👍 203
09/23 08:00
Object permanence and solidity are hallmarks of human cognitive priors. Recent studies show that video generation models, a paradigmatic class of current world models, have begun to show emerged reasoning abilities, making them ideal candidates for building human-like physical intelligence. Do video
中文介绍 研究如何训练视频世界模型获得物体恒存性。分析视频生成模型能否在遮挡后保持对物体连续存在的推断,探讨其与人类物理认知的差距,提出在训练中引入物体持久性目标的方法,期望赋予世界模型更贴近人类的物理智能与推理能力。
👍 84
09/22 08:00
Long-term conversational memory in multi-party settings requires more than retrieving relevant content from long-term conversations: it must distinguish who said what, whom each statement concerns, how individuals perceive one another, what information is shared by the group, and how states change o
中文介绍 提出 SpeakerMem-R1,一种面向多方对话的双轨记忆机制。长期对话不仅需检索相关片段,还需区分说话人身份、发言所指对象、个体间感知关系以及共享信息与状态变化。该方法以说话人为中心构建记忆结构,提升多方场景的长期推理能力。
👍 75
09/24 08:00
While Large Language Models (LLMs) rely on highly non-linear components, in this work we demonstrate that they exhibit fundamental linearity: when inputs from distinct text streams are linearly combined, the model outputs a superposition of the individual next-token distributions. We term this the S
中文介绍 揭示 LLM 存在线性叠加特性。尽管 Transformer 由高度非线性组件构成,但研究发现当不同文本流输入线性组合时,模型的 next-token 分布也呈叠加。这项工作为可解释性与模型行为分析提供了新的线性视角。
👍 48
09/19 08:00
Spatial reasoning is essential for vision-language models (VLMs) to understand and act in the physical world. Reasoning in dynamic environments requires VLMs to perceive local state transitions caused by object motion and viewpoint changes and integrate them over long trajectories to maintain an upd
中文介绍 提出 Spatial-Interactor,通过 VLM 与可观测物理世界交互学习空间推理能力。在动态环境中,模型需感知物体运动与视角变化引起的局部状态转换,并在长轨迹上整合以维持更新后的空间表征,用于改进物理世界的理解与行动。
👍 43
09/21 08:00
Evaluating world models requires assessing both the quality of the worlds they generate and their consistency and responsiveness under exploration, interaction, and modification. We introduce HappyWorld-Bench, a comprehensive benchmark that evaluates whether generated worlds remain reliable as agent
中文介绍 发布 HappyWorld-Bench,一个评估世界模型的综合基准。除衡量所生成世界的质量外,还考察其在探索、交互与修改下的连贯性与响应能力,验证生成的世界在 agent 持续操作时是否仍保持可靠,用于统一评估世界模型。
👍 39
09/23 08:00
Advances in generative models have improved video fidelity, enabling long-horizon generation, interactive world modeling, and evolving visual environments. Autoregressive (AR) video generation extends visual sequences through causal rollouts. However, a fundamental bottleneck emerges: as the generat
中文介绍 面向自回归(AR)视频生成研究记忆机制。AR 通过因果 rollout 延长视频序列,但随着生成推进出现关键瓶颈:早期内容影响衰减、长程一致性难以维持。本文提出记忆模块以提升长视野生成、交互世界建模与动态视觉环境的一致性。
👍 35
09/23 08:00
Agentic memory systems reuse past experience to improve future performance, yet most existing designs curate memory at write time: once a task is completed, its trajectory is distilled into a fixed artifact, such as a reflection, workflow, skill, or reasoning strategy, that is later retrieved by sim
中文介绍 提出 Just-in-Time Memory,让 LLM agent 学会在任务执行时即时挑选任务自适应记忆。现有方法多在写入阶段固化反思、技能或推理策略,缺乏灵活性。本文在运行时动态策展记忆,提升 agent 复用过往经验改进未来任务的能力。
👍 22
09/20 08:00
Humans can effortlessly localize the direction of a sound source and integrate it with visual cues for reasoning, yet this remains challenging for embodied agents. In particular, it is still unclear how to effectively evaluate and model spatial audio understanding in embodied settings. To address th
中文介绍 提出 OmniEcho,面向具身 agent 的空间音频理解。旨在解决如何有效评估与建模具身场景下的空间音频,帮助 agent 像人一样定位声源方向并将其与视觉线索整合进行推理,推进听觉感知的具身智能。
👍 19
09/23 08:00
Recent advances in large language models (LLMs) have enabled agents to tackle long-horizon tasks across diverse environments. To further improve agent performance, existing language world models typically predict environment observations, yet reconstructing high-entropy, execution-dependent tool res
中文介绍 提出 Agent-Editing World Model,重新思考 LLM agent 的世界建模。现有语言世界模型常预测环境观测,但难以重建高熵、执行依赖的工具结果。本文转而建模 agent 对环境的编辑与状态演化,改善长程任务的决策能力。
👍 17
09/24 08:00
Rufus-Air is an open and reproducible post-training recipe on GLM-4.5-Air-Base (106B-A12B), organized as a serial pipeline of eight stages: SFT, Reasoning RL, Coding RL, Instruction-Following RL, General Agent, Coding Agent, Search Agent, and RLHF. We document the data, reward design, infrastructure
中文介绍 提出 Rufus-Air,一个开源可复现的 LLM 后训练配方,基于 GLM-4.5-Air-Base(106B-A12B)。由 SFT、推理 RL、代码 RL、指令遵循 RL、通用/代码/搜索 agent 与 RLHF 八个串行阶段组成,公开数据、奖励设计与基础设施细节。
👍 16
09/22 08:00
Reinforcement learning has become a central component of large language model (LLM) post-training, yet token-level credit lacks a generally accepted mathematical definition, leaving its relationship to commonly used training signals unclear. We formulate three regularity conditions, namely Completen
中文介绍 提出 PACT,将 token 级 credit assignment 与 critic alignment 关联。作者指出 RL 后训练中 token 级信用缺乏公认的数学定义,提出三条正则性条件,理清信用分配与常见训练信号的关系,使强化学习后训练更可控。
👍 15
09/22 08:00
The rapid capability gains of frontier language models are widely attributed to improved reasoning abilities, yet this cannot be verified as raw CoT traces in closed-source systems are hidden. By registering a simple custom tool through a standard API feature, we induce frontier models to externaliz
中文介绍 研究前沿模型的隐藏思维链。由于闭源系统的原始 CoT 不可见,作者通过标准 API 注册自定义工具诱导模型外化其推理,从而提取并刻画隐藏 CoT,用以分析模型能力来源并验证其推理行为。
👍 12
09/23 08:00
Language-model agents increasingly face long-horizon tasks with evolving state, interdependent decisions, and delayed outcomes. Scaling their training requires diverse agentic environments, dependable outcome signals, and low extension cost. Existing generation pipelines commonly construct an enviro
中文介绍 提出可验证隐藏动态博弈,从已求解机制生成 agentic RL 环境。面对长程、状态演化与延迟回报的任务,作者提供多样化环境、可靠结果信号与低扩展成本的生成方式,用于规模化训练语言模型 agent。
👍 11
09/23 08:00
Recently, Large Language Models (LLMs) have been increasingly able to solve advanced mathematical problems, including many that have been open for decades. This opens the door to expansion of mathematical knowledge at unprecedented scale. Yet, while LLMs may be able to conjecture and prove more and
中文介绍 研究如何让 LLM 学会发现有趣的数学问题。尽管模型已能解决包括数十年未解的难题,但如何引导其提出有新意的猜想与定理仍存挑战。本文探索扩大数学知识边界的策略,使模型在推测与证明基础上主动发现新问题。
👍 11
09/24 08:00
Task and motion planning (TAMP) problems remain difficult even with full observability and object-centric states because discrete decisions are tightly coupled to geometric, kinematic, and dynamic constraints. Generalized TAMP addresses this difficulty by exploiting regularities across problem insta
中文介绍 将编码 agent 用于广义任务与运动规划(TAMP)问题。即使在全观测、以物体为中心的状态下,离散决策与几何、运动学、动力学约束高度耦合仍难解。该方法利用跨问题的规律性,用编码 agent 泛化处理 TAMP。
👍 11
09/24 08:00
The rapid progression of large language models is extending AI from passive content generation into the active workflows of engineering and scientific discovery. This shift raises a compelling question: can AI be both the object of development and an active participant in building next-generation AI
中文介绍 提出 Qwen-Planner-Agent,一个用于真实移动 planner agent 的闭环 AI-for-AI 框架。让 AI 既是开发对象又是参与构建下一代 AI 的主动方,将大模型能力从被动内容生成推向工程与科学发现的主动工作流。
👍 11
09/20 08:00
Robotic bin packing requires long-horizon sequential decision-making, as each object placement affects the available space for subsequent packing. Existing methods primarily rely on hand-crafted geometric heuristics that optimize predefined objectives or reinforcement learning policies learned throu
中文介绍 提出 PackLab,面向机器人装箱的多模态模型开发、训练与评估框架。装箱需长程顺序决策,每次放置都会影响后续可用空间。现有方法依赖手工几何启发或强化策略,本文统一流程以提升装箱决策能力。
👍 11
09/23 08:00
Vision-Language-Action models provide a strong foundation for general-purpose robot control, yet a vast majority of policies do not preserve and leverage episode-level information beyond the current observation. This limitation is consequential in history-dependent manipulation tasks that depend on
中文介绍 提出 MemBodied,为视觉-语言-动作模型引入循环联想记忆。多数策略无法保留并复用当前观测以外的 episode 级信息,在依赖历史的操控任务中受限。本文以关联记忆保留并利用整段经验,提升泛化操控性能。
👍 10
09/23 08:00
We present Hunyuan-A13B, an open-source large language model based on a Mixture-of-Experts architecture. It contains 80 billion total parameters but activates only 13 billion during inference, balancing model capability, computational efficiency, and deployment cost. The model is pretrained on a rig
中文介绍 发布 Hunyuan-A13B,基于 MoE 架构的开源大语言模型。总参数 800 亿,推理时仅激活 130 亿,平衡能力、计算效率与部署成本。模型在大规模数据上预训练,兼顾表现与经济性。
👍 9
09/24 08:00
Deep search requires LLM agents to decompose complex queries, search for evidence, and synthesize grounded answers, yet existing ReAct-style agents suffer from two limitations: role coupling, where one policy must handle planning, evidence use, and synthesis; and context accumulation, where growing
中文介绍 提出 IterSynth,通过角色解耦的迭代合成重新思考深度搜索 agent。现有 ReAct 风格 agent 存在角色耦合——单一策略同时负责规划、证据使用与合成——以及上下文不断累积的问题,本文解耦各角色并迭代合成以改进证据驱动的搜索回答。