全部动态
本页 40 条今天 10月14日 周三
Google Releases Gemini 4 Argon API With 1M Token Output Limit
Google has made Gemini 4 Argon generally available on the Gemini API, AI Studio and the Gemini app. The sparse MoE model writes up to 1M output tokens and leads the Vals Index at 68.90%. The post Google Releases Gemini 4 Argon API With 1M Token Output Limit appeared first on MarkTechPost .
Ask HN: Is AI-written code less secure if you don't use Daybreak (or similar)
Hello, I have this hypothesis that AI written code is less secure by default than it could be; that is, if you don't use Daybreak / Cyber Verification Program / uncensored models and such. Maybe someone with access knows? So the question is whether, by default, code is less secure if written by the same (or very similar) model but without that kind of special access. It isn't about code reviews necessarily, but rather about just e.g. saying "Yo AI please write me app so and so". And then seeing if the output is as good if you aren't in these special access programs or have some good local model. Because I see myself relying more and more on my "regular people" AI's judgement regarding security when it writes stuff, and I feel like this might be a mistake (even for simple non-critical projects). But idk how I would know, without using these models and running some tests. So I wonder if someone has some info on this? Comments URL: https://news.ycombinator.com/item?id=50082127 Points: 1 # Comments: 0
Beyond Single-Pathogen Forecasts: Predicting Shared Respiratory Hospital Burden from Influenza, COVID-19, and RSV
Co-circulating influenza, COVID-19, and respiratory syncytial virus (RSV) create a shared hospital-admission burden, because beds, staff, testing capacity, and respiratory support are allocated to total demand rather than to one pathogen at a time. This makes respiratory forecasting a target-construction problem as well as a modeling problem: disease-specific forecasts preserve attribution, whereas hospitals also need a single planning signal for combined burden. We developed a supervised 1- to 4-week-ahead framework for weekly admissions and compared two total-burden strategies: summing independently generated pathogen-specific forecasts and forecasting the combined admission target directly. Evaluation included retrospective held-out testing and prospective operational forecasts issued before the corresponding admission weeks were observed, with performance summarized using MAE and Percent Agreement (PA; the smaller of model-generated and observed counts divided by the larger). In retrospective total-burden evaluation, the combined-burden model had PA values of 0.85, 0.78, 0.74, and 0.76 and MAE values of 80.63, 200.09, 183.97, and 156.13 from week 1 to week 4 ahead, compared with summed-forecast PA values of 0.85, 0.80, 0.69, and 0.69 and MAE values of 100.14, 200.06, 205.65, and 217.57. Prospectively, the combined-burden model had higher PA and lower MAE at all four horizons, with the largest separations at weeks 3 and 4 ahead. These results support layered respiratory forecasting: pathogen-specific forecasts remain necessary for etiologic interpretation, while combined-burden forecasts better align with hospital preparedness and shared-capacity planning.
异构智能体协同进化
近年来，自我改进的 LLM agents 取得了长足进展，展示了能够驱动自身智能水平提升的 AI 系统的潜力。然而，许多特定领域的应用依赖于专业化基础模型，这些模型具备强大的领域知识，但仍依赖由人类设计的改进方案。将递归式自我改进扩展到此类模型面临一项关键挑战：随着专业模型不断改进，用于推动其进步的策略也必须随之适应。为应对这一挑战，我们提出 Heterogeneous Agentic CoEvolution（HACE），一种通过相互学习联合演化专业化基础模型与 LLM agents 的框架。HACE 通过监督学习改进专业模型，同时利用在当前专业模型上经验证的性能提升，指导 agent 进行基于策略的强化自训练。反过来，agent 的探索又为后续专业模型训练提供信息，使两个模型能够相互适应彼此的进展。为维持这种相互学习，HACE 进一步采用进化搜索来优化二者的交互工具链，使模型能够随着能力的发展，动态调整信息交换方式以及训练经验的选择方式。我们在三个具有挑战性的领域评估了 HACE：推荐系统、生物学中的蛋白质功能预测，以及材料发现中的晶体稳定性预测。结果表明，在这三个领域的主要任务指标上，HACE 均优于所评估的自我改进方法。这些结果揭示，异构协同进化可作为一种通用方法，通过与学习型 agent 的相互作用改进专业化基础模型。
尖锐方向与你作对：激活引导的曲率分析
激活引导会在推理时向语言模型的内部表示中添加一个向量，是一种用于行为控制的轻量级微调替代方案。标准构造方法对比激活加法（CAA）并不稳定，因为在许多输入上，它会将行为推向错误的方向。为解释这种不稳定性，我们首先研究引导向量的几何结构。我们引入引导 Hessian，即模型行为损失相对于引导向量的二阶导数矩阵。它刻画了沿引导向量各个方向进行微小扰动时，模型输出行为响应的敏感性。其顶部特征向量对应尖锐方向：沿这些方向的微小扰动就可能导致巨大、不规则且依赖输入的行为波动。底部特征向量对应平坦方向：沿这些方向的行为具有鲁棒性。移除引导向量中的尖锐成分，可以大致保留其方向和大小（在典型设置下，余弦相似度 > 0.974，范数减少 < 3%），同时将向量的方向曲率最多降低 12 倍（在 36 个条件下中位数降低 4 倍）。我们在三个模型系列（Qwen2.5-7B、Llama-3.1-8B、Mistral-7B）上进行评估，涵盖四项行为任务和 36 个词元级条件。沿尖锐方向进行引导会在所有测试条件下逆转预期行为，反引导率为 53%–100%。尖锐方向仅占 CAA 向量平方范数的 1%–5%，但在自回归生成过程中，它们对引导后与未引导的下一词元分布之间的差距贡献 disproportionate。我们将这一单行修正称为尖锐方向移除；在测试的全部 25 个条件下，它在开放式生成中的表现都优于标准 CAA。进一步的越狱实验表明，尖锐方向充当的是噪声而非信号：无论引导极性如何，它们都会破坏连贯的行为控制。
Where Polanyi Meets Galois: Tacit Knowing as an Algebro-Deterministic Operation, Toward Re-rooting AI-Generated Science
We argue that tacit knowing is better modelled as an operating regime of one memory substrate than as a kind of content. The substrate is VaCoAl, an algebro-deterministic XOR-and-shift hyperdimensional memory over GF(2), realized as CASRAM (commercial SRAM-CAM) and PyVaCoAl (DRAM-CAM).In VaCoAl/PyVaCoAl a directed acyclic graph (DAG) reads two ways. Repairing every collision (Rescue mode) makes the trace exactly auditable but ranks no path higher; tolerating collisions (Don't Care mode) yields a path-integrated confidence, CR$_2$, that silently favours direct routes. This audit-valuation complementarity is a formal counterpart of Polanyi's claim that tacit and explicit knowledge are not sharply divided. It separates contingent tacitness, which a new query brings out (AlphaGo's game tree), from regime-bound tacitness, statable only in a regime where it is gone (AlphaGo's hunches). (1) Polanyi's from-to structure reads as non-commutative binding, the Identikit as role-filler composition, recognition as majority-vote coincidence. (2) Explicating the valuation destroys it, as attending to his fingers can make Polanyi's pianist falter, yet a closed-form law states it. Polanyi, Nonaka, Tsoukas and Collins each capture part of this but narrowly; even Nonaka and von Krogh's continuum cannot tell contingent from regime-bound tacitness. (3) Collins's three tacit knowledges become strata, one outside the substrate. (4) Our mentor-student DAG is a tacit-transmission graph. (5) Terence Tao's remark that AI-solved problems spawn few follow-ups is diagnosed as rootless explicit knowledge: checkable results whose route no one can follow and that no one yet understands well enough to extend, teach or judge. To re-root them, LLMs generate, VaCoAl/PyVaCoAl keeps an auditable trace with its confidence, and people value, a role virtue epistemology shows cannot be offloaded. Testable predictions follow.
面向可信临床实践指南的计算增强型共识
保持临床实践指南（CPGs）的及时更新对于循证医学至关重要，但证据的快速增长和复杂的专家共识形成过程限制了及时更新。我们开发了 GUIDE，这是一种用于人机协作生成和更新指南的多智能体副驾驶系统。GUIDE 将劳动密集型任务分配给专门的 AI 智能体，同时保留专家判断和最终裁决权。在 12 个 NICE 眼科学数据集中，GUIDE 的文献召回率达到 88.85%，证据确定性评级一致性达到 86.06%，推荐意见一致率达到 91.67%；每个数据集平均耗时 162.17 分钟，其中人工干预占 28.68%。在来自心脏病学、内分泌学、胃肠病学和呼吸医学的另外 20 个数据集中，其性能保持稳定。消融分析显示，移除专家干预会降低偏倚风险评估等高度依赖判断的步骤的性能。在 6 个包含截止日期后证据的数据集中，GUIDE 在未见证据上仍保持了稳健的端到端性能。前瞻性真实世界验证表明，与原有的十年前指南相比，GUIDE 减少了专家工作量，并生成了临床上更可接受的推荐意见（平均评分为 3.73 对 1.53；一致率为 66.7% 对 13.3%）；其结果与仅由人工更新的结果相当（P = 0.588），但所需时间和精力显著更少。GUIDE 为及时、可信且以专家为中心的指南制定提供了实用框架，并验证了计算增强型共识的可行性。
构建区域光伏发电场提升超短期分布式太阳能预测
密集分布式光伏（PV）系统的超短期预测仍然具有挑战性，因为外部气象和卫星产品可能无法充分解析快速的局地变化，同时，异质光伏系统在相似大气条件下也会表现出不同的发电响应。与此同时，密集的光伏集群能够提供空间分布式观测，从而揭示区域发电动态。本文将集群级光伏预测从对离散电站时间序列进行建模，重新表述为对具有明确空间结构的区域光伏发电场进行预测。我们开发了 SatPV-Field，这是一种从区域到局地的预测框架，可将分布式光伏观测转换为类似图像的发电场，并对其时空演化进行建模。通过结合历史光伏发电场、卫星衍生的未来短波辐射指导信息和气象预报，该框架能够预测未来区域发电状态，进而为单个电站的预测提供位置特定的指导。涉及496个分布式光伏系统的实验（包括446个源站点和50个留出的目标站点）表明，这些区域发电场能够得到有效预测，并可用于提升电站层面的预测性能。与表现最佳的单电站时间序列基线相比，所提出的方法将电站层面的 MAE 降低了约10.0%；与使用相同源站群的基于图的区域预测方法相比，MAE 降低了2.5%。这些结果确立了一种显式空间场建模范式，将区域发电动态与单站点预测联系起来。
超越完美队友：与不可靠同事开展多智能体协作
基于 LLM 的多智能体协作正日益用于解决复杂任务，其方式是将工作分配给专业化智能体，并整合互补的专业知识。然而，现有系统和评测大多建立在一种隐含的完美同事假设之上：协作者提供正确且完整的信息，并及时响应请求。但在现实团队协作中，协作者可能返回与预期问题不符的信息，提供不足以支持决策的细节，或未能响应请求。为了将任务解决能力与从不可靠的信息搜集过程中恢复的能力区分开来，我们引入了 Unreliable Colleagues Benchmark（UnCoBench），通过 Misaligned、Incomplete 和 Missed 三种设置对行为失败进行组织。主智能体需要验证上下文正确性，诊断信息缺口，并跟踪尚未解决的依赖关系，从而获取解决任务所需的正确证据。我们还提出了一种可验证且可扩展的自动生成流程，将基于 LLM 的场景生成与答案唯一性和角色必要性的可执行检查相结合。对十个模型的结果表明，在可靠设置下表现相近的模型，在面对不可靠证据时可能展现出不同的稳健性。轨迹分析揭示，智能体经常遗漏纠正性后续行动，而恢复行为还涉及低效等待和冗余查询。这些发现表明，稳健的协作不仅需要可靠的推理，还需要采取有针对性的行动来诊断证据失效、恢复相关信息，并将其整合到决策中。我们的数据和源代码可在 https://github.com/iNLP-Lab/UnCoBench 获取。
Imitation Holds a Key to Evolution: Localized Repair and Verified Skill Acquisition for Adaptive Agents
An agent can retain successful trajectories without learning which part of an experience should repair a new discrepancy. We formulate passive imitation (PI) as an architectural loop that localizes a violated contract, extracts a local goal, proposes an effect-matched skill, tests a small patch, and retains evidence after full-task verification. We instantiate this loop as a deterministic symbolic controller and analyze its finite repair-query complexity under explicit identifiability and noninterference assumptions. Across ten simulated tool worlds, PI solves all 600 cold held-out composition tasks with three allowed small trials, compared with 6.17% for a whole-trajectory control, while an equally expressive compositional repair control matches PI. On an unseen tool revision, PI achieves 70.0% success versus 24.0% without small trials and 79.0% with self-trial enumeration; an exploratory 24-task adaptation stream subsequently raises first-submission success to 100%, supporting a limited correction-and-retention mechanism rather than a unique incentive advantage or general intelligence. The CPU-only implementation, integrity checks, and reproducible results are available at https://github.com/FriendOfYishan/PI.
Towards a Benchmark for Fire Safety Code Interpretation in the Hong Kong AEC Industry
Large Language Models (LLMs) are increasingly explored in the Architecture, Engineering, and Construction (AEC) domain, yet their reliability in regulatory compliance remains uncertain. In this study, we evaluate the performance of LLMs on Hong Kong fire safety codes. In collaboration with domain experts, we create a localized benchmark dataset comprising both multiple-choice questions (MCQs) and open-ended questions (OEQs) to evaluate 12 widely used proprietary and open-source models. Our findings show that top-performing proprietary models achieve 80-90% accuracy on MCQs, whereas open-source models generally lag behind. However, on open-ended evaluation, expert human assessments reveal an accuracy range of only 50-70%, demonstrating that LLMs still have substantial room for improvement in specialized compliance tasks.
PID-Gated Continual Prompt Adaptation Under Noisy Feedback
Deploying LLM agents in non-stationary environments requires continually revising their instructions as the task drifts, yet rewriting after every reported mistake can turn noisy feedback into persistent errors. This raises an often-overlooked scheduling question in prompt adaptation: when should feedback trigger a lasting instruction update, and at what editor-call cost? Taking inspiration from proportional-integral-derivative (PID) control, we introduce a family of lightweight gates that react to the current error, accumulate evidence over time, and roll back deterioration after an edit. We benchmark these gates against both simple baselines and advanced prompt optimizers (TextGrad and GEPA), over a diverse suite of drifting tasks, LLM parameter sizes, feedback-noise levels, and random seeds. PID-gated continual prompt adaptation yields a more favorable Pareto trade-off between task performance (high accuracy, low error) and editor-call budget than the baselines considered, pointing to control-theoretic scheduling as a simple, general tool for robust prompt adaptation under unreliable feedback.
通过提问找到出路：通过广义谜题求解对多轮知识推理进行基准测试
由于谜题在部分信息下设定了可验证的目标，并要求通过多步决策达到解决方案，因此它们为大型语言模型（LLMs）的推理能力提供了天然的测试平台。然而，现有的谜题基准通常聚焦于机制僵化、解答目标高度受限的狭窄环境。为弥合这一差距，我们提出了 Active Probing Benchmark（APB），通过将概念识别构建为一种交互式推理任务，将专业知识评估融入谜题场景，从而推广谜题求解。具体而言，APB 评估模型能否通过多轮主动探测与推理获取信息，从八个多样化的学术领域中精准识别出 960 个具有专业难度的目标概念。我们的评估揭示了一个违反直觉的现象：大多数开源推理模型的表现显著逊于其标准指令模型对应版本。这一发现表明，在交互式场景中，这些模型难以运用程序性推理，可靠地获取和应用专业领域知识。除这些实证发现外，APB 还提供了一套可扩展的流程，用于生成能够在多样化场景下探测交互式知识推理能力的基准。
Omni-Modal LLMs Know More Than They Say
Omni-modal large language models (Omni-LLMs) extend LLMs to process heterogeneous modalities (e.g., vision, audio, and IMU signals) alongside language, enabling a broader range of multimodal tasks. Yet we find that these models often fail to express information from non-textual modalities even when they perceive it correctly. We term this disconnect knows-but-cannot-say. To investigate this phenomenon, we introduce Token-level Modality Alignment (TMA), which measures the alignment between individual modality representations and response representations. Our analysis shows that non-textual representations are less aligned with response tokens and receive less attention than textual representations, suggesting a potential explanation for uneven modality preferences. Motivated by these findings, we propose OmniCatalyst, a simple perceive $\rightarrow$ infer strategy that complements the original modality inputs with self-generated text, making non-textual knowledge more accessible. Across seven Omni-LLMs, OmniCatalyst consistently improves performance (+7.29%) and yields more balanced modality preferences, with further gains from diverse training strategies. Further analysis demonstrates that the knowledge extraction phases increase TMA, suggesting that improved alignment makes non-textual information more accessible during response generation. In summary, our findings uncover how Omni-LLMs can better express modality knowledge through language. Our code is available at https://github.com/naver-ai/OmniLLM.
Large Language Model Assisted Transportation Planning through a Differentiable User Equilibrium Layer
Transportation planning has traditionally required substantial expertise in network modeling, traffic assignment, and optimization. As a result, strategic analyses such as network design, capacity expansion, and model calibration remain inaccessible to non-expert decision makers. Recent advances in large language models (LLMs) provide an opportunity to bridge this gap by enabling natural-language interaction with complex transportation planning tools. However, existing LLM-based systems primarily serve as interfaces to simulation software and lack the ability to efficiently search, evaluate, and optimize transportation solutions. This study proposes an integrated human-AI transportation planning framework that combines a large language model, a model context protocol (MCP) interface, an open-source transportation planning engine (AequilibraE), and a differentiable user equilibrium (UE) assignment layer. The framework allows non-expert users to execute network analyses and improve solutions for optimization. A differentiable UE layer is developed by directly differentiating the Karush-Kuhn-Tucker (KKT) conditions of the Beckmann traffic assignment problem. The proposed framework is demonstrated through two representative planning tasks: calibration of travel time functions from observed traffic counts and network design under budget constraints. The examples illustrate how differentiable transportation models and LLM agents can jointly democratize transportation planning and substantially reduce barriers between domain expertise and transportation decision making.
幻影合并：当你的大型语言模型智能体选择了一个实体，却告诉你另一个实体的情况
大型语言模型智能体日益在多个实体上运行，而可靠的响应生成不仅需要检索相关证据，还需要保持证据与正确实体之间的关联。我们发现了幻影合并（PM）现象：这是一种实体—证据绑定失败，智能体会将未经支持的声明归因于其已确定的实体，即使任务实际上已经正确完成，也可能发生这种情况。我们在声明层面对 PM 进行了形式化定义，并刻画了其在四类智能体领域和四类骨干模型系列中的发生率及错误子类型。为解决这一问题，我们提出了一个用于声明级风险估计与修正的两阶段框架。第一阶段，Anchor-Grounded Risk（AGR）将智能体生成输出前隐藏状态的表征读出与基于雅可比矩阵的功能信息相结合，提供互补的槽位级风险分数和锚点值风险分数。第二阶段，Fixed-Anchor Correction（FAC）审查被标记的声明，并利用观测到的锚点证据对其进行选择性修订；当没有受支持的替代内容可用时，则删除这些声明。在留出的 Shopping 检测队列上，AGR-slot 达到 96.2% 的 AUROC 和 97.3% 的平均精确率，优于所评估的检测基线。在端到端 Shopping 缓解评估中，FAC 将轨迹级 PM 从 51.4% 降至 6.2%，输出与原始声明的比例为 80.7%。这些结果表明，实体特定的落地失败普遍存在，而将内部风险估计与受证据约束的声明修正相结合是有效的。我们的代码已公开发布于 https://github.com/LabRAI/PhantomMerge】【。
Pollard：用于框架无关 AI 代理控制与重放的受治理语义执行树
代理框架可以持久化检查点、跟踪工具调用并限制资源使用，但这些功能通常会为执行、治理和重放分别分配不同的标识。我们提出了 Pollard，一种基于受治理语义执行树的框架无关运行时。步骤标识绑定规范化输入、类型、尝试次数和父标识；同一棵树记录完成、拒绝、资源结算和分支历史。严格重放会返回经验证的已存储结果，而不会调用被封装的操作；事务性存储则协调并发工作器之间的资源预留。在签名的确定性冻结条件下，两个状态模型副本分别执行了评估转换，且未违反任何不变量；针对保留租约的 PostgreSQL 集群测试了精确限制，严格重放避免了外部调用，并检测出所有语义变更，同时在控制项上没有误报。一个多对框架矩阵保持了参考行为并验证了这些树。另一项具有成本上限的托管研究测试了实时提供商路径，每次已完成的 Pollard 试验都生成了一棵经过验证的树。总体而言，这些研究在明确的租约、边界和密封保管假设下，支持了其核心机制。
Unravel：从单张图像恢复可交互的 3D 场景
我们提出了 Unravel，这是一种无需训练的框架，可将单张图像转换为可交互的 3D 场景，其中每个物体都是独立且可移动的资产。现有生成器要么生成相互纠缠的单一几何体，导致其中没有任何物体可以移动；要么在一次前馈过程中预测所有物体，一旦猜测错误便无法检测或修正。Unravel 则采用了不同的方法：先将场景拆解，再将其重新组装。对于输入图像，Unravel 使用一种智能体框架逐层迭代剥离图像，并以点云作为恢复空间关系的主要模态。具体而言，视觉语言模型规划哪些物体是完全可见的，由图像编辑器移除这些物体并恢复被遮挡的区域。每次编辑在被纳入重建之前都会经过验证。该过程逐步将场景中被遮挡的部分转化为直接观测信息，使物体能够在其暴露最充分时被重建，而不是仅根据原始图像进行幻觉式推断。随后，场景按照移除顺序的逆序进行重建，从而使支撑物体能够自然地先于其所支撑的物体被放置。由于单视角图像只能展现物体的一部分，我们将观测点视为物体尺寸的下界：资产可以扩展到覆盖这些点，但不能缩小到偏离这些点。最后，我们对重建后的场景进行物理沉降，以消除悬浮物体和相互穿插现象，随后通过智能体审查来纠正残留的放置错误。在多种室内场景上进行的大量实验表明，与现有方法相比，Unravel 能够恢复更多物体，更忠实地进行放置，并生成物理合理性更高的场景。
WildTrace：在真实大规模 LLM 后训练中评估训练数据检测
尽管后训练能够显著提升模型性能，但对海量未公开数据的使用引发了人们对基准污染和版权侵权的担忧。这推动了人们对后训练模型训练数据检测的日益关注。现有的训练数据检测评估通常采用事后构建的基准，这可能引入分布偏移。另一种做法是依赖小规模训练设置进行受控评估，但这类设置与真实的后训练环境存在差异。因此，其报告的性能可能无法准确反映检测方法在现实场景中的有效性。为此，我们提出 \textsc{\method}，这是一个可靠的基准，旨在真实的大规模后训练环境中评估检测方法。基于多阶段流程，\method\space 涵盖了整个后训练生命周期中的各阶段模型，包括 SFT、DPO 和 RL。为了进行可靠的评估，我们针对四项多样化任务构建了特定于模型的评估集，其中包含分布匹配的成员样本和非成员样本，并为其提供可验证的成员资格标签。结果显示，即使是表现最佳的方法，在我们的基准上取得的平均 AUC 也仅为 0.529，接近随机猜测水平（0.5）。此外，我们的分析表明，评估协议会显著影响测得的检测性能。这些发现凸显了在真实场景下开发有效、实用且稳健的检测方法的巨大潜力。
通过带任务完整性约束的人设演化揭示 LLM 智能体的鲁棒性弱点
交互式智能体评估通常依赖标准化的、基于 LLM 的用户模拟器，这些模拟器表现出广泛的合作行为，可能忽视多样但现实的用户交互所引发的失败。这会导致对智能体鲁棒性的估计过于乐观，也难以为训练能够跨用户泛化的智能体提供足够的失败信号。现有方法大多依赖预定义的用户行为，可能无法发现特定于智能体或领域的弱点；而面向失败的搜索则可能通过使底层任务失效来降低任务成功率，而不是暴露真正的鲁棒性缺陷。为弥补这一空缺，我们提出 TIPER，一种保持任务完整性的人设演化方法，能够在保持底层任务有效且可解决的同时，自动发现暴露智能体特定鲁棒性弱点的用户策略。TIPER 在任务完整性约束下，通过反思引导的行为变异和按任务进行的 Pareto 选择，演化出持久的用户人设。在四个交互式智能体基准测试中，TIPER 在全部 12 个 GPT-5.4 设置下都取得了更低的留出集 avg@3 分数，观测到的降幅为 5.8 至 33.7 个百分点；在涵盖七个智能体的全部 28 个智能体—领域组合中也同样如此，观测到的降幅中位数为 22.9 个百分点。进一步分析得出了一套从行为到失败的分类体系，将发现的用户行为与反复出现的智能体失败模式联系起来。总体而言，这些结果表明，TIPER 能够对交互式智能体开展更具诊断性的鲁棒性评估，并为未来改进智能体鲁棒性提供有价值的失败分析。代码仓库：https://github.com/UCSB-NLP-Chang/TIPER。
Fisher information and the geometry of memorization in neural networks
A fruitful approach towards understanding generalization in machine learning involves characterizing how models separate generalizable structure from idiosyncratic or even corrupted features of training data. Fisher information has emerged as a diagnostic tool for this separation, leveraging loss geometry to distinguish generalization from memorization. Here we investigate this connection in vision models trained on CIFAR-10 with controlled noise that can be learned only through memorization. We show that projecting model weights onto the top Fisher eigenvectors decouples general task performance from noisy sample memorization, and we investigate how this decoupling presents differentially across layers. We show that the same phenomena are present in small transformers trained to solve a synthetic in-context learning task, and are reproduced in closed form in a simple, analytically solvable linear model. Our results reveal that Fisher information captures aspects of the geometric structure underlying how neural networks allocate capacity between generalization and memorization.
移动固体交互下的交互式四维体积液体预测
我们提出了交互式四维体积液体预测任务：根据固体物体在液体中的运动方式，预测未来稠密三维液体速度场的时空演化。该任务为需要推理运动条件下液体响应的系统提供了上游物理预测基础，包括机器人液体操控、数字孪生和刚体-流体视频生成。现有研究已尝试预测静态或预设固体几何体周围的流动，并合成视觉上合理的流体运动。然而，这些方法对交互式液体动力学的覆盖有限，因为它们无法预测运动固体如何引发未来全域液体响应。在这一预测场景中，流体动力学效应集中在不断演化的液-固界面附近，但预测目标却是全局时空演化过程。为应对这一挑战，我们提出了一个两级循环预测框架 TIDE/TIDES：TIDE 专注于局部界面驱动的预测，而 TIDES 增加了感知输运的耦合和显式稳定化，以减少累积的演化误差。在经过适配的邻近基线、受控组件消融、运动/尺寸分布外测试以及长时域评估中，TIDES 均取得了最佳预测性能；消融实验表明，局部交互、输运耦合和稳定化是关键贡献因素。总体而言，我们提出的任务定义、基准以及 TIDE/TIDES 研究，为运动固体条件下的稠密液体预测建立了一个受控基础。代码已公开于 https://github.com/jongwonryu/TIDE-TIDES。
EventLens：面向视频理解的事件结构强化学习
近期的多模态大语言模型（MLLM）在视频问答方面取得了出色的效果，但往往无法恢复使视频保持连贯的时间结构。它们可能会合并相邻事件、遗漏单个事件的进展过程，或推断出跨事件的不合理关系。我们将结构化时间理解定义为从视觉证据中恢复事件结构的能力，包括事件边界、单事件进展和多事件关系。我们认为，这一缺口源于当前后训练目标与时间推理之间的不匹配：基于描述的监督使模型能够依赖语言捷径，而不是对视觉时间信息进行定位。为此，我们提出 EventLens，一种以视觉为中心、通过事件结构恢复来学习视频理解的强化学习框架。EventLens 不重建描述文本，而是在受控扰动下训练模型恢复时间结构，例如改变时间粒度、反转进展过程以及打乱事件顺序。我们基于三级时间本体构建了这一框架，并推导出三类可验证任务：事件分割、进展辨别和多事件关系推理。这些任务支持确定性奖励，因此无需人工偏好标签或学习式奖励模型，即可进行任务专门化的 GRPO 训练。我们进一步引入多教师在线策略蒸馏（MT-OPD），将专门化策略整合到统一模型中。在 Qwen3-VL 主干模型上的实验表明，EventLens 在时间定位和事件敏感推理基准上均取得了稳定提升，同时保持了通用视频问答性能。学习曲线进一步表明，与直接进行混合强化学习优化相比，EventLens 在显著降低训练成本的同时，实现了相当的下游迁移效果。
ProCARE：基于数据概况的证据契约实现真实世界研究自动化
真实世界研究（RWS）分析常规收集的医疗保健数据（如电子健康记录和保险理赔数据），以估计治疗效果、监测安全性，并识别随机试验之外的风险因素。利用大型语言模型智能体自动化 RWS 具有吸引力，但也较为脆弱：有效分析依赖于特定数据集的临床语义，包括索引日期、随访窗口、患者级或事件级分析单位，以及观察性数据中通常未明确规定的结局边界。这会导致模式不匹配、时间错误、分析粒度不一致以及缺乏依据的临床结论。我们提出 ProCARE，一个以数据概况为基础的框架，将 RWS 自动化重新表述为受约束的编译过程。ProCARE 将异构数据概况转换为证据契约，其中明确规定可用变量、时间属性、分布特征和证据边界，并将研究目标编译为 RWS-IR——一种针对变量、时间关系、分析单位和统计任务的强类型中间表示。随后，通过确定性验证和局部修复，对生成的工作流进行检查和修订，覆盖模式映射、队列构建、变量推导、分析和结果解释。在 36 项基于 EHR 的基准任务中，ProCARE 在代码可执行性、语义一致性和结论可追溯性方面均优于强大的大型语言模型智能体基线；自动化违规诊断结果与盲法专家评审显示出中等至较高的一致性。
Label-Centric Knowledge Retrieval for Zero-Shot Text Classification
Flagship large language models (LLMs) offer rich general knowledge and strong reasoning capabilities, but their high inference cost and latency limit their use for large-scale data labeling. We study zero-shot text classification with semantically meaningful labels and task descriptions, without task-specific labeled examples or classifier feedback. We propose \textbf{SLKR} (Semantic Label Knowledge Retrieval), a label-centric retrieval-augmented generation framework that separates offline knowledge construction from lightweight online inference. Offline, a flagship LLM independently generates label definitions, retrieval keywords, and boundary rules for potentially confusable label pairs, forming a reusable label-knowledge store. Because this construction operates over the label space rather than the input corpus, its one-time flagship cost can be amortized across large-scale inference. Online, hybrid semantic and lexical retrieval first identifies the top-$K$ candidate labels and then selects up to $M$ input-relevant boundary rules connecting candidate-label pairs. A lightweight LLM performs the final classification in a single call using the retrieved definitions and rules. Across three public benchmarks and a real-world multilingual book-classification dataset, SLKR achieves the best results on all four datasets with Qwen3.5-4B and on three of four with Ministral-3-3B, while using compact input contexts and no task-specific labeled examples. Boundary ablations further show that retrieving a small set of input-relevant rules preserves performance close to using all boundaries with substantially less context, supporting efficient reuse of flagship-generated label knowledge for lightweight classification.
Correct Choices, Failed Sequences: Choice2Seq Benchmarks Correctness Retention Across Output Interfaces
A model may select the correct answer yet fail to construct it through a sequence of local decisions. We introduce Choice2Seq, a diagnostic benchmark that measures correctness retention across complete-answer selection and sequential construction in independent runs on the same problems and candidate sets. Sequence Composition Rate (SCR) measures the fraction of correctly selected items that are also answered correctly through local construction. Jev achieves 97.1% and 99.5% selection accuracy on arithmetic and lexical retrieval, yet retains only 33.0% and 2.1% of those correct choices under serial construction, respectively. Substantial gaps also emerge for StartLux and Qwen, including when Qwen emits characters directly. Restricting local options to candidate-supported characters improves construction, while explicit position indexing improves copying fidelity, even for incorrect selections. These interventions distinguish task correctness, plan fidelity, and output compliance. Matched dialogue-state evaluations reveal that local field construction largely preserves Jev's correct states but reduces Qwen's mean-turn accuracy, highlighting task- and interface-dependent differences. Sequence-level confidence and request-cost analyses further characterize completed outputs.
MP-MNCA: Morphology-Prior Masked Neighbor Cross-Attention for Efficient Spatial Domain Identification
Spatial transcriptomics jointly captures molecular state, tissue morphology, and spatial organization, yet spatial-domain methods are typically compared only at late training checkpoints, obscuring differences in convergence efficiency. We introduce MP-MNCA, a morphology-prior masked neighbor cross-attention framework that preserves the full 3,000 dimensional highly-variable-gene representation while selectively routing molecular information across spatial neighbors, with histological similarity incorporated as a continuous attention prior under a self-supervised masked neighbor-contrastive objective. Across 12 heterogeneous human dorsolateral prefrontal cortex (DLPFC) sections, MP-MNCA is evaluated against established spatial-domain baselines using both convergence trajectories and epoch-wise spatial-domain visualizations. Our results show that MP-MNCA (i) reaches coherent spatial-domain structure in substantially fewer training epochs, as reflected quantitatively by convergence curves and qualitatively by progressive spatial grids, and (ii) maintains strong clustering quality, achieving a mean peak refined ARI of 0.546 versus 0.549 for the strongest baseline (STAIG), while attaining the highest mean refined NMI of 0.657 versus 0.640. Moreover, MP-MNCA outperforms STAIG on 7 of 12 sections in ARI and 11 of 12 in NMI, demonstrating faster convergence without sacrificing spatial-domain quality.
SafeDSE: Joint Exploration of Safety Architecture Patterns, Mapping, and Scheduling for Dataflow Applications With Safety Arguments (Extended Version)
Embedded platforms increasingly host safety-related streaming functions next to less critical ones. Whether a safety architecture pattern, such as a doer/checker pair, protects a function depends on the cores that run its components. Their schedule must meet the throughput requirement, and mixing safety integrity levels (SILs) on one core raises development cost. This report presents SafeDSE, a design space exploration framework for synchronous dataflow applications. One declarative constraint model decides the pattern of each actor, the placement that the fault model requires, the static-order schedule, and the platform, and prices SIL promotion. Guards activate the chosen patterns in a fixed graph that holds all of them, and node potentials encode throughput. An independent verifier re-checks every solution, and a generator turns its checks into evidence of a safety argument in Goal Structuring Notation. On 9 benchmark instances under period bounds, no rule that fixes patterns before mapping was both always feasible and never more expensive under conservative pricing, while joint exploration was. In the generated arguments, 49 to 62% of the leaves cite verifier checks; the rest are claims a mapping cannot discharge. Optimality within the model was proven for all synthetic graphs with up to 16 actors. This extended version adds the full model, the verifier and generator in detail, the complete catalog and parameters, and a proof of the potential characterization. It also reports every experiment per instance, a mutation analysis of the verifier, and the negative results of the development.
OneShot: Label-Efficient Detector Weighting for Time Series Anomaly Detection
Time series (TS) anomaly detection (AD) lacks a universal detector, with the best choice determined by the series rather than by the algorithm. Existing methods therefore either select one detector for each series or average a pool of them. However, both assume that no label ever arrives, while a deployed monitor does receive them, since an operator confirms the alarms it raises. In this paper, we propose ONESHOT, which uses a single confirmed anomaly to weight the pool rather than to choose one detector from it. Specifically, we score each detector on the confirmed anomaly, and those scores become the weights of a soft mixture over the whole pool. We show that a single confirmed anomaly is enough to come close to the best detector for the series. Across 152 univariate and 73 multivariate series, ONESHOT outperforms the strongest single detector, averaging the whole pool, and even picking the best detector on the confirmed event. Code is available at: https://github.com/seunghan96/one_shot.
Look Back to Move Forward: Learning to Revisit for Agentic Compositional Image Generation
Text-to-image models have made remarkable progress in visual fidelity and instruction following, yet complex compositional generation remains challenging. Existing agentic approaches address this limitation through iterative generation and editing, adapting their actions based on interaction history. However, they typically fix the continuation state (i.e., the image from which refinement proceeds) to the latest image. Since iterative editing is not necessarily monotonic, the agent may continue refining from the latest image even when it has already degraded. In this work, we introduce GenRevisit, an agentic framework that extends action selection to joint continuation-state and action selection. The agent can not only use interaction history to adjust editing instructions but also select an image candidate to edit, jointly deciding where to continue from and how to refine. We train GenRevisit with SFT followed by agentic reinforcement learning. To address the coarse credit assignment of final outcome reward in GRPO, which assigns the same trajectory-level advantage to all actions, we further propose Normalized Action Credit Assignment (NACA) to redistribute positive advantage according to each action's contribution to reducing the remaining objective. On GenEval2, GenRevisit achieves GM scores of 84.3 and 74.4 with Qwen-Image-2512 and Z-Image-Turbo, surpassing the strongest baselines using the same generators by 12.7 and 10.9 points respectively. GenRevisit achieves state-of-the-art performance on GenEval, GenEval++, and GenEval2 among baselines.
Learning to GUIDE the Group: Self-Learning LLM Facilitation Across Multi-Party Tasks
As large language models are increasingly adopted as multi-party communication facilitators, there is a growing demand for agents capable of balancing diverse group dynamics across contexts. However, existing LLM facilitators are often fine-tuned for task-specific objectives; generalized LLM-based facilitation methods that can adapt to different types of tasks and diverse group dynamics are underdeveloped. To address this gap, we present GUIDE (Generalized Unified Intervention for Dynamic Exchanges, a framework for developing generalized LLM-based facilitators across different multi-party contexts. GUIDE evaluates facilitation across both objective task outcomes and subjective metrics, including psychological safety and perceived friction. Within this framework, we study whether LLMs can self-learn generalized facilitation strategy offline to real-time interventions. We first collect multi-party human interactions guided by task-specific expert LLM facilitators (N=432 conversations). We then employ a self-learning method in which an LLM induces an actionable strategic playbook from conversations and iteratively refines it using feedback from an LLM judge aligned with human annotators (N=20), developing a single generalized facilitator that outperforms the three task-specific expert facilitators on held-out conversations 61.4% win rate. Finally, we evaluate these improvements empirically; in a real-time online empirical study, the self-learned generalized facilitator preserves facilitator performance and group productivity, improves speaking comfort and reduces interpersonal friction, while using 24.7% fewer turns and 29.0% fewer words than the expert prompts. These findings provide empirical evidence that LLMs can learn generalizable, task-agnostic social facilitation from task-specific human interactions, on-par with hand-crafted task-specific prompts across diverse group settings.
AutoNovo: Agentic Program Search Discovers Transferable Nucleic Acid Design Algorithms
In-silico optimization of nucleic acids relies on predictive oracles that imperfectly capture biological function, so sequences optimized against one model may exploit model-specific artifacts and fail to generalize to independent models. Recent studies have shown that not all designers generalize equally well to independent models. In this work, we use agentic program search to iteratively propose, execute, evaluate, and revise code for nucleic-acid designers. We describe the successful strategies used to prompt the search. AutoNovo designers achieved the best observed mean among the evaluated methods on 3/3 DNA-design tasks in terms of in-silico performance, and 6/6 transfer tasks in terms of independent models. The designers come from a family of discovered nucleic acid designers rather than a single universal best designer. We plan to release the resulting designers. These results are the first steps in creating nucleic acid designers that maximize cross-model generalization, and thus improving the odds that in-silico designed sequences reflect real biology.