论文

将一个确定性、无模型的流水线编译进 Agent 记忆：该流水线将本地采集流切分为类型化的活动帧与有界事件片段，携带应用、站点、时间、输入量以及回指原始行的证据指针，全程无模型参与。A deterministic, zero-model pipeline is compiled into agent memory with a deterministic, zero-model pipeline that segments a local capture stream into typed activity frames, bounded episodes carrying application, site, timing, input volume, and evidence pointers back to the raw rows, with no model in the loop.

该综述围绕"权重"与"技能"这一轴线组织领域，梳理了互补的"技能"一极——从无监督强化学习的技能发现，到大语言模型的技能库——并指出"skill"一词至少存在五种不同含义。This survey organises the field around that axis of weights versus skills, and maps the complementary"skills"pole, from unsupervised reinforcement-learning skill discovery to large-language-model skill libraries, and shows that the word "skill" is used in at least five distinct senses.

提出 FactorJEPA，将世界结构作为一等预测原语，并通过 visibility gate 与分离的子空间来组合布局、实体与交互，以保留部分可观测的 Agent 并抑制跨因子捷径。FactorJEPA is introduced, which makes world structure a first-class predictive primitive, and composes layout, entities, and interactions, using a visibility gate and separated subspaces to preserve partially observed agents and discourage cross-factor shortcuts.

GaussianSelector 在选择质量上与 SOTA 的多视角 SAM 方法相当，同时所需交互视角显著更少、计算开销明显更低，因而非常适合真实场景下 human-in-the-loop 的 3D 场景编辑与 3D 资产提取。GaussianSelector achieves competitive selection quality against state-of-the-art multi-view SAM-based methods, while requiring significantly fewer interaction views and substantially lower computational overhead, which makes it well suited for human-in-the-loop 3D scene editing and 3D asset extraction in real-world deployment scenarios.

本文详细阐述了驱动 PyTorch 实现的原则及其在架构中的体现，并解释了 runtime 关键组件的精心且务实的实现如何使其协同工作以获得出色的性能。This paper details the principles that drove the implementation of PyTorch and how they are reflected in its architecture, and explains how the careful and pragmatic implementation of the key components of its runtime enables them to work together to achieve compelling performance.

这项系统性研究在数十项语言理解任务上比较了预训练目标、架构、无标注数据集、迁移方法及其他因素，并在涵盖摘要、问答、文本分类等的许多基准上取得了 SOTA 结果。This systematic study compares pre-training objectives, architectures, unlabeled datasets, transfer approaches, and other factors on dozens of language understanding tasks and achieves state-of-the-art results on many benchmarks covering summarization, question answering, text classification, and more.

本文提出了一种基于迭代模型平均的深度网络联邦学习实践方法，并进行了广泛的实证评估，考虑了五种不同的模型架构和四个数据集。This work presents a practical method for the federated learning of deep networks based on iterative model averaging, and conducts an extensive empirical evaluation, considering five different model architectures and four datasets.

本文提出了 mixup，一种通过对样本对及其标签的凸组合来训练神经网络的简单学习原则，提升了 SOTA 神经网络架构的泛化能力。This work proposes mixup, a simple learning principle that trains a neural network on convex combinations of pairs of examples and their labels, which improves the generalization of state-of-the-art neural network architectures.

本文提出了一种通用的无监督学习方法——对比预测编码（Contrastive Predictive Coding），用于从高维数据中提取有用的表征，并在语音、图像、文本和 3D 环境中的强化学习四个不同领域取得了出色的性能。This work proposes a universal unsupervised learning approach to extract useful representations from high-dimensional data, which it calls Contrastive Predictive Coding, and demonstrates that the approach is able to learn useful representations achieving strong performance on four distinct domains: speech, images, text and reinforcement learning in 3D environments.

本文提出了一种双时间尺度更新规则（TTUR），用于在任意 GAN 损失函数下使用 SGD 训练 GAN，并引入了 Frechet Inception Distance（FID），相比 Inception Score 能更好地捕捉生成图像与真实图像之间的相似性。This work proposes a two time-scale update rule (TTUR) for training GANs with stochastic gradient descent on arbitrary GAN loss functions and introduces the "Frechet Inception Distance" (FID) which captures the similarity of generated images to real ones better than the Inception Score.

名为 PaLM 的 540 亿参数、密集激活的 Transformer 语言模型取得了突破性性能，在一系列多步推理任务上超越了微调后的 SOTA，并在最近发布的 BIG-bench 基准上超越了人类平均水平。A 540-billion parameter, densely activated, Transformer language model, which is called PaLM achieves breakthrough performance, outperforming the finetuned state-of-the-art on a suite of multi-step reasoning tasks, and outperforming average human performance on the recently released BIG-bench benchmark.

本文提出了一种通过创建在所有边缘设备之间全局共享的小型数据子集来改进非独立同分布数据训练的策略，并表明在 CIFAR-10 数据集上，仅共享 5% 的全局数据即可将准确率提升 30%。This work presents a strategy to improve training on non-IID data by creating a small subset of data which is globally shared between all the edge devices, and shows that accuracy can be increased by 30% for the CIFAR-10 dataset with only 5% globally shared data.

采用迭代的在线训练模式，按周节奏用新的人类反馈数据更新偏好模型与 RL 策略，并发现 RL 奖励与策略相对其初始化的 KL 散度平方根之间近似呈线性关系。An iterated online mode of training, where preference models and RL policies are updated on a weekly cadence with fresh human feedback data, and a roughly linear relation between the RL reward and the square root of the KL divergence between the policy and its initialization is identified.

研究表明，APE 生成的提示词既可引导模型趋向真实性和/或信息量，也可通过将其前置拼接到标准上下文学习提示词之前来提升少样本学习性能。It is shown that APE-engineered prompts can be applied to steer models toward truthfulness and/or informativeness, as well as to improve few-shot learning performance by simply prepending them to standard in-context learning prompts.

本文提出 Atlas，一个经过精心设计并预训练的检索增强大语言模型，能以极少训练样例学习知识密集型任务，并研究了文档索引内容的影响，表明该索引可便捷地更新。This work presents Atlas, a carefully designed and pre-trained retrieval augmented language model able to learn knowledge intensive tasks with very few training examples, and studies the impact of the content of the document index, showing that it can easily be updated.

提出 PHOENIX（Predictive Health On-orbit Edge Neural Intelligence eXtension），为卫星赋予自主故障推理能力，并在 ESA Anomaly Detection Benchmark 上报告了初步结果。PHOENIX (Predictive Health On-orbit Edge Neural Intelligence eXtension) is proposed to give the satellite its own fault reasoning capability, and preliminary results on the ESA Anomaly Detection Benchmark are reported.

提出 SimWAM，一种简单而有效的 WAM，仅将未来视频预测用作训练时的监督信号，并通过联合 flow matching 协同训练一个预训练视频专家与一个轻量动作专家。SimWAM is presented, a simple yet effective WAM that leverages future-video prediction solely as a training-time supervision signal, and co-trains a pretrained video expert and a lightweight action expert with joint flow matching.

提出 YOLO-PEFT，一个结构感知的框架，将 adapter 的放置建模为可审计的约束规划问题，以显式、可审查的规划取代手工对目标模块的试错，同时保留已验证的 train-save-merge-export 路径。YOLO-PEFT is proposed, a structure-aware framework that formulates adapter placement as an auditable constraint-planning problem that replaces manual target-module trial and error with explicit, inspectable planning while preserving verified train-save-merge-export paths.

本文研究 11 项重大生活事件引发的人格变化，以大五人格作为心理测量锚点，并将所得轨迹与人类人格心理学的纵向证据进行对照，指出当前 PC-Agents 模拟了人类人格动态的均值，但未能模拟其形态。This work studies event-induced personality change after 11 major life events, using the Big Five traits as a psychometric anchor and interpreting the resulting trajectories against longitudinal evidence from human personality psychology, and suggests that current PC-Agents simulate the mean of human personality dynamics, but not its shape.

实验表明，对敏感信息的不必要获取广泛存在，并观察到任务完成能力与获取阶段泄露之间的相关性，因此对获取阶段隐私进行审计既紧迫也必要。The experiments show that the unnecessary acquisition of sensitive information is widespread, and a correlation between the task-completion capability and acquisition-stage leakage, and auditing acquisition-stage privacy both urgent and necessary is observed.

提出 C4，一个受认知启发的成语跨概念创造力评估框架，揭示了当前 MLLM 在通过跨概念关系解码创造性编码语义方面存在的显著差距。C4, a cognition-inspired evaluation framework for Chengyu (Chinese idiom)-based Cross-Concept Creativity, is introduced, exposing a substantial gap in how current MLLMs decode creatively encoded meaning through cross-concept relations.

SMRC-SD（State-Matched Routing and Contextualized Self-Distillation）显式地决定特权轨迹应在何时、以何种方式指导 on-policy student，其表现始终优于无条件的成功全路径蒸馏。State-Matched Routing and Contextualized Self-Distillation (SMRC-SD), which explicitly determines when and how a privileged trajectory should guide an on-policy student, consistently outperforms unconditional successful full-path distillation.

往返一致性将可逆性转化为生成式模型一种实用的可信信号；双向训练带来负成本，在两个方向上均优于单向专家模型；其中反向还可作为快速的逆问题求解器。Round-trip consistency turns reversibility into a practical trust signal for generative models, and Bidirectional training comes at negative cost, beating direction specialists in both directions, and the backward direction doubles as a fast inverse solver.

本文在低 prefill 延迟约束下优化 Pareto 前沿以最大化精度，提出 CoinRAG（Contextualized Information Nugget KV Cache Reuse for Long-Context RAG），通过 chunk 级上下文无缝拼接其切片化的 KV 表示。This paper optimizes the Pareto frontier under low prefill latency constraints while maximizing accuracy by proposing CoinRAG (Contextualized Information Nugget KV Cache Reuse for Long-Context RAG), which seamlessly assembles their sliced KV representations with a chunk-level context.

本文提出一种结构化协议，通过对开源 LLM 评估与安全工具的分类驱动分析来自动化 AI 风险缓解，并给出一个可同时适用于开源与商用方案的分类驱动框架。This paper proposes a structured protocol to automate AI risk mitigation through a taxonomy-driven analysis of open-source LLM evaluation and security tools, and presents a taxonomy-driven framework applicable to open-source and proprietary solutions.

通过分别评估论文检索、证据 grounding 与答案准确性，LitTraceQA 为生成可验证答案、而非无依据摘要的科学 QA 系统提供了测试基准。By evaluating paper retrieval, evidence grounding, and answer accuracy separately, LitTraceQA provides a testbed for scientific QA systems that produce verifiable answers rather than unsupported summaries.

本文提出 Exact Adaptive Hybrid Retrieval（EAHR），将完整列表加权 RRF 定义的有序 Top-K 固定为检索目标，并将通道深度视为请求特定的执行状态，在全部 150 组 query-snapshot 组合中复现了完整列表的有序 Top-20。This work proposes Exact Adaptive Hybrid Retrieval (EAHR), which fixes the ordered Top-$K$ defined by complete-list weighted RRF as the retrieval target and treats channel depth as request-specific execution state and reproduced the complete-list ordered Top-20 in all 150 query-snapshot combinations.

HiSparse 已合入上游 SGLang，并在 H200、B200 与 GH200 平台上针对 DSA、NSA、Quest 三类稀疏注意力家族进行评估：在长上下文负载下峰值生成吞吐提升最高达 4.7×，同时保持可比的 per-token 延迟，并降低高负载下的 time-to-first-token。HiSparse is merged into upstream SGLang and evaluated across three sparse-attention families (DSA, NSA, and Quest) on H200, B200, and GH200 platforms: it improves peak generation throughput by up to 4.7x on long-context workloads while preserving comparable per-token latency and reducing time-to-first-token at high load.

本文是一项关于如何提升蒸馏训练效率的实践研究，围绕两项系统贡献展开，并提出一种融合的 chunked KL loss，使峰值内存随序列长度线性增长。A practitioner's study of how to make distillation training efficient is presented, organised around two systems contributions, and a fused, chunked KL loss is introduced, making peak memory linear in the sequence length.

研究发现，微调后的 AO 可能变成概念特异的 anti-reader：它们会选择性地无法恢复自身训练过程中持续存在的概念，从而对习得的可解释性接口提出可靠性担忧。It is found that fine-tuned AOs can become concept-specific anti-readers: they selectively fail to recover the concept persistently present during their own training, raising a reliability concern for learned interpretability interfaces.

本文通过深度多模态嵌入视觉与自然语言数据，提出了一种用于图文双向检索的模型，并引入结构化的最大间隔目标，使该模型能够显式地跨模态关联片段。This work introduces a model for bidirectional retrieval of images and sentences through a deep, multi-modal embedding of visual and natural language data and introduces a structured max-margin objective that allows this model to explicitly associate fragments across modalities.

论文证明，使用标注数据进行微调，并允许模型查询外部知识源，能够在安全性和事实性这两个关键挑战上带来显著提升。It is demonstrated that fine-tuning with annotated data and enabling the model to consult external knowledge sources can lead to significant improvements towards the two key challenges of safety and factual grounding.

本文利用源自 CNN 的同一基础信息重新审视初始搜索与重排序两个检索阶段，显著改进了现有基于 CNN 的识别流水线。This work revisits both retrieval stages, namely initial search and re-ranking, by employing the same primitive information derived from the CNN, and significantly improves existing CNN-based recognition pipeline.

通过引入扩散卷积运算，本文展示了如何从图结构数据中学习基于扩散的表示，并将其作为节点分类的有效基础。Through the introduction of a diffusion-convolution operation, it is shown how diffusion-based representations can be learned from graph-structured data and used as an effective basis for node classification.

Manifold Mixup 在监督学习、对单步对抗攻击的鲁棒性、半监督学习以及留出样本的负对数似然（NLL）上，相较强基线均取得了大幅提升。Manifold Mixup achieves large improvements over strong baselines in supervised learning, robustness to single-step adversarial attacks, semi-supervised learning, and Negative Log-Likelihood on held out samples.

本文介绍了 CodeXGLUE，一个基准数据集，旨在推动面向程序理解与生成的机器学习研究，涵盖 14 个数据集上的 10 项任务，并提供模型评估与比较的平台。This paper introduces CodeXGLUE, a benchmark dataset to foster machine learning research for program understanding and generation that includes a collection of 10 tasks across 14 datasets and a platform for model evaluation and comparison.