👍 272
10/09 20:48
Multimodal language models (MLLMs) exhibit promising spatial reasoning capabilities, yet the internal mechanisms supporting them remain poorly understood. We uncover S-Space, a linear representation subspace in which object-token representations encode continuous coordinates along horizontal, vertic
中文介绍 针对多模态大模型空间推理机制不透明问题,提出 S-Space:在 object token 表征中发现一个线性子空间,可编码水平、垂直及深度连续坐标。无需额外参数即可显式解码物体位置,为理解与增强 MLLM 空间推理提供新视角。
👍 247
09/30 08:00
Precise grounding matters. It specifies which object is the target and where that object is, even in clutter and for tiny objects, and it has to be fast enough for closed-loop control. Yet vision-language-action (VLA) and world-action models (WAMs) take perception from general-purpose vision-languag
中文介绍 面向物理智能提出 GroundingPI,以视觉基元为基础实现高精度 grounding 基础模型。针对 VLA/WAM 模型在杂乱场景和微小物体上定位不准且速度不足的问题,统一物体识别与定位,并满足闭环控制对实时性的要求。
👍 215
10/08 08:00
Large Language Models (LLMs) increasingly ship with explicit "thinking modes", yet their counterpart, "no-thinking", has received far less attention. We study LLMs' no-thinking behavior along two axes. a. How to measure no-thinking? Prior work typically defines no-thinking through proxies such as a
中文介绍 研究 LLM 的 no-thinking 行为。提出从两方面度量「不思考」,发现即使被要求不思考,模型仍持续进行内部推理,形成「thinking inertia」。揭示了显式思考模式之外被忽视的对偶现象,为推理控制提供新证据。
👍 201
10/09 08:00
As the production of software passes into the hands of software itself, coding agents are increasingly commissioned to build agents fitted to their users' own workflows. This delegation carries a quiet assumption, that the delivered agent will faithfully serve the requirement it was built for, rathe
中文介绍 提出 AgentSWE,评测编码智能体能否真正构建用户所需的 Agent。指出委托编码智能体代建 Agent 时存在交付物无法忠实满足需求的隐忧,通过系统化评测检验其可靠性与可用性。
👍 133
10/09 08:00
Terminal agents are moving beyond local operations toward completing real-world professional tasks. Short trajectories can demonstrate commands and tool calls, but they do not fully show how agents accumulate evidence through exploration, use it to inform judgments, continually adjust strategies bas
中文介绍 面向长时终端智能,提出扩展 agentic 数据的规模化方案。强调短轨迹无法展示智能体在探索中累积证据、据以判断并持续调整策略的能力,通过构建长时轨迹数据提升终端智能体完成真实专业任务的能力。
👍 78
10/07 08:00
Video interactive storytelling enables viewers to actively steer how a video unfolds. However, once we allow viewers to intervene during generation, a new challenge arises: The viewer's request can have latent dependencies on both the grounding source and the current rendered video state. These depe
中文介绍 提出 VIS-Ground,实现带上下文锚定的视频交互式叙事。解决观众在生成过程中干预时,请求对原始素材和当前视频状态的双重潜在依赖难以跟踪的问题,保证交互后叙事的一致性。
👍 73
09/30 08:00
Autoregressive (AR) grounding models serialize spatial predictions, introducing sequential latency and imposing a causal order on output tokens. We view grounding as visual evidence extraction: objects, locations, and spatial relations are jointly constrained by the image and query, yet their depend
中文介绍 提出 GroundAnything,将 grounding 视为视觉证据提取,用并行解码替代自回归串行预测。消除 AR 模型的顺序延迟,使物体、位置和空间关系联合约束,实现接近实时的精准视觉定位。
👍 34
10/09 08:00
Attention Residuals (AttnRes) enables content-dependent retrieval over network depth, which can be viewed as single-stream softmax attention along the depth axis. Inspired by the success of multi-head attention along the sequence axis, we introduce Multi-Stream Attention Residuals (MSAR), an extensi
中文介绍 提出 Multi-Stream Attention Residuals (MSAR)。在 Attention Residuals 沿深度轴做单流注意力基础上,借鉴多头注意力的成功,引入多流扩展,实现深度方向的内容依赖检索,提升网络容量利用效率。
👍 30
09/27 08:00
Personal Graphical User Interface (GUI) agents rely on interaction history to infer what a user wants from ambiguous instructions and to anticipate recurring routines. Existing approaches retrieve task-relevant history and append it to the policy's context, implicitly assuming that experience releva
中文介绍 针对个人 GUI 智能体,指出相关性不等于可用性。现有方法直接检索相关历史并拼入上下文,隐含假设相关经验可用。本文提出经验激活机制,区分检索与利用,提升智能体从交互历史中推断用户意图的能力。
👍 26
10/08 08:00
Speculative decoding accelerates large language model inference by using a lightweight drafter to propose multiple tokens that are subsequently verified in parallel by the target model. Recent diffusion-based drafters make this paradigm particularly attractive: deep attention-based neural backbones
中文介绍 面向推测解码中的扩散草稿模型,提出对并行草稿的上下文校正方法。利用扩散草稿器并行生成多个 token 的优势,通过上下文感知校正提升草稿接受率,从而在保持推理加速的同时改善目标模型验证效率。
👍 21
10/08 08:00
Humans perceive far more in a scene than what is explicitly depicted: a single glance captures past causes and future trajectories; a quick peek determines if a vehicle can fit between two parked cars; a few seconds of video reveals who holds authority in a room; and a fleeting clip highlights subtl
中文介绍 提出 Humanity's Sixth Sense 基准,评测多模态模型的直觉视觉推理能力。覆盖人类能快速感知但场景未显式呈现的信息,如车辆能否通过、视频中权力关系、微妙社会线索等,衡量模型的深度场景理解水平。
👍 20
09/29 08:00
Language-model agents are usually trained by reinforcement learning from one reward per episode, and privileged self-distillation enriches it by letting the same policy, given a skill, teach its skill-free self through token probabilities. However, we identify two phenomena that question this channe
中文介绍 提出 PR-OPD,面向 agentic 强化学习的特权表征在策略自蒸馏。指出每回合仅一个奖励信号信息稀疏,现有特权自蒸馏依赖 token 概率存在局限,通过特权表征丰富学习信号以提升智能体训练效果。
👍 20
09/29 08:00
Historical street-view imagery records urban evolution, but uneven coverage leaves substantial gaps in historical records. Generating plausible past appearances requires restoring changed structures while preserving persistent scene content. We construct VIGOR-his, a decade-spanning cross-view datas
中文介绍 构建 VIGOR-his,一个跨十年级别的街景跨视角数据集,并提出奖励引导的历史街景生成方法。在不均匀覆盖条件下恢复已变更结构同时保留持久场景内容,填补历史城市影像记录的空白。
👍 17
10/04 08:00
Recent GPU kernel optimization agents retain what they learn in knowledge bases or as distilled skills. Kernel benchmarks score each attempt's implementation for correctness and speed but leave the reuse value of retained experience unmeasured. We introduce MetaKernelBench, which measures whether ex
中文介绍 提出 MetaKernelBench,衡量 GPU kernel 优化智能体的知识迁移能力。现有基准仅评分单次实现的正确性与速度,忽视知识库中留存经验的可复用价值。本文专门评估超出代码本身的知识转移效果。
👍 16
10/09 08:00
When solving complex spatial reasoning tasks, ordinary humans naturally revisit the scenes from task-relevant viewpoints and mentally reconstruct the spatial layouts from those viewpoints. But can Large Vision-Language Models (LVLMs) proactively think in the same way? We term this ability Active Spa
中文介绍 提出 SpatialASM,将主动心理模拟能力内化到空间推理中。受人类从任务相关视角重建空间布局启发,让 LVLM 主动切换视角、重构空间关系,提升复杂空间推理任务的表现。
👍 11
10/07 08:00
The evaluation of coding agents generally strives to be as realistic as possible. In our study, we collect 4,782 agent sessions of real software engineers in JetBrains IDEs, which we call Production Sessions. Since our subject is interactive agents, we study the sessions with at least three user mes
中文介绍 指出编码智能体评测应匹配用户真实工作流。收集 JetBrains IDE 中 4,782 段真实工程师会话(Production Sessions),筛选含至少三轮用户交互的会话进行评测,使 benchmark 更贴近实际使用场景。
👍 11
10/08 08:00
Knowledge distillation (KD) transfers knowledge from stronger Teacher models to weaker Student models, but most methods require training the Student parameters, thereby binding the distilled knowledge to a specific architecture and checkpoint. This implicit representation is difficult to interpret o
中文介绍 提出 Universal Textual Teaching,以纯文本方式将教师模型知识传递给学生 LLM。区别于需训练学生参数的传统蒸馏,无需绑定特定架构或 checkpoint,提升蒸馏知识的可解释性与可移植性。
👍 10
10/02 08:00
We reconstruct high-fidelity 3D scene meshes from posed multi-view images without per-scene optimization, across scales ranging from single objects to large outdoor scenes. Per-scene optimization methods lack the learned 3D prior needed when observations are sparse or surfaces are glossy or transpar
中文介绍 提出 T3lescope,从多视角图像无逐场景优化地重建高保真 3D 网格,尺度覆盖单物体到大型室外场景。利用学到的 3D 先验,在观测稀疏、表面高光或透明时仍保持重建质量,突破逐场景优化方法的局限。
👍 10
10/07 08:00
Vision-and-Language Navigation (VLN) has largely focused on a single agent following a single instruction, yet many real-world applications require teams of robots to tackle tasks beyond the capabilities of any individual agent. We present Systematic Multi-Agent Vision-and-Language Navigation, provi
中文介绍 提出系统性多智能体视觉语言导航(Multi-Agent VLN),提供问题形式化、基准与解决方法。面向多机器人团队协作完成任务的场景,超越单智能体跟随单条指令的设定,适配现实多智能体协作需求。
👍 7
10/09 08:00
Large language models have achieved remarkable advances in complex reasoning. However, scaling reasoning post-training with ever-larger corpora can substantially increase training cost without proportional performance gains. In this study, we therefore ask: how far can reasoning supervision be compr
中文介绍 提出不确定性引导的渐进收缩(Progressive Contraction),实现高效推理后训练。指出现有方法通过不断扩大数据集提升推理能力,训练成本增长远超收益,本文探索在更少监督数据下达到相近推理效果。