continuous derivative functions of the model computed for the loss functions. The results of numerical experiments demonstrate the effective performance and efficient training convergence speed of the proposed method.
优化|敛散性(3篇)
【1】Byzantine-Robust and Differentially Private Federated Optimization under Weaker Assumptions
标题：较弱假设下的拜占庭稳健和差异私有联邦优化
链接：https://arxiv.org/abs/2603.23472
备注：12 pages, 3 figures
摘要：Federated Learning (FL) enables heterogeneous clients to collaboratively train a shared model without centralizing their raw data, offering an inherent level of privacy. However, gradients and model updates can still leak sensitive information, while malicious servers may mount adversarial attacks such as Byzantine manipulation. These vulnerabilities highlight the need to address differential privacy (DP) and Byzantine robustness within a unified framework. Existing approaches, however, often rely on unrealistic assumptions such as bounded gradients, require auxiliary server-side datasets, or fail to provide convergence guarantees. We address these limitations by proposing Byz-Clip21-SGD2M, a new algorithm that integrates robust aggregation with double momentum and carefully designed clipping. We prove high-probability convergence guarantees under standard $L$-smoothness and $σ$-sub-Gaussian gradient noise assumptions, thereby relaxing conditions that dominate prior work. Our analysis recovers state-of-the-art convergence rates in the absence of adversaries and improves utility guarantees under Byzantine and DP settings. Empirical evaluations on CNN and MLP models trained on MNIST further validate the effectiveness of our approach.
【2】A Schrödinger Eigenfunction Method for Long-Horizon Stochastic Optimal Control
标题：长期随机最优控制的Schrödinger特征函数方法
链接：https://arxiv.org/abs/2603.23173
备注：Accepted to ICLR 2026, code available in https://github.com/lclaeys/eigenfunction-solver
摘要：High-dimensional stochastic optimal control (SOC) becomes harder with longer planning horizons: existing methods scale linearly in the horizon $T$, with performance often deteriorating exponentially. We overcome these limitations for a subclass of linearly-solvable SOC problems-those whose uncontrolled drift is the gradient of a potential. In this setting, the Hamilton-Jacobi-Bellman equation reduces to a linear PDE governed by an operator $\mathcal{L}$. We prove that, under the gradient drift assumption, $\mathcal{L}$ is unitarily equivalent to a Schrödinger operator $\mathcal{S} = -Δ+ \mathcal{V}$ with purely discrete spectrum, allowing the long-horizon control to be efficiently described via the eigensystem of $\mathcal{L}$. This connection provides two key results: first, for a symmetric linear-quadratic regulator (LQR), $\mathcal{S}$ matches the Hamiltonian of a quantum harmonic oscillator, whose closed-form eigensystem yields an analytic solution to the symmetric LQR with \emph{arbitrary} terminal cost. Second, in a more general setting, we learn the eigensystem of $\mathcal{L}$ using neural networks. We identify implicit reweighting issues with existing eigenfunction learning losses that degrade performance in control tasks, and propose a novel loss function to mitigate this. We evaluate our method on several long-horizon benchmarks, achieving an order-of-magnitude improvement in control accuracy compared to state-of-the-art methods, while reducing memory usage and runtime complexity from $\mathcal{O}(Td)$ to $\mathcal{O}(d)$.
【3】Double Coupling Architecture and Training Method for Optimization Problems of Differential Algebraic Equations with Parameters
标题：带参数微代数方程优化问题的双耦合结构和训练方法
链接：https://arxiv.org/abs/2603.22724
备注：19pages, 11 figures
摘要 ：Simulation and modeling are essential in product development, integrated into the design and manufacturing process to enhance efficiency and quality. They are typically represented as complex nonlinear differential algebraic equations. The growing diversity of product requirements demands multi-task optimization, a key challenge in simulation modeling research. A dual physics-informed neural network architecture has been proposed to decouple constraints and objective functions in parametric differential algebraic equation optimization problems. Theoretical analysis shows that introducing a relaxation variable with a global error bound ensures solution equivalence between the network and optimization problem. A genetic algorithm-enhanced training framework for physics-informed neural networks improves training precision and efficiency, avoiding redundant solving of differential algebraic equations. This approach enables generalization for multi-task objectives with a single, training maintaining real-time responsiveness to product requirements.
预测|估计(6篇)
【1】A Comparative Study of Machine Learning Models for Hourly Forecasting of Air Temperature and Relative Humidity
标题：每小时气温和相对湿度预测的机器学习模型比较研究
链接：https://arxiv.org/abs/2603.23282
摘要：Accurate short-term forecasting of air temperature and relative humidity is critical for urban management, especially in topographically complex cities such as Chongqing, China. This study compares seven machine learning models: eXtreme Gradient Boosting (XGBoost), Random Forest, Support Vector Regression (SVR), Multi-Layer Perceptron (MLP), Decision Tree, Long Short-Term Memory (LSTM) networks, and Convolutional Neural Network (CNN)-LSTM (CNN-LSTM), for hourly prediction using real-world open data. Based on a unified framework of data preprocessing, lag-feature construction, rolling statistical features, and time-series validation, the models are systematically evaluated in terms of predictive accuracy and robustness. The results show that XGBoost achieves the best overall performance, with a test mean absolute error (MAE) of 0.302 °C for air temperature and 1.271% for relative humidity, together with an average R2 of 0.989 across the two forecasting tasks. These findings demonstrate the strong effectiveness of tree-based ensemble learning for structured meteorological time-series forecasting and provide practical guidance for intelligent meteorological forecasting in mountainous cities.
【2】MsFormer: Enabling Robust Predictive Maintenance Services for Industrial Devices
标题：MsFormer：为工业设备提供强大的预测性维护服务
链接：https://arxiv.org/abs/2603.23076
摘要：Providing reliable predictive maintenance is a critical industrial AI service essential for ensuring the high availability of manufacturing devices. Existing deep-learning methods present competitive results on such tasks but lack a general service-oriented framework to capture complex dependencies in industrial IoT sensor data. While Transformer-based models show strong sequence modeling capabilities, their direct deployment as robust AI services faces significant bottlenecks. Specifically, streaming sensor data collected in real-world service environments often exhibits multi-scale temporal correlations driven by machine working principles. Besides, the datasets available for training time-to-failure predictive services are typically limited in size. These issues pose significant challenges for directly applying existing models as robust predictive services. To address these challenges, we propose MsFormer, a lightweight Multi-scale Transformer designed as a unified AI service model for reliable industrial predictive maintenance. MsFormer incorporates a Multi-scale Sampling (MS) module and a tailored position encoding mechanism to capture sequential correlations across multi-streaming service data. Additionally, to accommodate data-scarce service environments, MsFormer adopts a lightweight attention mechanism with straightforward pooling operations instead of self-attention. Extensive experiments on real-world datasets demonstrate that the proposed framework achieves significant performance improvements over state-of-the-art methods. Furthermore, MsFormer outperforms across industrial devices and operating conditions, demonstrating strong generalizability while maintaining a highly reliable Quality of Service (QoS).
【3】Multitask-Informed Prior for In-Context Learning on Tabular Data: Application to Steel Property Prediction
标题：表格数据内上下文学习的多任务知情先验：应用于钢铁资产预测
链接：https://arxiv.org/abs/2603.22738
摘要：Accurate prediction of mechanical properties of steel during hot rolling processes, such as Thin Slab Direct Rolling (TSDR), remains challenging due to complex interactions among chemical compositions, processing parameters, and resultant microstructures. Traditional empirical and experimental methodologies, while effective, are often resource-intensive and lack adaptability to varied production conditions. Moreover, most existing approaches do not explicitly leverage the strong correlations among key mechanical properties, missing an opportunity to improve predictive accuracy through multitask learning. To address this, we present a multitask learning framework that injects multitask awareness into the prior of TabPFN--a transformer-based foundation model for in-context learning on tabular data--through novel fine-tuning strategies. Originally designed for single-target regression or classification, we augment TabPFN's prior with two complementary approaches: (i) target averaging, which provides a unified scalar signal compatible with TabPFN's single-target architecture, and (ii) task-specific adapters, which introduce task-specific supervision during fine-tuning. These strategies jointly guide the model toward a multitask-informed prior that captures cross-property relationships among key mechanical metrics. Extensive experiments on an industrial TSDR dataset demonstrate that our multitask adaptations outperform classical machine learning methods and recent state-of-the-art tabular learning models across multiple evaluation metrics. Notably, our approach enhances both predictive accuracy and computational efficiency compared to task-specific fine-tuning, demonstrating that multitask-aware prior adaptation enables foundation models for tabular data to deliver scalable, rapid, and reliable deployment for automated industrial quality control and process optimization in TSDR.
【4】Cloud-Edge Collaborative Large Models for Robust Photovoltaic Power Forecasting
标题：云边缘协作大型模型用于稳健的太阳能电力预测
链接：https://arxiv.org/abs/2603.22343
摘要：Photovoltaic (PV) power forecasting in edge-enabled grids requires balancing forecasting accuracy, robustness under weather-driven distribution shifts, and strict latency constraints. Local specialized models are efficient for routine conditions but often degrade under rare ramp events and unseen weather patterns, whereas always relying on cloud-side large models incurs substantial communication delay and cloud overhead. To address this challenge, we propose a risk-aware cloud-edge collaborative framework for latency-sensitive PV forecasting. The framework integrates a site-specific expert predictor for routine cases, a lightweight edge-side model for enhanced local inference, and a cloud-side large retrieval model that provides matched historical context when needed through a retrieval-prediction pipeline. A lightweight screening module estimates predictive uncertainty, out-of-distribution risk, weather mutation intensity, and model disagreement, while a Lyapunov-guided router selectively escalates inference to the edge-small or cloud-assisted branches under long-term latency, communication, and cloud-usage constraints. The outputs of the activated branches are combined through adaptive fusion. Experiments on two real-world PV datasets demonstrate a favorable overall trade-off among forecasting accuracy, routing quality, robustness, and system efficiency.
【5】Enhancing AI-Based Tropical Cyclone Track and Intensity Forecasting via Systematic Bias Correction
标题：通过系统偏差修正加强基于人工智能的热带气旋路径和强度预测
链接：https://arxiv.org/abs/2603.22314
摘要：Tropical cyclones (TCs) pose severe threats to life, infrastructure, and economies in tropical and subtropical regions, underscoring the critical need for accurate and timely forecasts of both track and intensity. Recent advances in AI-based weather forecasting have shown promise in improving TC track forecasts. However, these systems are typically trained on coarse-resolution reanalysis data (e.g., ERA5 at 0.25 degree), which constrains predicted TC positions to a fixed grid and introduces significant discretization errors. Moreover, intensity forecasting remains limited especially for strong TCs by the smoothing effect of coarse meteorological fields and the use of regression losses that bias predictions toward conditional means. To address these limitations, we propose BaguanCyclone, a novel, unified framework that integrates two key innovations: (1) a probabilistic center refinement module that models the continuous spatial distribution of TC centers, enabling finer track precision; and (2) a region-aware intensity forecasting module that leverages high-resolution internal representations within dynamically defined sub-grid zones around the TC core to better capture localized extremes. Evaluated on the global IBTrACS dataset across six major TC basins, our system consistently outperforms both operational numerical weather prediction (NWP) models and most AI-based baselines, delivering a substantial enhancement in forecast accuracy. Remarkably, BaguanCyclone excels in navigating meteorological complexities, consistently delivering accurate forecasts for re-intensification, sweeping arcs, twin cyclones, and meandering events. Our code is available at https://github.com/DAMO-DI-ML/Baguan-cyclone.
【6】Fair splits flip the leaderboard: CHANRG reveals limited generalization in RNA secondary-structure prediction
标题：公平的分歧颠覆了排行榜：CHRANRG揭示了RNA二级结构预测的普遍性有限
链接：https://arxiv.org/abs/2603.22330
摘要：Accurate prediction of RNA secondary structure underpins transcriptome annotation, mechanistic analysis of non-coding RNAs, and RNA therapeutic design. Recent gains from deep learning and RNA foundation models are difficult to interpret because current benchmarks may overestimate generalization across RNA families. We present the Comprehensive Hierarchical Annotation of Non-coding RNA Groups (CHANRG), a benchmark of 170{,}083 structurally non-redundant RNAs curated from more than 10 million sequences in Rfam~15.0 using structure-aware deduplication, genome-aware split design and multiscale structural evaluation. Across 29 predictors, foundation-model methods achieved the highest held-out accuracy but lost most of that advantage out of distribution, whereas structured decoders and direct neural predictors remained markedly more robust. This gap persisted after controlling for sequence length and reflected both loss of structural coverage and incorrect higher-order wiring. Together, CHANRG and a padding-free, symmetry-aware evaluation stack provide a stricter and batch-invariant framework for developing RNA structure predictors with demonstrable out-of-distribution robustness.
其他神经网络|深度学习|模型|建模(24篇)
【1】VTAM: Video-Tactile-Action Models for Complex Physical Interaction Beyond VLAs
标题：VLAM：超越VLA的复杂物理交互的视频-触觉-动作模型
链接：https://arxiv.org/abs/2603.23481
备注：https://plan-lab.github.io/projects/vtam/
摘要：Video-Action Models (VAMs) have emerged as a promising framework for embodied intelligence, learning implicit world dynamics from raw video streams to produce temporally consistent action predictions. Although such models demonstrate strong performance on long-horizon tasks through visual reasoning, they remain limited in contact-rich scenarios where critical interaction states are only partially observable from vision alone. In particular, fine-grained force modulation and contact transitions are not reliably encoded in visual tokens, leading to unstable or imprecise behaviors. To bridge this gap, we introduce the Video-Tactile Action Model (VTAM), a multimodal world modeling framework that incorporates tactile perception as a complementary grounding signal. VTAM augments a pretrained video transformer with tactile streams via a lightweight modality transfer finetuning, enabling efficient cross-modal representation learning without tactile-language paired data or independent tactile pretraining. To stabilize multimodal fusion, we introduce a tactile regularization loss that enforces balanced cross-modal attention, preventing visual latent dominance in the action model. VTAM demonstrates superior performance in contact-rich manipulation, maintaining a robust success rate of 90 percent on average. In challenging scenarios such as potato chip pick-and-place requiring high-fidelity force awareness, VTAM outperforms the pi 0.5 baseline by 80 percent. Our findings demonstrate that integrating tactile feedback is essential for correcting visual estimation errors in world action models, providing a scalable approach to physically grounded embodied foundation models.
【2】Similarity-Aware Mixture-of-Experts for Data-Efficient Continual Learning
标题：具有相似性的专家混合，实现数据高效的持续学习
链接：https://arxiv.org/abs/2603.23436
备注：9 pages
摘要：Machine learning models often need to adapt to new data after deployment due to structured or unstructured real-world dynamics. The Continual Learning (CL) framework enables continuous model adaptation, but most existing approaches either assume each task contains sufficiently many data samples or that the learning tasks are non-overlapping. In this paper, we address the more general setting where each task may have a limited dataset, and tasks may overlap in an arbitrary manner without a priori knowledge. This general setting is substantially more challenging for two reasons. On the one hand, data scarcity necessitates effective contextualization of general knowledge and efficient knowledge transfer across tasks. On the other hand, unstructured task overlapping can easily result in negative knowledge transfer. To address the above challenges, we propose an adaptive mixture-of-experts (MoE) framework over pre-trained models that progressively establishes similarity awareness among tasks. Our design contains two innovative algorithmic components: incremental global pooling and instance-wise prompt masking. The former mitigates prompt association noise through gradual prompt introduction over time. The latter decomposes incoming task samples into those aligning with current prompts (in-distribution) and those requiring new prompts (out-of-distribution). Together, our design strategically leverages potential task overlaps while actively preventing negative mutual interference in the presence of per-task data scarcity. Experiments across varying data volumes and inter-task similarity show that our method enhances sample efficiency and is broadly applicable.
【3】General Machine Learning: Theory for Learning Under Variable Regimes
标题：通用机器学习：可变状态下的学习理论
链接：https://arxiv.org/abs/2603.23220
备注：56 pages
摘要：We study learning under regime variation, where the learner, its memory state, and the evaluative conditions may evolve over time. This paper is a foundational and structural contribution: its goal is to define the core learning-theoretic objects required for such settings and to establish their first theorem-supporting consequences. The paper develops a regime-varying framework centered on admissible transport, protected-core preservation, and evaluator-aware learning evolution. It records the immediate closure consequences of admissibility, develops a structural obstruction argument for faithful fixed-ontology reduction in genuinely multi-regime settings, and introduces a protected-stability template together with explicit numerical and symbolic witnesses on controlled subclasses, including convex and deductive settings. It also establishes theorem-layer results on evaluator factorization, morphisms, composition, and partial kernel-level alignment across semantically commensurable layers. A worked two-regime example makes the admissibility certificate, protected evaluative core, and regime-variation cost explicit on a controlled subclass. The symbolic component is deliberately restricted in scope: the paper establishes a first kernel-level compatibility result together with a controlled monotonic deductive witness. The manuscript should therefore be read as introducing a structured learning-theoretic framework for regime-varying learning together with its first theorem-supporting layer, not as a complete quantitative theory of all learning systems.
【4】A Bayesian Learning Approach for Drone Coverage Network: A Case Study on Cardiac Arrest in Scotland
标题：无人机覆盖网络的贝叶斯学习方法：苏格兰心脏骤停案例研究
链接：https://arxiv.org/abs/2603.23134
摘要：Drones are becoming popular as a complementary system for \ac{ems}. Although several pilot studies and flight trials have shown the feasibility of drone-assisted \ac{aed} delivery, running a full-scale operational network remains challenging due to high capital expenditure and environmental uncertainties. In this paper, we formulate a reliability-informed Bayesian learning framework for designing drone-assisted \ac{aed} delivery networks under environmental and operational uncertainty. We propose our objective function based on the survival probability of \ac{ohca} patients to identify the ideal locations of drone stations. Moreover, we consider the coverage of existing \ac{ems} infrastructure to improve the response reliability in remote areas. We illustrate our proposed method using geographically referenced cardiac arrest data from Scotland. The result shows how environmental variability and spatial demand patterns influence optimal drone station placement across urban and rural regions. In addition, we assess the robustness of the network and evaluate its economic viability using a cost-effectiveness analysis based on expected \ac{qaly}. The findings suggest that drone-assisted \ac{aed} delivery is expected to be cost-effective and has the potential to significantly improve the emergency response coverage in rural and urban areas with longer ambulance response times.
【5】Policy-based Tuning of Autoregressive Image Models with Instance- and Distribution-Level Rewards
标题：基于策略的自回归图像模型调整，并提供实例级和分布级奖励
链接：https://arxiv.org/abs/2603.23086
摘要 ：Autoregressive (AR) models are highly effective for image generation, yet their standard maximum-likelihood estimation training lacks direct optimization for sample quality and diversity. While reinforcement learning (RL) has been used to align diffusion models, these methods typically suffer from output diversity collapse. Similarly, concurrent RL methods for AR models rely strictly on instance-level rewards, often trading off distributional coverage for quality. To address these limitations, we propose a lightweight RL framework that casts token-based AR synthesis as a Markov Decision Process, optimized via Group Relative Policy Optimization (GRPO). Our core contribution is the introduction of a novel distribution-level Leave-One-Out FID (LOO-FID) reward; by leveraging an exponential moving average of feature moments, it explicitly encourages sample diversity and prevents mode collapse during policy updates. We integrate this with composite instance-level rewards (CLIP and HPSv2) for strict semantic and perceptual fidelity, and stabilize the multi-objective learning with an adaptive entropy regularization term. Extensive experiments on LlamaGen and VQGAN architectures demonstrate clear improvements across standard quality and diversity metrics within only a few hundred tuning iterations. The results also show that the model can be updated to produce competitive samples even without Classifier-Free Guidance, and bypass its 2x inference cost.
【6】Generalization Bounds for Physics-Informed Neural Networks for the Incompressible Navier-Stokes Equations
标题：不可压Navier-Stokes方程物理信息神经网络的推广界
链接：https://arxiv.org/abs/2603.23072
摘要：This work establishes rigorous first-of-its-kind upper bounds on the generalization error for the method of approximating solutions to the (d+1)-dimensional incompressible Navier-Stokes equations by training depth-2 neural networks trained via the unsupervised Physics-Informed Neural Network (PINN) framework. This is achieved by bounding the Rademacher complexity of the PINN risk. For appropriately weight bounded net classes our derived generalization bounds do not explicitly depend on the network width and our framework characterizes the generalization gap in terms of the fluid's kinematic viscosity and loss regularization parameters. In particular, the resulting sample complexity bounds are dimension-independent. Our generalization bounds suggest using novel activation functions for solving fluid dynamics. We provide empirical validation of the suggested activation functions and the corresponding bounds on a PINN setup solving the Taylor-Green vortex benchmark.
【7】Assessing the Robustness of Climate Foundation Models under No-Analog Distribution Shifts
标题：评估非模拟分布漂移下气候基金会模型的稳健性
链接：https://arxiv.org/abs/2603.23043
备注：Accepted at Machine Learning Earth
摘要：The accelerating pace of climate change introduces profound non-stationarities that challenge the ability of Machine Learning based climate emulators to generalize beyond their training distributions. While these emulators offer computationally efficient alternatives to traditional Earth System Models, their reliability remains a potential bottleneck under "no-analog" future climate states, which we define here as regimes where external forcing drives the system into conditions outside the empirical range of the historical training data. A fundamental challenge in evaluating this reliability is data contamination; because many models are trained on simulations that already encompass future scenarios, true out-of-distribution (OOD) performance is often masked. To address this, we benchmark the OOD robustness of three state-of-the-art architectures: U-Net, ConvLSTM, and the ClimaX foundation model specifically restricted to a historical-only training regime (1850-2014). We evaluate these models using two complementary strategies: (i) temporal extrapolation to the recent climate (2015-2023) and (ii) cross-scenario forcing shifts across divergent emission pathways. Our analysis within this experimental setup reveals an accuracy vs. stability trade-off: while the ClimaX foundation model achieves the lowest absolute error, it exhibits higher relative performance changes under distribution shifts, with precipitation errors increasing by up to 8.44% under extreme forcing scenarios. These findings suggest that when restricted to historical training dynamics, even high-capacity foundation models are sensitive to external forcing trajectories. Our results underscore the necessity of scenario-aware training and rigorous OOD evaluation protocols to ensure the robustness of climate emulators under a changing climate.
【8】Asymptotic Learning Curves for Diffusion Models with Random Features Score and Manifold Data
标题：具有随机特征得分和流形数据的扩散模型的渐近学习曲线
链接：https://arxiv.org/abs/2603.22962
备注：12 pages
摘要：We study the theoretical behavior of denoising score matching--the learning task associated to diffusion models--when the data distribution is supported on a low-dimensional manifold and the score is parameterized using a random feature neural network. We derive asymptotically exact expressions for the test, train, and score errors in the high-dimensional limit. Our analysis reveals that, for linear manifolds the sample complexity required to learn the score function scales linearly with the intrinsic dimension of the manifold, rather than with the ambient dimension. Perhaps surprisingly, the benefits of low-dimensional structure starts to diminish once we have a non-linear manifold. These results indicate that diffusion models can benefit from structured data; however, the dependence on the specific type of structure is subtle and intricate.
【9】From Arithmetic to Logic: The Resilience of Logic and Lookup-Based Neural Networks Under Parameter Bit-Flips
标题：从算术到逻辑：参数位翻转下逻辑和基于查找的神经网络的弹性
链接：https://arxiv.org/abs/2603.22770
摘要 ：The deployment of deep neural networks (DNNs) in safety-critical edge environments necessitates robustness against hardware-induced bit-flip errors. While empirical studies indicate that reducing numerical precision can improve fault tolerance, the theoretical basis of this phenomenon remains underexplored. In this work, we study resilience as a structural property of neural architectures rather than solely as a property of a dataset-specific trained solution. By deriving the expected squared error (MSE) under independent parameter bit flips across multiple numerical formats and layer primitives, we show that lower precision, higher sparsity, bounded activations, and shallow depth are consistently favored under this corruption model. We then argue that logic and lookup-based neural networks realize the joint limit of these design trends. Through ablation studies on the MLPerf Tiny benchmark suite, we show that the observed empirical trends are consistent with the theoretical predictions, and that LUT-based models remain highly stable in corruption regimes where standard floating-point models fail sharply. Furthermore, we identify a novel even-layer recovery effect unique to logic-based architectures and analyze the structural conditions under which it emerges. Overall, our results suggest that shifting from continuous arithmetic weights to discrete Boolean lookups can provide a favorable accuracy-resilience trade-off for hardware fault tolerance.
【10】Reconstruction-Guided Slot Curriculum: Addressing Object Over-Fragmentation in Video Object-Centric Learning
标题：重建引导老虎机课程：解决视频以对象为中心学习中的对象过度碎片化问题
链接：https://arxiv.org/abs/2603.22758
备注：CVPR 2026 paper. Our code is available at github.com/wjun0830/SlotCurri
摘要：Video Object-Centric Learning seeks to decompose raw videos into a small set of object slots, but existing slot-attention models often suffer from severe over-fragmentation. This is because the model is implicitly encouraged to occupy all slots to minimize the reconstruction objective, thereby representing a single object with multiple redundant slots. We tackle this limitation with a reconstruction-guided slot curriculum (SlotCurri). Training starts with only a few coarse slots and progressively allocates new slots where reconstruction error remains high, thus expanding capacity only where it is needed and preventing fragmentation from the outset. Yet, during slot expansion, meaningful sub-parts can emerge only if coarse-level semantics are already well separated; however, with a small initial slot budget and an MSE objective, semantic boundaries remain blurry. Therefore, we augment MSE with a structure-aware loss that preserves local contrast and edge information to encourage each slot to sharpen its semantic boundaries. Lastly, we propose a cyclic inference that rolls slots forward and then backward through the frame sequence, producing temporally consistent object representations even in the earliest frames. All combined, SlotCurri addresses object over-fragmentation by allocating representational capacity where reconstruction fails, further enhanced by structural cues and cyclic inference. Notable FG-ARI gains of +6.8 on YouTube-VIS and +8.3 on MOVi-C validate the effectiveness of SlotCurri. Our code is available at github.com/wjun0830/SlotCurri.
【11】Causal Discovery in Action: Learning Chain-Reaction Mechanisms from Interventions
标题：行动中的因果发现：从干预中学习连锁反应机制
链接：https://arxiv.org/abs/2603.22620
备注：Accepted to the 5th Conference on Causal Learning and Reasoning (CLeaR 2026)
摘要：Causal discovery is challenging in general dynamical systems because, without strong structural assumptions, the underlying causal graph may not be identifiable even from interventional data. However, many real-world systems exhibit directional, cascade-like structure, in which components activate sequentially and upstream failures suppress downstream effects. We study causal discovery in such chain-reaction systems and show that the causal structure is uniquely identifiable from blocking interventions that prevent individual components from activating. We propose a minimal estimator with finite-sample guarantees, achieving exponential error decay and logarithmic sample complexity. Experiments on synthetic models and diverse chain-reaction environments demonstrate reliable recovery from a few interventions, while observational heuristics fail in regimes with delayed or overlapping causal effects.
【12】A Foundation Model for Instruction-Conditioned In-Context Time Series Tasks
标题：受教学条件影响的上下文时间序列任务的基础模型
链接：https://arxiv.org/abs/2603.22586
摘要：In-context learning (ICL) allows a model to adapt at inference time by conditioning on examples rather than updating parameters. Existing time-series foundation models use implicit positional context, retrieval, or task-specific objectives, but rarely explicit instruction-conditioned demonstrations. We present a foundation model for instruction-conditioned in-context time-series tasks based on a quantile-regression T5 encoder-decoder. Historical examples and queries are encoded with a structured tokenization scheme that marks target series, covariates, context, and task-specific future information. A hierarchical Transformer with per-example encoding, example-level fusion, and cross-example attention conditions decoding on demonstration pairs, enabling forecasting and related tasks without task-specific fine-tuning. We train on large-scale real and synthetic time series using supervised forecasting plus self-supervised tasks, including imputation, reconstruction, classification, anomaly detection, and source demixing. This multi-task training learns a distribution over task mappings and improves adaptation to local structure at inference time. Across diverse datasets, frequencies, and horizons, our method outperforms strong foundation baselines on point and probabilistic forecasting benchmarks, including fev-bench and GIFT-Eval, while remaining competitive on classification and anomaly detection.
【13】Wake Up to the Past: Using Memory to Model Fluid Wake Effects on Robots
标题：唤醒过去：使用记忆建模机器人的流体尾流效应
链接：https://arxiv.org/abs/2603.22472
备注：8 pages, 7 figures. Submitted to IROS 2026. Project website: https://sites.google.com/view/wake-up-to-the-past
摘要 ：Autonomous aerial and aquatic robots that attain mobility by perturbing their medium, such as multicopters and torpedoes, produce wake effects that act as disturbances for adjacent robots. Wake effects are hard to model and predict due to the chaotic spatio-temporal dynamics of the fluid, entangled with the physical geometry of the robots and their complex motion patterns. Data-driven approaches using neural networks typically learn a memory-less function that maps the current states of the two robots to a force observed by the "sufferer" robot. Such models often perform poorly in agile scenarios: since the wake effect has a finite propagation time, the disturbance observed by a sufferer robot is some function of relative states in the past. In this work, we present an empirical study of the properties a wake-effect predictor must satisfy to accurately model the interactions between two robots mediated by a fluid. We explore seven data-driven models designed to capture the spatio-temporal evolution of fluid wake effects in four different media. This allows us to introspect the models and analyze the reasons why certain features enable improved accuracy in prediction across predictors and fluids. As experimental validation, we develop a planar rectilinear gantry for two spinning monocopters to test in real-world data with feedback control. The conclusion is that support of history of previous states as input and transport delay prediction substantially helps to learn an accurate wake-effect predictor.
【14】Computational Arbitrage in AI Model Markets
标题：人工智能模型市场中的计算套利
链接：https://arxiv.org/abs/2603.22404
摘要：Consider a market of competing model providers selling query access to models with varying costs and capabilities. Customers submit problem instances and are willing to pay up to a budget for a verifiable solution. An arbitrageur efficiently allocates inference budget across providers to undercut the market, thus creating a competitive offering with no model-development risk. In this work, we initiate the study of arbitrage in AI model markets, empirically demonstrating the viability of arbitrage and illustrating its economic consequences. We conduct an in-depth case study of SWE-bench GitHub issue resolution using two representative models, GPT-5 mini and DeepSeek v3.2. In this verifiable domain, simple arbitrage strategies generate net profit margins of up to 40%. Robust arbitrage strategies that generalize across different domains remain profitable. Distillation further creates strong arbitrage opportunities, potentially at the expense of the teacher model's revenue. Multiple competing arbitrageurs drive down consumer prices, reducing the marginal revenue of model providers. At the same time, arbitrage reduces market segmentation and facilitates market entry for smaller model providers by enabling earlier revenue capture. Our results suggest that arbitrage can be a powerful force in AI model markets with implications for model development, distillation, and deployment.
【15】COMPASS-Hedge: Learning Safely Without Knowing the World
标题：指南针对冲：在不了解世界的情况下安全学习
链接：https://arxiv.org/abs/2603.22348
摘要：Online learning algorithms often faces a fundamental trilemma: balancing regret guarantees between adversarial and stochastic settings and providing baseline safety against a fixed comparator. While existing methods excel in one or two of these regimes, they typically fail to unify all three without sacrificing optimal rates or requiring oracle access to problem-dependent parameters. In this work, we bridge this gap by introducing COMPASS-Hedge. Our algorithm is the first full-information method to simultaneously achieve: i) Minimax-optimal regret in adversarial environments; ii) Instance-optimal, gap-dependent regret in stochastic environments; and iii) $\tilde{\mathcal{O}}(1)$ regret relative to a designated baseline policy, up to logarithmic factors. Crucially, COMPASS-Hedge is parameter-free and requires no prior knowledge of the environment's nature or the magnitude of the stochastic sub optimality gaps. Our approach hinges on a novel integration of adaptive pseudo-regret scaling and phase-based aggression, coupled with a comparator-aware mixing strategy. To the best of our knowledge, this provides the first "best-of-three-world" guarantee in the full-information setting, establishing that baseline safety does not have to come at the cost of worst-case robustness or stochastic efficiency.
【16】A Multi-Task Targeted Learning Framework for Lithium-Ion Battery State-of-Health and Remaining Useful Life
标题：锂离子电池健康状况和剩余使用寿命的多任务目标学习框架
链接：https://arxiv.org/abs/2603.22323
备注：https://github.com/wch1121/Joint-prediction-of-SOH-and-RUL
摘要：Accurately predicting the state-of-health (SOH) and remaining useful life (RUL) of lithium-ion batteries is crucial for ensuring the safe and efficient operation of electric vehicles while minimizing associated risks. However, current deep learning methods are limited in their ability to selectively extract features and model time dependencies for these two parameters. Moreover, most existing methods rely on traditional recurrent neural networks, which have inherent shortcomings in long-term time-series modeling. To address these issues, this paper proposes a multi-task targeted learning framework for SOH and RUL prediction, which integrates multiple neural networks, including a multi-scale feature extraction module, an improved extended LSTM, and a dual-stream attention module. First, a feature extraction module with multi-scale CNNs is designed to capture detailed local battery decline patterns. Secondly, an improved extended LSTM network is employed to enhance the model's ability to retain long-term temporal information, thus improving temporal relationship modeling. Building on this, the dual-stream attention module-comprising polarized attention and sparse attention to selectively focus on key information relevant to SOH and RUL, respectively, by assigning higher weights to important features. Finally, a many-to-two mapping is achieved through the dual-task layer. To optimize the model's performance and reduce the need for manual hyperparameter tuning, the Hyperopt optimization algorithm is used. Extensive comparative experiments on battery aging datasets demonstrate that the proposed method reduces the average RMSE for SOH and RUL predictions by 111.3\% and 33.0\%, respectively, compared to traditional and state-of-the-art methods.
【17】Bridging the Gap Between Climate Science and Machine Learning in Climate Model Emulation
标题：弥合气候模型仿真中气候科学和机器学习之间的差距
链接：https://arxiv.org/abs/2603.22320
摘要：While climate models provide insights for climate decision-making, their use is constrained by significant computational and technical demands. Although machine learning (ML) emulators offer a way to bypass the high computational costs, their effective use remains challenging. The hurdles are diverse, ranging from limited accessibility and a lack of specialized knowledge to a general mistrust of ML methods that are perceived as insufficiently physical. Here, we introduce a framework to overcome these barriers by integrating both climate science and machine learning perspectives. We find that designing easy-to-adopt emulators that address a clearly defined task and demonstrating their reliability offers a promising path for bridging the gap between our two fields.
【18】UniFluids: Unified Neural Operator Learning with Conditional Flow-matching
标题：UniFluids：具有条件流匹配的统一神经操作员学习
链接：https://arxiv.org/abs/2603.22309
备注：Preprint version. Work in progress
摘要：Partial differential equation (PDE) simulation holds extensive significance in scientific research. Currently, the integration of deep neural networks to learn solution operators of PDEs has introduced great potential. In this paper, we present UniFluids, a conditional flow-matching framework that harnesses the scalability of diffusion Transformer to unify learning of solution operators across diverse PDEs with varying dimensionality and physical variables. Unlike the autoregressive PDE foundation models, UniFluids adopts flow-matching to achieve parallel sequence generation, making it the first such approach for unified operator learning. Specifically, the introduction of a unified four-dimensional spatiotemporal representation for the heterogeneous PDE datasets enables joint training and conditional encoding. Furthermore, we find the effective dimension of the PDE dataset is much lower than its patch dimension. We thus employ $x$-prediction in the flow-matching operator learning, which is verified to significantly improve prediction accuracy. We conduct a large-scale evaluation of UniFluids on several PDE datasets covering spatial dimensions 1D, 2D and 3D. Experimental results show that UniFluids achieves strong prediction accuracy and demonstrates good scalability and cross-scenario generalization capability. The code will be released later.
【19】Full waveform inversion method based on diffusion model
标题：基于扩散模型的全波反求方法
链接：https://arxiv.org/abs/2603.22307
摘要：Seismic full-waveform inversion is a core technology for obtaining high-resolution subsurface model parameters. However, its highly nonlinear characteristics and strong dependence on the initial model often lead to the inversion process getting trapped in local minima. In recent years, generative diffusion models have provided a way to regularize full-waveform inversion by learning implicit prior distributions. However, existing methods mostly use unconditional diffusion processes, ignoring the inherent physical coupling relationship between velocity and density and other physical properties. This paper proposes a full-waveform inversion method based on conditional diffusion model regularization. By improving the backbone network structure of the diffusion model, two-dimensional density information is introduced as a conditional input into the U-Net network. Experimental results show that the full-waveform inversion method based on the conditional diffusion model significantly improves the resolution and structural fidelity of the inversion results, and exhibits stronger stability and robustness when dealing with complex situations. This method effectively utilizes density information to constrain the inversion and has good practical application value. Keywords: Deep learning; Diffusion model; Full waveform inversion.
【20】High-Resolution Tensor-Network Fourier Methods for Exponentially Compressed Non-Gaussian Aggregate Distributions
标题：指数压缩非高斯聚集分布的高分辨率张量网络傅里叶方法
链接：https://arxiv.org/abs/2603.23106
备注：22 pages, 13 figures
摘要：Characteristic functions of weighted sums of independent random variables exhibit low-rank structure in the quantized tensor train (QTT) representation, also known as matrix product states (MPS), enabling up to exponential compression of their fully non-Gaussian probability distributions. Under variable independence, the global characteristic function factorizes into local terms. Its low-rank QTT structure arises from intrinsic spectral smoothness in continuous models, or from spectral energy concentration as the number of components $D$ grows in discrete models. We demonstrate this on weighted sums of Bernoulli and lognormal random variables. In the former, despite an adversarial, incompressible small-$D$ regime, the characteristic function undergoes a sharp bond-dimension collapse for $D \gtrsim 300$ components, enabling polylogarithmic time and memory scaling. In the latter, the approach reaches high-resolution discretizations of $N = 2^{30}$ frequency modes on standard hardware, far beyond the $N = 2^{24}$ ceiling of dense implementations. These compressed representations enable efficient computation of Value at Risk (VaR) and Expected Shortfall (ES), supporting applications in quantitative finance and beyond.
【21】Post-Selection Distributional Model Evaluation
标题：选择后分布模型评估
链接：https://arxiv.org/abs/2603.23055
摘要：Formal model evaluation methods typically certify that a model satisfies a prescribed target key performance indicator (KPI) level. However, in many applications, the relevant target KPI level may not be known a priori, and the user may instead wish to compare candidate models by analyzing the full trade-offs between performance and reliability achievable at test time by the models. This task, requiring the reliable estimate of the test-time KPI distributions, is made more complicated by the fact that the same data must often be used both to pre-select a subset of candidate models and to estimate their KPI distributions, causing a potential post-selection bias. In this work, we introduce post-selection distributional model evaluation (PS-DME), a general framework for statistically valid distributional model assessment after arbitrary data-dependent model pre-selection. Building on e-values, PS-DME controls post-selection false coverage rate (FCR) for the distributional KPI estimates and is proved to be more sample efficient than a baseline method based on sample splitting. Experiments on synthetic data, text-to-SQL decoding with large language models, and telecom network performance evaluation demonstrate that PS-DME enables reliable comparison of candidate configurations across a range of reliability levels, supporting the statistically reliable exploration of performance--reliability trade-offs.
【22】A PAC-Bayesian approach to generalization for quantum models
标题：量子模型推广的PAC-Bayesian方法
链接：https://arxiv.org/abs/2603.22964
备注：15+29 pages, 4 figures
摘要：Generalization is a central concept in machine learning theory, yet for quantum models, it is predominantly analyzed through uniform bounds that depend on a model's overall capacity rather than the specific function learned. These capacity-based uniform bounds are often too loose and entirely insensitive to the actual training and learning process. Previous theoretical guarantees have failed to provide non-uniform, data-dependent bounds that reflect the specific properties of the learned solution rather than the worst-case behavior of the entire hypothesis class. To address this limitation, we derive the first PAC-Bayesian generalization bounds for a broad class of quantum models by analyzing layered circuits composed of general quantum channels, which include dissipative operations such as mid-circuit measurements and feedforward. Through a channel perturbation analysis, we establish non-uniform bounds that depend on the norms of learned parameter matrices; we extend these results to symmetry-constrained equivariant quantum models; and we validate our theoretical framework with numerical experiments. This work provides actionable model design insights and establishes a foundational tool for a more nuanced understanding of generalization in quantum machine learning.
【23】Off-Policy Evaluation and Learning for Survival Outcomes under Censoring
标题：截尾下生存结果的非政策评估与学习
链接：https://arxiv.org/abs/2603.22900
备注：Preprint
摘要：Optimizing survival outcomes, such as patient survival or customer retention, is a critical objective in data-driven decision-making. Off-Policy Evaluation~(OPE) provides a powerful framework for assessing such decision-making policies using logged data alone, without the need for costly or risky online experiments in high-stakes applications. However, typical estimators are not designed to handle right-censored survival outcomes, as they ignore unobserved survival times beyond the censoring time, leading to systematic underestimation of the true policy performance. To address this issue, we propose a novel framework for OPE and Off-Policy Learning~(OPL) tailored for survival outcomes under censoring. Specifically, we introduce IPCW-IPS and IPCW-DR, which employ the Inverse Probability of Censoring Weighting technique to explicitly deal with censoring bias. We theoretically establish that our estimators are unbiased and that IPCW-DR achieves double robustness, ensuring consistency if either the propensity score or the outcome model is correct. Furthermore, we extend this framework to constrained OPL to optimize policy value under budget constraints. We demonstrate the effectiveness of our proposed methods through simulation studies and illustrate their practical impacts using public real-world data for both evaluation and learning tasks.
【24】Probabilistic modeling over permutations using quantum computers
标题：使用量子计算机对排列进行概率建模
链接：https://arxiv.org/abs/2603.22401
备注：36 pages, 4 Figures
摘要：Quantum computers provide a super-exponential speedup for performing a Fourier transform over the symmetric group, an ability for which practical use cases have remained elusive so far. In this work, we leverage this ability to unlock spectral methods for machine learning over permutation-structured data, which appear in applications such as multi-object tracking and recommendation systems. It has been shown previously that a powerful way of building probabilistic models over permutations is to use the framework of non-Abelian harmonic analysis, as the model's group Fourier spectrum captures the interaction complexity: "low frequencies" correspond to low order correlations, and "high frequencies" to more complex ones. This can be used to construct a Markov chain model driven by alternating steps of diffusion (a group-equivariant convolution) and conditioning (a Bayesian update). However, this approach is computationally challenging and hence limited to simple approximations. Here we construct a quantum algorithm that encodes the exact probabilistic model -- a classically intractable object -- into the amplitudes of a quantum state by making use of the Quantum Fourier Transform (QFT) over the symmetric group. We discuss the scaling, limitations, and practical use of such an approach, which we envision to be a first step towards useful applications of non-Abelian QFTs.
其他(28篇)
【1】VISion On Request: Enhanced VLLM efficiency with sparse, dynamically selected, vision-language interactions
标题：Vision On Demand：通过稀疏的、动态选择的视觉语言交互来增强VLLM效率
链接：https://arxiv.org/abs/2603.23495
备注：Accepted at CVPR 2026
摘要：Existing approaches for improving the efficiency of Large Vision-Language Models (LVLMs) are largely based on the concept of visual token reduction. This approach, however, creates an information bottleneck that impairs performance, especially on challenging tasks that require fine-grained understanding and reasoning. In this work, we challenge this paradigm by introducing VISion On Request (VISOR), a method that reduces inference cost without discarding visual information. Instead of compressing the image, VISOR improves efficiency by sparsifying the interaction between image and text tokens. Specifically, the language model attends to the full set of high-resolution visual tokens through a small, strategically placed set of attention layers: general visual context is provided by efficient cross-attention between text-image, while a few well-placed and dynamically selected self-attention layers refine the visual representations themselves, enabling complex, high-resolution reasoning when needed. Based on this principle, we first train a single universal network on a range of computational budgets by varying the number of self-attention layers, and then introduce a lightweight policy mechanism that dynamically allocates visual computation based on per-sample complexity. Extensive experiments show that VISOR drastically reduces computational cost while matching or exceeding state-of-the-art results across a diverse suite of benchmarks, and excels in challenging tasks that require detailed visual understanding.
【2】End-to-End Efficient RL for Linear Bellman Complete MDPs with Deterministic Transitions
标题：具有确定性转变的线性Bellman完全MDP的端到端高效RL
链接：https://arxiv.org/abs/2603.23461
摘要：We study reinforcement learning (RL) with linear function approximation in Markov Decision Processes (MDPs) satisfying \emph{linear Bellman completeness} -- a fundamental setting where the Bellman backup of any linear value function remains linear. While statistically tractable, prior computationally efficient algorithms are either limited to small action spaces or require strong oracle assumptions over the feature space. We provide a computationally efficient algorithm for linear Bellman complete MDPs with \emph{deterministic transitions}, stochastic initial states, and stochastic rewards. For finite action spaces, our algorithm is end-to-end efficient; for large or infinite action spaces, we require only a standard argmax oracle over actions. Our algorithm learns an $\varepsilon$-optimal policy with sample and computational complexity polynomial in the horizon, feature dimension, and $1/\varepsilon$.
【3】GEM: Guided Expectation-Maximization for Behavior-Normalized Candidate Action Selection in Offline RL
标题：GEM：离线RL中行为规范化候选动作选择的引导期望最大化
链接：https://arxiv.org/abs/2603.23232
摘要：Offline reinforcement learning (RL) can fit strong value functions from fixed datasets, yet reliable deployment still hinges on the action selection interface used to query them. When the dataset induces a branched or multimodal action landscape, unimodal policy extraction can blur competing hypotheses and yield "in-between" actions that are weakly supported by data, making decisions brittle even with a strong critic. We introduce GEM (Guided Expectation-Maximization), an analytical framework that makes action selection both multimodal and explicitly controllable. GEM trains a Gaussian Mixture Model (GMM) actor via critic-guided, advantage-weighted EM-style updates that preserve distinct components while shifting probability mass toward high-value regions, and learns a tractable GMM behavior model to quantify support. During inference, GEM performs candidate-based selection: it generates a parallel candidate set and reranks actions using a conservative ensemble lower-confidence bound together with behavior-normalized support, where the behavior log-likelihood is standardized within each state's candidate set to yield stable, comparable control across states and candidate budgets. Empirically, GEM is competitive across D4RL benchmarks, and offers a simple inference-time budget knob (candidate count) that trades compute for decision quality without retraining.
【4】SpecXMaster Technical Report
标题：SpecXMaster技术报告
链接：https://arxiv.org/abs/2603.23101
备注：Technical report from DP Technology.21 pages, 5 figures
摘要：Intelligent spectroscopy serves as a pivotal element in AI-driven closed-loop scientific discovery, functioning as the critical bridge between matter structure and artificial intelligence. However, conventional expert-dependent spectral interpretation encounters substantial hurdles, including susceptibility to human bias and error, dependence on limited specialized expertise, and variability across interpreters. To address these challenges, we propose SpecXMaster, an intelligent framework leveraging Agentic Reinforcement Learning (RL) for NMR molecular spectral interpretation. SpecXMaster enables automated extraction of multiplicity information from both 1H and 13C spectra directly from raw FID (free induction decay) data. This end-to-end pipeline enables fully automated interpretation of NMR spectra into chemical structures. It demonstrates superior performance across multiple public NMR interpretation benchmarks and has been refined through iterative evaluations by professional chemical spectroscopists. We believe that SpecXMaster, as a novel methodological paradigm for spectral interpretation, will have a profound impact on the organic chemistry community.
【5】HUydra: Full-Range Lung CT Synthesis via Multiple HU Interval Generative Modelling
标题：HUydra：通过多HU间隔生成建模的全范围肺部CT合成
链接：https://arxiv.org/abs/2603.23041
备注：Submitted to iEEE TPAMI (Transactions on Pattern Analysis and Machine Intelligence)
摘要：Currently, a central challenge and bottleneck in the deployment and validation of computer-aided diagnosis (CAD) models within the field of medical imaging is data scarcity. For lung cancer, one of the most prevalent types worldwide, limited datasets can delay diagnosis and have an impact on patient outcome. Generative AI offers a promising solution for this issue, but dealing with the complex distribution of full Hounsfield Unit (HU) range lung CT scans is challenging and remains as a highly computationally demanding task. This paper introduces a novel decomposition strategy that synthesizes CT images one HU interval at a time, rather than modelling the entire HU domain at once. This framework focuses on training generative architectures on individual tissue-focused HU windows, then merges their output into a full-range scan via a learned reconstruction network that effectively reverses the HU-windowing process. We further propose multi-head and multi-decoder models to better capture textures while preserving anatomical consistency, with a multi-head VQVAE achieving the best performance for the generative task. Quantitative evaluation shows this approach significantly outperforms conventional 2D full-range baselines, achieving a 6.2% improvement in FID and superior MMD, Precision, and Recall across all HU intervals. The best performance is achieved by a multi-head VQVAE variant, demonstrating that it is possible to enhance visual fidelity and variability while also reducing model complexity and computational cost. This work establishes a new paradigm for structure-aware medical image synthesis, aligning generative modelling with clinical interpretation.
【6】Privacy-Preserving EHR Data Transformation via Geometric Operators: A Human-AI Co-Design Technical Report
标题
：通过几何运算符保护隐私的EHR数据转换：人机协同设计技术报告
链接：https://arxiv.org/abs/2603.22954
摘要：Electronic health records (EHRs) and other real-world clinical data are essential for clinical research, medical artificial intelligence, and life science, but their sharing is severely limited by privacy, governance, and interoperability constraints. These barriers create persistent data silos that hinder multi-center studies, large-scale model development, and broader biomedical discovery. Existing privacy-preserving approaches, including multi-party computation and related cryptographic techniques, provide strong protection but often introduce substantial computational overhead, reducing the efficiency of large-scale machine learning and foundation-model training. In addition, many such methods make data usable for restricted computation while leaving them effectively invisible to clinicians and researchers, limiting their value in workflows that still require direct inspection, exploratory analysis, and human interpretation. We propose a real-world-data transformation framework for privacy-preserving sharing of structured clinical records. Instead of converting data into opaque representations, our approach constructs transformed numeric views that preserve medical semantics and major statistical properties while, under a clearly specified threat model, provably breaking direct linkage between those views and protected patient-level attributes. Through collaboration between computer scientists and the AI agent \textbf{SciencePal}, acting as a constrained tool inventor under human guidance, we design three transformation operators that are non-reversible within this threat model, together with an additional mixing strategy for high-risk scenarios, supported by theoretical analysis and empirical evaluation under reconstruction, record linkage, membership inference, and attribute inference attacks.
【7】Confidence Calibration under Ambiguous Ground Truth
标题：模糊基本事实下的置信度校准
链接：https://arxiv.org/abs/2603.22879
摘要：Confidence calibration assumes a unique ground-truth label per input, yet this assumption fails wherever annotators genuinely disagree. Post-hoc calibrators fitted on majority-voted labels, the standard single-label targets used in practice, can appear well-calibrated under conventional evaluation yet remain substantially miscalibrated against the underlying annotator distribution. We show that this failure is structural: under simplifying assumptions, Temperature Scaling is biased toward temperatures that underestimate annotator uncertainty, with true-label miscalibration increasing monotonically with annotation entropy. To address this, we develop a family of ambiguity-aware post-hoc calibrators that optimise proper scoring rules against the full label distribution and require no model retraining. Our methods span progressively weaker annotation requirements: Dirichlet-Soft leverages the full annotator distribution and achieves the best overall calibration quality across settings; Monte Carlo Temperature Scaling with a single annotation per example (MCTS S=1) matches full-distribution calibration across all benchmarks, demonstrating that pre-aggregated label distributions are unnecessary; and Label-Smooth Temperature Scaling (LS-TS) operates with voted labels alone by constructing data-driven pseudo-soft targets from the model's own confidence. Experiments on four benchmarks with real multi-annotator distributions (CIFAR-10H, ChaosNLI) and clinically-informed synthetic annotations (ISIC~2019, DermaMNIST) show that Dirichlet-Soft reduces true-label ECE by 55-87% relative to Temperature Scaling, while LS-TS reduces ECE by 9-77% without any annotator data.
【8】The Coordinate System Problem in Persistent Structural Memory for Neural Architectures
标题：神经结构持久结构记忆中的坐标系问题
链接：https://arxiv.org/abs/2603.22858
摘要：We introduce the Dual-View Pheromone Pathway Network (DPPN), an architecture that routes sparse attention through a persistent pheromone field over latent slot transitions, and use it to discover two independent requirements for persistent structural memory in neural networks. Through five progressively refined experiments using up to 10 seeds per condition across 5 model variants and 4 transfer targets, we identify a core principle: persistent memory requires a stable coordinate system, and any coordinate system learned jointly with the model is inherently unstable. We characterize three obstacles -- pheromone saturation, surface-structure entanglement, and coordinate incompatibility -- and show that neither contrastive updates, multi-source distillation, Hungarian alignment, nor semantic decomposition resolves the instability when embeddings are learned from scratch. Fixed random Fourier features provide extrinsic coordinates that are stable, structure-blind, and informative, but coordinate stability alone is insufficient: routing-bias pheromone does not transfer (10 seeds, p>0.05). DPPN outperforms transformer and random sparse baselines for within-task learning (AULC 0.700 vs 0.680 vs 0.670). Replacing routing bias with learning-rate modulation eliminates negative transfer: warm pheromone as a learning-rate prior achieves +0.003 on same-family tasks (17 seeds, p<0.05) while never reducing performance. A structure completion function over extrinsic coordinates produces +0.006 same-family bonus beyond regularization, showing the catch-22 between stability and informativeness is partially permeable to learned functions. The contribution is two independent requirements for persistent structural memory: (a) coordinate stability and (b) graceful transfer mechanism.
【9】Towards The Implicit Bias on Multiclass Separable Data Under Norm Constraints
标题：规范约束下多类可分离数据的隐性偏差
链接：https://arxiv.org/abs/2603.22824
摘要 ：Implicit bias induced by gradient-based algorithms is essential to the generalization of overparameterized models, yet its mechanisms can be subtle. This work leverages the Normalized Steepest Descent} (NSD) framework to investigate how optimization geometry shapes solutions on multiclass separable data. We introduce NucGD, a geometry-aware optimizer designed to enforce low rank structures through nuclear norm constraints. Beyond the algorithm itself, we connect NucGD with emerging low-rank projection methods, providing a unified perspective. To enable scalable training, we derive an efficient SVD-free update rule via asynchronous power iteration. Furthermore, we empirically dissect the impact of stochastic optimization dynamics, characterizing how varying levels of gradient noise induced by mini-batch sampling and momentum modulate the convergence toward the expected maximum margin solutions.Our code is accessible at: https://github.com/Tsokarsic/observing-the-implicit-bias-on-multiclass-seperable-data.
【10】Combinatorial Privacy: Private Multi-Party Bitstream Grand Sum by Hiding in Birkhoff Polytopes
标题：组合隐私：隐藏在伯克霍夫多边形中的私人多方比特流总计
链接：https://arxiv.org/abs/2603.22808
摘要：We introduce PolyVeil, a protocol for private Boolean summation across $k$ clients that encodes private bits as permutation matrices in the Birkhoff polytope. A two-layer architecture gives the server perfect simulation-based security (statistical distance zero) while a separate aggregator faces \#P-hard likelihood inference via the permanent and mixed discriminant. Two variants (full and compressed) differ in what the aggregator observes. We develop a finite-sample $(\varepsilon,δ)$-DP analysis with explicit constants. In the full variant, where the aggregator sees a doubly stochastic matrix per client, the log-Lipschitz constant grows as $n^4 K_t$ and a signal-to-noise analysis shows the DP guarantee is non-vacuous only when the private signal is undetectable. In the compressed variant, where the aggregator sees a single scalar, the univariate density ratio yields non-vacuous $\varepsilon$ at moderate SNR, with the optimal decoy count balancing CLT accuracy against noise concentration. This exposes a fundamental tension. \#P-hardness requires the full matrix view (Birkhoff structure visible), while non-vacuous DP requires the scalar view (low dimensionality). Whether both hold simultaneously in one variant remains open. The protocol needs no PKI, has $O(k)$ communication, and outputs exact aggregates.
【11】Exposure-Normalized Bed and Chair Fall Rates via Continuous AI Monitoring
标题：通过持续人工智能监测暴露标准化床和椅子坠落率
链接：https://arxiv.org/abs/2603.22785
备注：23 pages, 6 figures
摘要：This retrospective cohort study used continuous AI monitoring to estimate fall rates by exposure time rather than occupied bed-days. From August 2024 to December 2025, 3,980 eligible monitoring units contributed 292,914 hourly rows, yielding probability-weighted rates of 17.8 falls per 1,000 chair exposure-hours and 4.3 per 1,000 bed exposure-hours. Within the study window, 43 adjudicated falls matched the monitoring pipeline, and 40 linked to eligible exposure hours for the primary Poisson model, producing an adjusted chair-versus-bed rate ratio of 2.35 (95% confidence interval 0.87 to 6.33; p=0.0907). In a separate broader observation cohort (n=32 deduplicated events), 6 of 7 direct chair falls involved footrest-positioning failures. Because this was an observational study in a single health system, these findings remain hypothesis-generating and support testing safer chair setups rather than using chairs less.
【12】Algorithmic warm starts for Hamiltonian Monte Carlo
标题：汉密尔顿蒙特卡洛的辉煌温暖开局
链接：https://arxiv.org/abs/2603.22741
摘要：Generating samples from a continuous probability density is a central algorithmic problem across statistics, engineering, and the sciences. For high-dimensional settings, Hamiltonian Monte Carlo (HMC) is the default algorithm across mainstream software packages. However, despite the extensive line of work on HMC and its widespread empirical success, it remains unclear how many iterations of HMC are required as a function of the dimension $d$. On one hand, a variety of results show that Metropolized HMC converges in $O(d^{1/4})$ iterations from a warm start close to stationarity. On the other hand, Metropolized HMC is significantly slower without a warm start, e.g., requiring $Ω(d^{1/2})$ iterations even for simple target distributions such as isotropic Gaussians. Finding a warm start is therefore the computational bottleneck for HMC. We resolve this issue for the well-studied setting of sampling from a probability distribution satisfying strong log-concavity (or isoperimetry) and third-order derivative bounds. We prove that \emph{non-Metropolized} HMC generates a warm start in $\tilde{O}(d^{1/4})$ iterations, after which we can exploit the warm start using Metropolized HMC. Our final complexity of $\tilde{O}(d^{1/4})$ is the fastest algorithm for high-accuracy sampling under these assumptions, improving over the prior best of $\tilde{O}(d^{1/2})$. This closes the long line of work on the dimensional complexity of MHMC for such settings, and also provides a simple warm-start prescription for practical implementations.
【13】MIOFlow 2.0: A unified framework for inferring cellular stochastic dynamics from single cell and spatial transcriptomics data
标题：MIOFlow 2.0：从单细胞和空间转录组学数据推断细胞随机动力学的统一框架
链接：https://arxiv.org/abs/2603.22564
摘要 ：Understanding cellular trajectories via time-resolved single-cell transcriptomics is vital for studying development, regeneration, and disease. A key challenge is inferring continuous trajectories from discrete snapshots. Biological complexity stems from stochastic cell fate decisions, temporal proliferation changes, and spatial environmental influences. Current methods often use deterministic interpolations treating cells in isolation, failing to capture the probabilistic branching, population shifts, and niche-dependent signaling driving real biological processes. We introduce Manifold Interpolating Optimal-Transport Flow (MIOFlow) 2.0. This framework learns biologically informed cellular trajectories by integrating manifold learning, optimal transport, and neural differential equations. It models three core processes: (1) stochasticity and branching via Neural Stochastic Differential Equations; (2) non-conservative population changes using a learned growth-rate model initialized with unbalanced optimal transport; and (3) environmental influence through a joint latent space unifying gene expression with spatial features like local cell type composition and signaling. By operating in a PHATE-distance matching autoencoder latent space, MIOFlow 2.0 ensures trajectories respect the data's intrinsic geometry. Empirical comparisons show expressive trajectory learning via neural differential equations outperforms existing generative models, including simulation-free flow matching. Validated on synthetic datasets, embryoid body differentiation, and spatially resolved axolotl brain regeneration, MIOFlow 2.0 improves trajectory accuracy and reveals hidden drivers of cellular transitions, like specific signaling niches. MIOFlow 2.0 thus bridges single-cell and spatial transcriptomics to uncover tissue-scale trajectories.
【14】Multimodal Training to Unimodal Deployment: Leveraging Unstructured Data During Training to Optimize Structured Data Only Deployment
标题：多模式训练到单模式部署：在训练期间利用非结构化数据来优化纯结构化数据部署
链接：https://arxiv.org/abs/2603.22530
备注：10 pages,3 figures
摘要：Unstructured Electronic Health Record (EHR) data, such as clinical notes, contain clinical contextual observations that are not directly reflected in structured data fields. This additional information can substantially improve model learning. However, due to their unstructured nature, these data are often unavailable or impractical to use when deploying a model. We introduce a multimodal learning framework that leverages unstructured EHR data during training while producing a model that can be deployed using only structured EHR data. Using a cohort of 3,466 children evaluated for late talking, we generated note embeddings with BioClinicalBERT and encoded structured embeddings from demographics and medical codes. A note-based teacher model and a structured-only student model were jointly trained using contrastive learning and contrastive knowledge distillation loss, producing a strong classifier (AUROC = 0.985). Our proposed model reached an AUROC of 0.705, outperforming the structured-only baseline of 0.656. These results demonstrate that incorporating unstructured data during training enhances the model's capacity to identify task-relevant information within structured EHR data, enabling a deployable structured-only phenotype model.
【15】mmFHE: mmWave Sensing with End-to-End Fully Homomorphic Encryption
标题：mmFHE：采用端到端全同质加密的mmWave感知
链接：https://arxiv.org/abs/2603.22437
备注：Under review
摘要：We present mmFHE, the first system that enables fully homomorphic encryption (FHE) for end-to-end mmWave radar sensing. mmFHE encrypts raw range profiles on a lightweight edge device and executes the entire mmWave signal-processing and ML inference pipeline homomorphically on an untrusted cloud that operates exclusively on ciphertexts. At the core of mmFHE is a library of seven composable, data-oblivious FHE kernels that replace standard DSP routines with fixed arithmetic circuits. These kernels can be flexibly composed into different application-specific pipelines. We demonstrate this approach on two representative tasks: vital-sign monitoring and gesture recognition. We formally prove two cryptographic guarantees for any pipeline assembled from this library: input privacy, the cloud learns nothing about the sensor data; and data obliviousness, the execution trace is identical on the cloud regardless of the data being processed. These guarantees effectively neutralize various supervised and unsupervised privacy attacks on raw data, including re-identification and data-dependent privacy leakage. Evaluation on three public radar datasets (270 vital-sign recordings, 600 gesture trials) shows that encryption introduces negligible error: HR/RR MAE <10^-3 bpm versus plaintext, and 84.5% gesture accuracy (vs. 84.7% plaintext) with end-to-end cloud GPU latency of 103s for a 10s vital-sign window and 37s for a 3s gesture window. These results show that privacy-preserving end-to-end mmWave sensing is feasible on commodity hardware today.
【16】Instruction-Tuned, but Not More Verifiable Instruction-Following: A Cross-Task Diagnosis for LoRA Adapters
标题：经过指令调整，但不是更具可验证的指令遵循：LoRA适配器的交叉任务诊断
链接：https://arxiv.org/abs/2603.22379
备注：12 pages, 5 figures, 6 tables
摘要：Adapters are often selected and deployed based on nominal labels (e.g., instruction-tuned), which implicitly suggest what capability improves after adaptation. We test whether nominal training objectives reliably align with realized cross-task capability gains by evaluating the same LoRA adapter across tasks. Our strongest evidence is tied to strict, automatically verifiable instruction following as measured by IFEval: across multiple seeds, base models, and LoRA settings, nominal labels recurrently but not universally fail to predict improvements on this verifiable target, with clear configuration sensitivity including a near-zero or negative case. As an illustrative strongest-case example in a controlled instruction-versus-numeric setting, an instruction-tuned adapter substantially improves off-target NM-based numeric benchmark performance from 0.133 to 0.632 while not improving verifiable instruction following on IFEval (ILA: 0.313 to 0.271; PLA: 0.250 to 0.143; values rounded to three decimals). We refer to this nominal-versus-realized mismatch pattern as capability drift as a descriptive label. The mismatch is visible in the raw cross-task performance matrix; we use a drift score only as a compact summary in the same units as the underlying metrics, not as a new formal metric contribution. Evidence from broader instruction-following benchmarks is benchmark-dependent and mixed, reflecting heterogeneity in how instruction following is operationalized; we therefore do not treat cross-benchmark agreement as a premise. Overall, the practical takeaway is to perform routine cross-task evaluation before deployment and to avoid treating nominal labels as reliable capability proxies.
【17】Three Creates All: You Only Sample 3 Steps
标题：三创建全部：您只需尝试3个步骤
链接：https://arxiv.org/abs/2603.22375
摘要 ：Diffusion models deliver high-fidelity generation but remain slow at inference time due to many sequential network evaluations. We find that standard timestep conditioning becomes a key bottleneck for few-step sampling. Motivated by layer-dependent denoising dynamics, we propose Multi-layer Time Embedding Optimization (MTEO), which freeze the pretrained diffusion backbone and distill a small set of step-wise, layer-wise time embeddings from reference trajectories. MTEO is plug-and-play with existing ODE solvers, adds no inference-time overhead, and trains only a tiny fraction of parameters. Extensive experiments across diverse datasets and backbones show state-of-the-art performance in the few-step sampling and substantially narrow the gap between distillation-based and lightweight methods. Code will be available.
【18】Rethinking Multimodal Fusion for Time Series: Auxiliary Modalities Need Constrained Fusion
标题：重新思考时间序列的多模式融合：辅助模式需要约束融合
链接：https://arxiv.org/abs/2603.22372
摘要：Recent advances in multimodal learning have motivated the integration of auxiliary modalities such as text or vision into time series (TS) forecasting. However, most existing methods provide limited gains, often improving performance only in specific datasets or relying on architecture-specific designs that limit generalization. In this paper, we show that multimodal models with naive fusion strategies (e.g., simple addition or concatenation) often underperform unimodal TS models, which we attribute to the uncontrolled integration of auxiliary modalities which may introduce irrelevant information. Motivated by this observation, we explore various constrained fusion methods designed to control such integration and find that they consistently outperform naive fusion methods. Furthermore, we propose Controlled Fusion Adapter (CFA), a simple plug-in method that enables controlled cross-modal interactions without modifying the TS backbone, integrating only relevant textual information aligned with TS dynamics. CFA employs low-rank adapters to filter irrelevant textual information before fusing it into temporal representations. We conduct over 20K experiments across various datasets and TS/text models, demonstrating the effectiveness of the constrained fusion methods including CFA. Code is publicly available at: https://github.com/seunghan96/cfa/.
【19】Intelligence Inertia: Physical Principles and Applications
标题：智力惯性：物理原理与应用
链接：https://arxiv.org/abs/2603.22347
备注：53 pages, 9 figures
摘要：While Landauer's principle establishes the fundamental thermodynamic floor for information erasure and Fisher Information provides a metric for local curvature in parameter space, these classical frameworks function effectively only as approximations within regimes of sparse rule-constraints. They fail to explain the super-linear, and often explosive, computational and energy costs incurred when maintaining symbolic interpretability during the reconfiguration of advanced intelligent systems. This paper introduces the property of intelligence inertia and its underlying physical principles as foundational characteristics for quantifying the computational weight of intelligence. We demonstrate that this phenomenon is not merely an empirical observation but originates from the fundamental non-commutativity between rules and states, a root cause we have formally organized into a rigorous mathematical framework. By analyzing the growing discrepancy between actual adaptation costs and static information-theoretic estimates, we derive a non-linear cost formula that mirrors the Lorentz factor, characterizing a relativistic J-shaped inflation curve -- a "computational wall" that static models are blind to. The validity of these physical principles is examined through a trilogy of decisive experiments: (1) a comparative adjudication of this J-curve inflation against classical Fisher Information models, (2) a geometric analysis of the "Zig-Zag" trajectory of neural architecture evolution, and (3) the implementation of an inertia-aware scheduler wrapper that optimizes the training of deep networks by respecting the agent's physical resistance to change. Our results suggest a unified physical description for the cost of structural adaptation, offering a first-principle explanation for the computational and interpretability-maintenance overhead in intelligent agents.
【20】Problems with Chinchilla Approach 2: Systematic Biases in IsoFLOP Parabola Fits
标题：龙猫方法2的问题：IsoFLOP Parabola Fits中的系统偏差
链接：https://arxiv.org/abs/2603.22339
摘要：Chinchilla Approach 2 is among the most widely used methods for fitting neural scaling laws. Its parabolic approximation introduces systematic biases in compute-optimal allocation estimates, even on noise-free synthetic data. Applied to published Llama 3 IsoFLOP data at open frontier compute scales, these biases imply a parameter underallocation corresponding to 6.5% of the $3.8\times10^{25}$ FLOP training budget and \$1.4M (90% CI: \$412K-\$2.9M) in unnecessary compute at 50% H100 MFU. Simulated multimodal model misallocations show even greater opportunity costs due to higher loss surface asymmetry. Three sources of this error are examined: IsoFLOP sampling grid width (Taylor approximation accuracy), uncentered IsoFLOP sampling, and loss surface asymmetry ($α\neq β$). Chinchilla Approach 3 largely eliminates these biases but is often regarded as less data-efficient, numerically unstable, prone to local minima, and harder to implement. Each concern is shown to be unfounded or addressable, especially when the partially linear structure of the objective is exploited via Variable Projection, enabling unbiased inference on all five loss surface parameters through a two-dimensional optimization that is well-conditioned, analytically differentiable, and amenable to dense, or even exhaustive, grid search. It may serve as a more convenient replacement for Approach 2 or a more scalable alternative for adaptations of Approach 3 to richer scaling law formulations.
【21】Beyond the Mean: Distribution-Aware Loss Functions for Bimodal Regression
标题：超越平均值：双峰回归的分布感知损失函数
链接：https://arxiv.org/abs/2603.22328
备注：28 pages, 27 figures
摘要 ：Despite the strong predictive performance achieved by machine learning models across many application domains, assessing their trustworthiness through reliable estimates of predictive confidence remains a critical challenge. This issue arises in scenarios where the likelihood of error inferred from learned representations follows a bimodal distribution, resulting from the coexistence of confident and ambiguous predictions. Standard regression approaches often struggle to adequately express this predictive uncertainty, as they implicitly assume unimodal Gaussian noise, leading to mean-collapse behavior in such settings. Although Mixture Density Networks (MDNs) can represent different distributions, they suffer from severe optimization instability. We propose a family of distribution-aware loss functions integrating normalized RMSE with Wasserstein and Cramér distances. When applied to standard deep regression models, our approach recovers bimodal distributions without the volatility of mixture models. Validated across four experimental stages, our results show that the proposed Wasserstein loss establishes a new Pareto efficiency frontier: matching the stability of standard regression losses like MSE in unimodal tasks while reducing Jensen-Shannon Divergence by 45% on complex bimodal datasets. Our framework strictly dominates MDNs in both fidelity and robustness, offering a reliable tool for aleatoric uncertainty estimation in trustworthy AI systems.
【22】Hybrid Associative Memories
标题：混合联想记忆
链接：https://arxiv.org/abs/2603.22325
备注：30 pages, 10 figures
摘要：Recurrent neural networks (RNNs) and self-attention are both widely used sequence-mixing layers that maintain an internal memory. However, this memory is constructed using two orthogonal mechanisms: RNNs compress the entire past into a fixed-size state, whereas self-attention's state stores every past time step growing its state (the KV cache) linearly with the sequence length. This results in orthogonal strengths and weaknesses. Self-attention layers excel at retrieving information in the context but have large memory and computational costs, while RNNs are more efficient but degrade over longer contexts and underperform for precise recall tasks. Prior work combining these mechanisms has focused primarily on naively interleaving them to reduce computational cost without regard to their complementary mechanisms. We propose the Hybrid Associative Memory (HAM) layer, which combines self-attention and RNNs while leveraging their individual strengths: the RNN compresses the entire sequence, while attention supplements it *only* with information that is difficult for the RNN to predict, which is hence the most valuable information to explicitly store. HAM layers enable data-dependent growth of the KV cache, which can be precisely controlled by the user with a single, continuous threshold. We find that this fine-grained control of the KV cache growth rate has a smooth trade-off with loss and performance. Empirically, we show that our hybrid architecture offers strong, competitive performance relative to RNNs and Transformers even at substantially lower KV-cache usage.
【23】ST-GDance++: A Scalable Spatial-Temporal Diffusion for Long-Duration Group Choreography
标题：ST-GDance++：长时间团体编舞的可扩展时空扩散
链接：https://arxiv.org/abs/2603.22316
摘要：Group dance generation from music requires synchronizing multiple dancers while maintaining spatial coordination, making it highly relevant to applications such as film production, gaming, and animation. Recent group dance generation models have achieved promising generation quality, but they remain difficult to deploy in interactive scenarios due to bidirectional attention dependencies. As the number of dancers and the sequence length increase, the attention computation required for aligning music conditions with motion sequences grows quadratically, leading to reduced efficiency and increased risk of motion collisions. Effectively modeling dense spatial-temporal interactions is therefore essential, yet existing methods often struggle to capture such complexity, resulting in limited scalability and unstable multi-dancer coordination. To address these challenges, we propose ST-GDance++, a scalable framework that decouples spatial and temporal dependencies to enable efficient and collision-aware group choreography generation. For spatial modeling, we introduce lightweight distance-aware graph convolutions to capture inter-dancer relationships while reducing computational overhead. For temporal modeling, we design a diffusion noise scheduling strategy together with an efficient temporal-aligned attention mask, enabling stream-based generation for long motion sequences and improving scalability in long-duration scenarios. Experiments on the AIOZ-GDance dataset show that ST-GDance++ achieves competitive generation quality with significantly reduced latency compared to existing methods.
【24】The Efficiency Attenuation Phenomenon: A Computational Challenge to the Language of Thought Hypothesis
标题：效率衰减现象：对思维语言假设的计算挑战
链接：https://arxiv.org/abs/2603.22312
备注：11 pages
摘要：This paper computationally investigates whether thought requires a language-like format, as posited by the Language of Thought (LoT) hypothesis. We introduce the ``AI Private Language'' thought experiment: if two artificial agents develop an efficient, inscrutable communication protocol via multi-agent reinforcement learning (MARL), and their performance declines when forced to use a human-comprehensible language, this Efficiency Attenuation Phenomenon (EAP) challenges the LoT. We formalize this in a cooperative navigation task under partial observability. Results show that agents with an emergent protocol achieve 50.5\% higher efficiency than those using a pre-defined, human-like symbolic protocol, confirming the EAP. This suggests optimal collaborative cognition in these systems is not mediated by symbolic structures but is naturally coupled with sub-symbolic computations. The work bridges philosophy, cognitive science, and AI, arguing for pluralism in cognitive architectures and highlighting implications for AI ethics.
【25】Mitigating Premature Discretization with Progressive Quantization for Robust Vector Tokenization
标题：通过渐进量化缓解过早离散化以实现鲁棒的载体令牌化
链接：https://arxiv.org/abs/2603.22304
摘要 ：Vector Quantization (VQ) has become the cornerstone of tokenization for many multimodal Large Language Models and diffusion synthesis. However, existing VQ paradigms suffer from a fundamental conflict: they enforce discretization before the encoder has captured the underlying data manifold. We term this phenomenon Premature Discretization. To resolve this, we propose Progressive Quantization (ProVQ), which incorporates the dynamics of quantization hardness as a fundamental yet previously overlooked axis in VQ training. By treating quantization as a curriculum that smoothly anneals from a continuous latent space to a discrete one, ProVQ effectively guides the codebook toward the well-expanded manifolds. Extensive experimental results demonstrate the broad effectiveness of ProVQ across diverse modalities. We report improved reconstruction and generative performance on the ImageNet-1K and ImageNet-100 benchmarks, highlighting the ProVQ's boost for generative modeling. Furthermore, ProVQ proves highly effective for modeling complex biological sequences, establishing a new performance ceiling for protein structure tokenization on the StrutTokenBench leaderboard.
【26】Scaling Attention via Feature Sparsity
标题：通过特征稀疏性增加注意力
链接：https://arxiv.org/abs/2603.22300
备注：26 pages, 11 figures; Accepted at ICLR 2026
摘要：Scaling Transformers to ultra-long contexts is bottlenecked by the $O(n^2 d)$ cost of self-attention. Existing methods reduce this cost along the sequence axis through local windows, kernel approximations, or token-level sparsity, but these approaches consistently degrade accuracy. In this paper, we instead explore an orthogonal axis: feature sparsity. We propose Sparse Feature Attention (SFA), where queries and keys are represented as $k$-sparse codes that preserve high-dimensional expressivity while reducing the cost of attention from $Θ(n^2 d)$ to $Θ(n^2 k^2/d)$. To make this efficient at scale, we introduce FlashSFA, an IO-aware kernel that extends FlashAttention to operate directly on sparse overlaps without materializing dense score matrices. Across GPT-2 and Qwen3 pretraining, SFA matches dense baselines while improving speed by up to $2.5\times$ and reducing FLOPs and KV-cache by nearly 50\%. On synthetic and downstream benchmarks, SFA preserves retrieval accuracy and robustness at long contexts, outperforming short-embedding baselines that collapse feature diversity. These results establish feature-level sparsity as a complementary and underexplored axis for efficient attention, enabling Transformers to scale to orders-of-magnitude longer contexts with minimal quality loss. Code is available at https://github.com/YannX1e/Sparse-Feature-Attention.
【27】Generative Inversion of Spectroscopic Data for Amorphous Structure Elucidation
标题：非晶体结构解析光谱数据的生成倒置
链接：https://arxiv.org/abs/2603.23210
备注：10 pages; SI: 51 pages
摘要：Determining atomistic structures from characterization data is one of the most common yet intricate problems in materials science. Particularly in amorphous materials, proposing structures that balance realism and agreement with experiments requires expert guidance, good interatomic potentials, or both. Here, we introduce GLASS, a generative framework that inverts multi-modal spectroscopic measurements into realistic atomistic structures without knowledge of the potential energy surface. A score-based model learns a structural prior from low-fidelity data and samples out-of-distribution structures conditioned on differentiable spectral targets. Reconstructions using pair distribution functions (PDFs), X-ray absorption spectroscopy, and diffraction measurements quantify the complementarity between spectral modalities and demonstrate that PDFs is the most informative probe for our framework. We use GLASS to rationalize three contested experimental problems: paracrystallinity in amorphous silicon, a liquid-liquid phase transition in sulfur, and ball-milled amorphous ice. In each case, generated structures reproduce experimental measurements and reveal mechanisms inaccessible to diffraction analysis alone.
【28】SPDE Methods for Nonparametric Bayesian Posterior Contraction and Laplace Approximation
标题：非参数Bayesian后验压缩和拉普拉斯逼近的SPED方法
链接：https://arxiv.org/abs/2603.22468
备注：32 pages, under review
摘要：We derive posterior contraction rates (PCRs) and finite-sample Bernstein von Mises (BvM) results for non-parametric Bayesian models by extending the diffusion-based framework of Mou et al. (2024) to the infinite-dimensional setting. The posterior is represented as the invariant measure of a Langevin stochastic partial differential equation (SPDE) on a separable Hilbert space, which allows us to control posterior moments and obtain non-asymptotic concentration rates in Hilbert norms under various likelihood curvature and regularity conditions. We also establish a quantitative Laplace approximation for the posterior. The theory is illustrated in a nonparametric linear Gaussian inverse problem.
机器翻译由腾讯交互翻译提供，仅供参考
点击“阅读原文”获取带摘要的学术速递
Python社区是高质量的Python/Django开发社区
本文地址：http://www.python88.com/topic/194344
本文地址：http://www.python88.com/topic/194344