👍 97
06/17 08:00
While 10B-level industrial foundation models have pushed the boundaries of image inpainting, their prohibitive computational costs severely hinder practical deployment. Constructing a highly optimized task-specific specialist offers a promising solution; however, extreme structural compression inevi
中文介绍 提出 Moebius,一种仅 0.2B 参数的轻量图像修复框架,性能媲美 10B 级大模型。通过极端结构压缩实现高效部署,解决了工业大模型计算成本过高的问题。
👍 61
06/13 08:00
Dexterous interaction with articulated objects is important for household, assistive, and humanoid manipulation, where multi-finger hands can provide compliant contact patterns beyond parallel-jaw grasping. However, articulated-object manipulation differs from static-object manipulation: the target
中文介绍 DragMesh-2 实现了对铰接物体的灵巧手-物交互,生成物理合理的抓取与操作。区别于静态物体,该方法处理多关节物体的动态接触模式,适用于家庭、辅助和人形机器人场景。
👍 41
06/17 08:00
Motion forecasting is central to visual intelligence: agents must anticipate how objects will move in order to plan actions, reason about physical interactions, and synthesize realistic futures. We argue that 3D points in world coordinates provide a general representation that is class-agnostic, vie
中文介绍 MolmoMotion 提出用语言指令预测世界坐标系下 3D 点的运动轨迹,实现类别无关的通用运动预测。该方法将点轨迹作为视觉智能的通用表示,支持动作规划、物理推理和未来合成。
👍 35
06/17 08:00
Current agentic robot systems can write executable Code-as-Policy programs, observe feedback, and revise behavior across multiple attempts, but they remain largely task-driven: reusable skills are acquired only after explicit instructions. We study Playful Agentic Robot Learning, where an embodied c
中文介绍 研究「玩耍式具身智能体学习」,让机器人通过自由探索而非明确指令来获取可复用技能。该方法使编码智能体在无任务驱动下自主习得技能库,拓展了机器人学习范式的边界。
👍 27
06/18 08:00
Real-world spatial intelligence requires reasoning over a continuous and evolving 3D world, yet existing VLMs and tool-augmented agents largely remain tied to static, stateless inference from isolated visual observations. We introduce \textsc{S-Agent}, a spatial tool-use agentic paradigm for underst
中文介绍 S-Agent 提出空间工具使用代理范式,让 VLM 通过调用外部工具(如地图、测量)在连续、动态的 3D 世界中进行空间推理。解决了现有模型依赖静态、无状态观察的局限。
👍 26
06/18 08:00
LiveCodeBench (LCB) has recently become a widely adopted benchmark for evaluating large language models (LLMs) on code-generation tasks. By curating competitive programming problems, constantly adding fresh problems to the set, and filtering them by release dates, LCB provides contamination-aware ev
中文介绍 Multi-LCB 将 LiveCodeBench 扩展至多编程语言,持续收集多语言竞赛题目并按发布时间过滤,提供防污染评估。填补了现有代码生成 benchmark 仅覆盖单一语言的空白。
👍 20
06/18 08:00
Agent benchmarks are growing fast, but no single benchmark touches more than four or five of the dimensions that deployment exposes. This paper aggregates the largest coordinated deep-dive of one MCP-based industrial-agent benchmark to date: fourteen parallel implementation studies covering new asse
中文介绍 超越静态排行榜,提出基于 MCP 的工业智能体 benchmark 的预测有效性验证框架。通过 14 项并行实现研究,评估基准在部署维度上的泛化能力,为智能体评测提供新视角。
👍 18
06/18 08:00
Style-content dual-reference generation aims to synthesize an image that preserves the structure and semantics of a content reference while adopting the style of a separate style reference.Despite recent progress, this setting remains challenging because models must balance content fidelity, style a
中文介绍 FreeStyle 实现风格-内容双参考图像生成,通过挖掘社区 LoRA 实现灵活的风格控制。在保持内容结构的前提下精确迁移风格,解决了双参考生成中内容保真与风格迁移的平衡难题。
👍 9
06/18 08:00
Advances in radiance fields have enabled photorealistic novel view synthesis. In several domains, large-scale real-world datasets have been developed to support comprehensive benchmarking and to facilitate progress beyond scene-specific reconstruction. However, for distractor-free radiance fields, a
中文介绍 发布 DF3DV-1K 大规模无干扰新视角合成数据集与基准,支持全面测评。填补了无干扰辐射场领域大规模真实世界数据的空白,推动超越场景特定的新视角合成研究。
👍 8
06/16 08:00
As an increasing majority of global video content is consumed on social platforms for interactive social purposes, video generation models built for social worlds are important but largely overlooked by previous studies. In this work, we define the position of social world models and build a prototy
中文介绍 MaineCoon 构建实时音视频社交世界模型,为社交平台上的互动视频生成提供原型。定义了社交世界模型的位置,填补了以往视频生成模型忽视社交交互用途的空白。
👍 7
06/18 08:00
Achieving dexterous robotic manipulation in the real world heavily relies on human supervision and algorithm engineering, which becomes a central bottleneck in the pursuit of general physical intelligence. Although emerging coding agents can generate code to automate algorithm search, their successe
中文介绍 ENPIRE 提出具身机器人策略的自主改进框架,让编码智能体在真实世界中自我优化算法代码。减少对人工监督的依赖,向通用物理智能迈出关键一步。
👍 6
06/12 08:00
Vision Transformers (ViTs) have become a dominant architecture for visual representation learning, providing exceptionally strong and broadly reusable backbone features. However, ViTs are commonly operated on relatively small patch-token grids due to the quadratic cost of global self-attention, whic
中文介绍 ViT-Up 提出忠实特征上采样方法,解决 Vision Transformer 因全局注意力二次复杂度而只能使用小尺寸 patch 网格的问题。在不引入额外参数下恢复高分辨率细节特征。
👍 4
06/15 08:00
Large language models (LLMs) often fail when answering requires identifying a small but decisive piece of evidence within a long or complex context, such as a single line in a tool trace or a subtle detail in an image. We propose ContextRL, a context-aware reinforcement learning (RL) method that imp
中文介绍 ContextRL 提出上下文感知强化学习方法,提升 LLM 在长复杂上下文中定位关键证据的能力(如工具日志中的一行、图像中的细微细节)。通过 RL 优化模型对决定性线索的敏感度。
👍 4
06/15 08:00
Recent retrieval-augmented generation (RAG) approaches have demonstrated strong capability in handling complex queries, yet current research overlooks a critical challenge: different retrievers require fundamentally different query formulation strategies for optimal performance. In this work, we pre
中文介绍 系统研究环境感知信息检索的行为,揭示不同检索器需要不同的查询生成策略。提出针对性的检索优化方法,填补了 RAG 中检索器特性对齐的空白。
👍 4
06/17 08:00
Multi-step LLM pipelines fail through interactions among retrieval, reasoning, and formatting steps, so prompt-only optimization can miss bottlenecks in the chain. We present FAPO (Fully Autonomous Prompt Optimization), a framework that lets Claude Code optimize an LLM pipeline inside a standardized
中文介绍 FAPO 实现多步 LLM 管线的全自动提示优化。让 Claude Code 在标准化框架内自主优化检索、推理、格式化等步骤的交互瓶颈,无需人工干预。
👍 4
06/15 08:00
Visual thinking should not only sound right; it should show its evidence. While recent vision-language models (VLMs) can produce natural-language reasoning traces, these traces often leave the supporting image regions implicit, making them hard to verify and difficult to supervise. We introduce visu
中文介绍 提出视觉思维的 grounding 推理方法,让 VLM 在生成自然语言推理轨迹时明确标注所依据的图像区域。增强可验证性,便于监督,解决了隐式视觉证据的问题。
👍 3
06/18 08:00
Policy-adherent tool-calling agents in customer-service domains must maintain task states across turns while calling tools and obeying domain policies. Task states consist of relevant facts, identifiers, constraints, and conditions observed through user interaction and tool calls. In standard agents
中文介绍 LedgerAgent 为客服领域的策略合规工具调用智能体设计结构化状态管理。维护跨轮对话中的事实、约束与条件,确保工具调用遵循领域策略,提升任务稳定性。
👍 3
06/17 08:00
Progress in legal AI increasingly depends on access to authoritative legal text at scale. Yet one of the most consequential layers of American law remains largely absent from existing machine-readable corpora: local ordinances. Local codes govern zoning, housing, business licensing, public health, n
中文介绍 发布 LOCUS 美国地方条例语料库,填补了法律 AI 中地方级法律文本的空白。涵盖分区、住房、商业许可等法规,为法律 NLP 提供可扩展的权威数据基础。
👍 3
06/18 08:00
Embodied foundation models are expected to benefit from data scaling like large language models, but face a much tighter data bottleneck. Teleoperated real-robot trajectories remain the dominant pretraining source due to their precise action supervision and embodiment alignment, yet their scalabilit
中文介绍 HumanScale 证明第一人称人类视频在具身预训练中可超越真实机器人数据。利用大规模人类日常活动视频进行预训练,缓解真实机器人数据稀缺与扩展瓶颈。
👍 3
06/16 08:00
Offline reinforcement learning is typically analyzed under process-level reward supervision, yet many sequential decision datasets
record only trajectory-level outcomes. We develop a statistical theory for offline policy optimization from such outcome-level
supervision. We first study the canoni
中文介绍 离线强化学习中,当只有轨迹级结果(而非过程奖励)时如何高效学习?提出统计理论,分析从包含成功/失败信号的轨迹中优化策略的条件与算法,扩展了离线 RL 的适用场景。