👍 372
10/26 08:00
Long-horizon LLM agents interact with their environments through complete assistant turns while generating each action token by token. We observe that critic values vary primarily across turns, motivating optimization anchored at turns with finer credit allocation among tokens. We propose Multiscale
中文介绍 该论文针对长视程LLM智能体提出多尺度策略优化(Multiscale Policy Optimization)。由于批评值主要在回合间变化,该方法以回合为锚点,在回合内进行更细粒度的信用分配,优化令牌生成。此策略有望提升长程决策性能,但摘要未提供量化结果。
👍 311
10/26 08:00
Multimodal intelligence can be viewed as human-guided co-evolution between worlds and agents through interaction. However, scaling this process faces two bottlenecks: constructing heterogeneous, verifiable environments is costly, and manual Harness iteration cannot keep pace with growing task divers
中文介绍 该论文提出「Trinity」框架,通过能力驱动的递归自改进实现多模态智能体与环境的协同进化。为解决环境构建成本高和人工Harness迭代慢两大瓶颈,Trinity采用能力驱动的自改进机制,使智能体能够自主构建可验证环境并递归提升自身能力,以适应任务多样性增长。
👍 171
10/27 08:00
Diffusion video models produce striking clips, but they cannot reliably render an exact formula, a correct chart, a rule-consistent game or narration synchronized to on-screen events. A different route is emerging: multimodal agents that write code to render every frame and synthesize every sound. W
中文介绍 论文提出新基准VibeVideoBench,用于评估通过编写代码来生成视频的多模态智能体。相较于扩散模型,该方法能精确渲染公式、图表及规则一致的游戏,并同步音视频。基准涵盖多种代码驱动视频生成任务,以评估智能体理解和执行复杂指令的能力。
👍 111
10/25 08:00
Memory is increasingly used in long-horizon agents, yet the net value of building memory over direct retrieval from raw history remains largely unclear. Through a large-scale evaluation across multiple benchmarks and model backbones, we find many dedicated memory systems underperform naive RAG in ma
中文介绍 论文通过大规模多基准评估,剖析了长视程智能体中构建记忆网络的净价值。研究发现,许多专用记忆系统在多数任务上表现不如直接对原始历史进行RAG检索,揭示了「记忆税」问题。该结果挑战了当前广泛使用的复杂记忆架构,并为设计高效记忆机制提供了实证依据。
👍 109
10/23 08:00
Recent test-time training (TTT) frameworks have enabled open-weight large language models (LLMs) to outperform inference-only frameworks powered by frontier models on scientific discovery tasks. However, TTT introduces additional computational cost for online weight updates during the discovery proc
中文介绍 论文比较了测试时记忆演化与测试时训练(TTT)在科学发现任务中的性能。TTT虽能提升开源LLM性能,但需在线更新权重,计算成本高。研究者提出测试时记忆演化方法,在不更新权重的情况下动态调整记忆,旨在以更低成本达到或超越TTT效果,但需验证具体性能差异。
👍 88
10/27 08:00
Discrete visual tokens represent vision in a discrete form analogous to language, offering a native interface for multimodal modeling under a shared next-token prediction objective. However, quantization error is widely assumed to cause unavoidable semantic loss, fundamentally limiting the semantic
中文介绍 论文探讨离散视觉令牌是否可作为视觉信号的无损语义表示。传统观点认为量化误差会导致不可避免的语义损失。该工作挑战此假设,提出新的离散化方法或训练策略,可能通过改善量化误差或利用其结构信息,使离散令牌在语义上接近连续表示,从而提升多模态模型的视觉理解能力。
👍 78
10/20 08:00
In long-term interactions with large language model (LLM) agents, users reveal habits through their choices and actions. These habits capture implicit user information that agents must infer across conversations. However, individual observations may provide only partial evidence about a habit or whe
中文介绍 论文提出HABIT-Bench基准,评估LLM智能体从长时交互中推断用户习惯的能力。由于个体观察仅提供部分证据,习惯推断颇具挑战。该基准提供纵向弱证据环境,测试智能体能否整合跨对话信息,识别隐性用户模式,并可能评估其预测和适应能力。
👍 50
10/26 08:00
Action-conditioned world models offer a scalable alternative to costly real-world interaction for improving robot policies. However, existing world-model-based methods largely treat world models as forward-only surrogates of the physical environment, leaving untapped a capability no physical environ
中文介绍 论文提出Rollback方法,利用世界模型中的经验回溯来改进机器人策略。传统世界模型仅作前向代理,忽略回溯能力。Rollback通过回溯过往经验,使智能体能够从错误中学习并优化决策。在机器人控制任务中,该方法有望减少真实交互成本,提升策略收敛速度。
👍 49
10/27 08:00
Video captioning aims to preserve visual information through compact descriptions, predominantly using natural language. However, precisely expressing fine-grained spatial structures and temporal dynamics in natural language often leads to lengthy and potentially ambiguous descriptions. In this work
中文介绍 论文提出CodeCap-Bench基准,探索代码是否可作为视频描述的有效替代方案。自然语言描述在表达细粒度空间结构和时间动态时往往冗长且模糊。该研究评估代码描述在视频理解中的表现,发现其可能在保持精确性的同时更紧凑,为视频理解提供新范式。
👍 47
10/27 08:00
Vision-Language-Action (VLA) models have made substantial progress in robotic manipulation, but their generalization remains constrained by the diversity and coverage of action-labeled robot demonstrations. Human videos provide much broader behavioral and environmental experience, while learning fro
中文介绍 论文提出ThinkWorld框架,通过世界想象强化链式动力学推理。为缓解VLA模型因动作标注数据不足导致的泛化瓶颈,该方法利用人类视频丰富经验,构建世界模型以模拟动力学,从而增强机器人操作策略的泛化能力,无需额外动作标签。
👍 24
10/27 08:00
Multimodal perception forms the foundation of multimodal intelligence. In particular, spatial perception and multimodal long-context understanding are two pivotal tasks that underpin embodied AI and multimodal agents. However, synthesizing targeted, high-quality training data for these tasks remains
中文介绍 该论文介绍Zooming without Zooming v2,用于提升多模态感知的泛化能力。该方法专注于空间感知和多模态长上下文理解,通过合成高质量训练数据来连接证据,使模型能从细节扩展到全局证据关联。此框架有望在无需大规模人工标注下增强多模态智能体的感知能力。
👍 22
10/26 08:00
Diffusion language models (DLMs) promise parallel text generation, but in practice face an unresolved speed-quality trade-off. Exact decoding fills in only one masked position per model pass, while faster multi-token decoding schemes are lossy. Speculative decoding is the standard lossless accelerat
中文介绍 论文提出SPOT方法,增强扩散语言模型(DLM)的投机解码。DLM存在速度与质量权衡困境,而SPOT通过预测提交顺序实现无损加速。该方法能并行解码多个位置,同时保持与精确解码相同的质量,有效提升DLM生成速度,解决其速度瓶颈。
👍 19
10/24 08:00
Can AI agents autonomously train useful language models and improve them using feedback from users? We introduce RSI Arena, a two-stage study of autonomous post-training as a step toward recursive self-improvement. Ten frontier models act as training agents, each starting from NVIDIA Nemotron 3.5 Li
中文介绍 论文介绍RSI Arena,研究基于人类反馈的递归自改进。十个前沿模型作为训练代理,从NVIDIA Nemotron 3.5开始,通过用户反馈自主训练和提升语言模型。该两阶段研究探索AI能否自主完成后训练并利用人类反馈实现递归改进,为AGI研究提供实证基础。
👍 18
10/27 08:00
Mixture-of-Experts (MoE) models are bottlenecked by loading routed-expert weights from memory during interactive decoding at small batch sizes. Because most intermediate channels in a routed expert contribute little to any single token, channel selectors predict which channels matter and read only t
中文介绍 论文提出无需校准数据的通道选择方法,用于加速MoE模型解码。由于路由专家权重加载是瓶颈,且多数中间通道对单令牌贡献小,该方法预测关键通道并仅加载它们,从而减少内存访问。此方法无需校准数据集,直接提升小批量解码效率,降低推理延迟。
👍 15
10/25 08:00
Generative models increasingly expose a reasoning effort setting that lets users scale inference-time compute, trading higher cost for potentially improved performance. Users must choose an effort level before knowing whether extra reasoning will help, and even a successful run does not reveal wheth
中文介绍 论文从经济学角度研究LLM智能体中推理努力级别的选择问题。用户需在不知额外推理是否有益的情况下预选努力级别,且成功运行不揭示其性价比。该研究建立决策框架,分析推理成本与性能的权衡,帮助用户选择最优努力级别,最大化效用。
👍 13
10/25 08:00
While progress in informal mathematics has come largely from scaling the chain-of-thought and agentic reasoning of frontier language models, formal theorem proving has increasingly relied on human-designed workflows around these models, such as planning, decomposition, and orchestrating specialized
中文介绍 论文借鉴「苦涩的教训」,提出h1方法,用于自动定理证明。与依赖人工设计的规划和分解流程不同,h1将规模扩展与简单搜索结合,利用学习到的启发式引导证明搜索。该方法有望减少人工干预,提升形式化定理证明的自动化水平,但需验证实际性能。
👍 11
10/26 08:00
Omnimodal large language models (Omni-LLMs) are increasingly expected to assist users while observing continuous audiovisual streams. However, existing evaluations provide limited insight into how reliably these models connect stream evidence and user needs to timely, sustained interventions. We int
中文介绍 论文提出Omni-InteractScope基准,评估全模态LLM在连续音视频流中的主动交互能力。该基准关注模型如何连接流证据与用户需求,并及时、持续地进行干预。通过测试主动性和代理性行为,揭示现有模型在动态多模态环境下的可靠性不足。
👍 9
10/26 08:00
Learning velocity fields directly in pixel space is difficult and slow. We show that coarse second-order image statistics can ease this task: frozen Wiener estimates improve velocity learning, with gains extending beyond the linear predictions they explicitly provide. Moreover, we observe similar bu
中文介绍 论文提出Wiener-Informed Decoding方法,简化像素空间流模型训练。利用冻结的Wiener估计(二阶图像统计量)辅助速度场学习,能改善训练效率,且增益超越线性预测。此方法降低像素空间流模型训练难度,加速收敛,且可能推广到其他生成任务。
👍 6
10/27 08:00
Language model agents are post-trained and deployed through a harness, the prompts, tools, and control logic that present a task to the model and execute its outputs. Harnesses are often revised and judged by how an already trained model performs under them, which assumes that what the model learned
中文介绍 论文揭示语言模型智能体的策略本质上是一种「Harness」,即提示、工具和控制逻辑的接口。当前Harness的修改和评估基于已训练模型的性能,忽略了策略对接口的依赖。该研究通过决策几何分析,说明策略与Harness的不可分性,为智能体部署设计提供新视角。
👍 6
10/26 08:00
AI agents must decide when another search is worth its cost. Answer-success probabilities capture question difficulty, while this decision requires estimating the additional benefit of search. We introduce SearchHorizonBench, a benchmark for learning and evaluating search value from execution experi
中文介绍 论文引入SearchHorizonBench基准,用于评估AI智能体的搜索价值决策能力。该基准要求智能体根据执行经验学习何时值得进行额外搜索,以平衡成本与收益。通过此基准,研究者可系统衡量智能体在不确定环境中优化搜索策略的能力。