👍 50
06/16 20:00
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 针对现有10B级别图像修复模型在计算成本上的障碍,提出了一种名为Moebius的轻量级框架。该框架通过极端结构压缩与任务特定优化相结合,达到接近10B级性能。实验结果显示,与传统模型相比,Moebius在保持效果的同时显著降低了计算需求。意义:为图像修复和轻量化模型设计提供了新思路。
👍 43
06/12 20:00
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 提出了DragMesh-2,一个可实现物理上合理的多指手与可动物体交互的方法。该方法针对动作变化自动生成相应的动态接触模式,提升了家用和辅助设备的操作能力。通过实验验证,DragMesh-2在处理复杂物体时表现出更高的灵活性与效率。意义:增强机器人对环境的适应性,推动人机交互技术的发展。
👍 40
06/16 20:00
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提出了一种基于语言指令的3D点轨迹预测方法,旨在提升机器视觉智能中的动作预测能力。该方法利用世界坐标中的3D点提供了一种通用表示,支持更复杂的物体运动规划与物理交互。实验结果表明,该方法在多种真实场景中均有效能显著提升。意义:为智能体计划与推理提供了新的范式。
👍 31
06/16 20:00
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 研究了Playful Agentic Robot Learning,提出了一种新方法,使机器人能够自主编写并执行可执行的策略程序。与传统的任务驱动学习相比,新方法强调灵活的技能获取和行为调整,有利于提升机器人的适应性与效率。实验表明,该系统在多个任务中显示出更强的学习能力。意义:推动智能体自我学习与适应能力的研究。
👍 25
06/17 20:00
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引入了一种新的空间工具使用智能体框架,以应对现实世界中的动态三维空间推理挑战。与静态视觉观察不同,该框架使得智能体能够利用工具进行推理与决策。实验结果显示,S-Agent在几种复杂任务中显著提高了空间理解能力。意义:旨在改善智能体在动态环境中的操作能力。
👍 20
06/16 20:00
Reinforcement learning (RL) has become a representative post-training paradigm for LLMs, enabling strong reasoning and agentic capabilities. However, rollout generation remains a dominant latency bottleneck because autoregressive sampling decodes responses sequentially and a small number of long-tai
中文介绍 EfficientRollout提出了一种系统感知的自我推测解码方法,以改善强化学习(RL)中的回合生成延迟。通过改变自回归采样策略,该方法可加速长尾任务的处理,提升响应速度和效率。实验结果显示,该方法能够在较短时间内实现更优的决策质量。意义:推动强化学习在实际应用中的效率提升。
👍 19
06/17 20:00
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 Beyond Static Leaderboards这项研究聚焦于LLM代理评估的预测有效性,整合了多个维度的评估,以更全面地了解其在实际部署中的表现。通过对多个并行实施的新资产和性能指标进行大规模协调,研究推进了LLM评估的全面性和科学性。意义:为LLM的评估标准化提供新的方向,推动实际应用落地。
👍 16
06/15 20:00
Reinforcement learning pipelines for Large Language Model (LLM) training often rely on manually redesigned environments between stages, requiring practitioners to heuristically infer which configuration will best improve the current policy. To automate this process, we propose the LLM-as-Environment
中文介绍 提出了一种新的LLM训练环境设计方法,旨在自动化强化学习中的环境重定义过程。通过引入LLM作为环境设计者,研究显示该方法能有效简化环境配置,提高当前策略的优化效率。实验验证了所提方法在复杂多代理场景中的有效性。意义:促进强化学习与大型语言模型的协作发展。
👍 15
06/17 20:00
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旨在实现社区LoRA挖掘下风格-内容双重参考生成的自由控制。通过在风格与内容之间平衡,该方法能够合成更具表现力和美感的图像。实验结果表明,相较于传统方法,FreeStyle在内容保真度与风格呈现上均表现出更高的灵活性。意义:为生成艺术和图像处理领域提供新方法。
👍 12
06/16 20:00
Passive models for long video understanding typically rely on a "watch-it-all" paradigm, processing frames uniformly regardless of query difficulty, causing computational cost to grow with video duration. Although interactive frameworks have emerged, they often rely on global pre-scanning, and their
中文介绍 研究Native Active Perception的可交互框架,用于增强全模态理解的推理能力。通过使用个性化预扫描方法,动态调整模型处理策略,从而降低复杂查询带来的计算成本。实验验证该方法在长视频理解任务中的有效性。意义:为提升视频智能理解和处理效率提供新思路。
👍 9
06/04 06:26
Multicultural multi-agent systems are increasingly deployed in globally diverse settings, where different agents are grounded in different cultural backgrounds. Existing cultural evaluation focuses on value alignment: how closely a single agent matches a target culture. Yet alignment is a per-agent
中文介绍 Beyond Alignment探讨了多文化多代理系统中价值多样性的问题,提出与单一文化对齐不同,必须考虑多个代理的文化背景。研究显示,价值多样性可提高系统的适应性和协作能力,为评估和优化多文化代理提供了新的视角。意义:推动智能系统的文化敏感性和多样性考虑。
👍 8
06/15 20:00
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定义了社交世界模型在实时音视频生成中的应用,致力于为社交互动提供更为真实的生成内容。通过构建原型,该模型能够更好地理解和反应社交环境的动态特征。实验结果显示,在社交内容生成上表现出色。意义:促进社交智能体与内容生成的相互影响和发展。
👍 8
06/16 20:00
Reinforcement Learning with Verifiable Rewards algorithms like GRPO have emerged as the dominant post-training paradigm for complex reasoning in LLMs, yet commonly suffer from policy entropy collapse during training. We conduct a first-order gradient analysis of token-level entropy dynamics under GR
中文介绍 STARE提出了一种基于惊讶度引导的策略熵重加权方法,旨在解决强化学习中策略熵崩溃的问题。通过对令牌级熵动态进行分析,提出了一种新的训练策略,有效提升了训练过程的稳定性与策略质量。实验结果显示,该方法可显著改善LLM的推理性能。意义:为强化学习稳定性问题提供了新策略。
👍 7
06/17 20:00
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 DF3DV-1K推出了一个大型数据集及基准,专注于无干扰的新的视角合成。这一创造性的数据集为研究领域提供了一个可靠的基准,推动光场技术在现实场景中的应用测试。实验表明,新数据集推动了光场技术的进一步发展。意义:为新视角合成研究开拓了更广泛的应用方向。
👍 7
06/17 20:00
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针对现实环境中的灵巧机器人政策自我改善进行了研究。该方法通过生成自适应代码以优化算法搜索,减少了对人为监督的依赖,提升了机器人在复杂任务中的表现。实验结果验证了该方法的有效性和机器人能力的提升。意义:助力实现更全面的物理智能。
👍 6
06/11 20:00
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提出了一种可靠的特征上采样方法,以解决视觉Transformer(ViT)在处理小补丁令牌网格时的局限性。该方法优化了自注意机制,以平衡性能与计算效率。实验结果表明,ViT-Up在多个视觉任务中均表现出色。意义:为视觉学习架构的高效性提供了新思路。
👍 5
06/17 20:00
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扩展了LiveCodeBench以支持多种编程语言,利用竞争性问题创建不断更新的测试集。通过集成多种语言的评估,Multi-LCB提升了对大型语言模型(LLMs)在代码生成任务中的性能评估的全面性。实验结果验证了其在多语言环境中的有效性。意义:为多语言代码生成的评估提供了新标准。
👍 4
06/14 20:00
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 Understanding the Behaviors of Environment-aware Information Retrieval探讨了检索增强生成(RAG)模型在处理复杂查询时的性能差异,强调不同检索器对查询表达策略的需求。该研究提供了一种前期分析方法,优化了信息检索在不同查询中的应用。意义:提升信息检索与生成过程的协同作用。
👍 4
06/14 20:00
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 Thinking with Visual Grounding提出了一种方法,将视觉推理与图像证据结合起来,解决现有视觉-语言模型(VLM)中推理痕迹的不可验证性。该模型能够明确表示支持推理的图像区域,提升了推理的透明度与可监督性。实验结果证明了其在多种应用场景中的有效性。意义:推动视觉推理研究走向更深层次的应用探索。
👍 4
06/14 11:45
Frontier scientific reasoning remains a major challenge for large language models (LLMs), where even the strongest commercial systems fall short of expert-level performance. A closer look at model behavior reveals substantial complementarity that single-model evaluation hides: different frontier mod
中文介绍 SciOrch致力于协调多种专家LLM以解决前沿多模态科学推理任务。研究表明,不同模型之间的互补性可以显著增强整体性能,提供了在复杂科学任务中的合作新模型。实验结果显示,该方法比单一模型在多模态推理中表现优异。意义:促进多模态学习和科学推理的交叉研究。