👍 108
06/15 20:00
Looped Transformers scale latent computation by repeatedly applying shared blocks, but sequential looping increases latency and KV-cache memory with the loop count. Parallel loop Transformers (PLT) alleviate this cost through cross-loop position offsets (CLP) and shared-KV gated sliding-window atten
中文介绍 LoopCoder-v2 提出了针对 Looped Transformers 的优化方法,通过 Cross-Loop Position Offsets (CLP) 和共享 KV cache 的滑动窗口机制,显著降低了并行循环的延迟和内存消耗。实验表明,该方法在计算效率方面优于传统的循环机制,具有更好的扩展性,适用于高效计算及深度学习的实际应用中。意义:这项研究推动了 Transformer 模型在高效计算方面的发展,尤其适合大规模深度学习任务。
👍 101
06/14 20:00
Generalist robot policies must follow user instructions while reasoning about how objects, cameras, and robot actions interact in the 3D physical world. Recent vision-language-action models (VLAs) and video world-action models (WAMs) inherit strong semantic or temporal priors from large-scale founda
中文介绍 Geometric Action Model 提出了一种针对机器人政策学习的几何动作模型,旨在让通用机器人政策在执行用户指令时,合理推理物体、摄像头和机器人动作在三维物理世界中的相互作用。该模型结合了大型基础模型中的语义和时序先验,有助于提高机器人在复杂环境中的决策能力。意义:该研究推动了机器人智能化的发展,对自动化和人机交互应用具有重要影响。
👍 94
06/14 20:00
DreamX-World 1.0 is a general-purpose interactive text/image-to-video world model for controllable long-horizon generation. It supports camera navigation, revisits to previously observed regions, and promptable events across photorealistic, game-style, and stylized domains. Our data engine combines
中文介绍 DreamX-World 1.0 是一种通用互动世界模型,支持文本/图像到视频的可控长时间生成。该模型能够实现摄像机导航、重新访问以前观察到的区域以及基于提示的事件生成,适用于照片现实、游戏风格和风格化域的互动生成。其数据引擎结合了多种输入源,展示了强大的生成能力。意义:这项研究为虚拟现实和互动媒体生成领域提供了新的工具和方法,促进了生成模型的应用扩展。
👍 68
06/12 20:00
Efficient and scalable agentic intelligence requires models that can deliver both low-latency responses and strong reasoning capabilities while remaining practical to train, serve, and deploy. In this report, we present Ling-2.6 and Ring-2.6, a family of models designed to address this challenge at
中文介绍 Ling-2.6 和 Ring-2.6 是专为高效和可扩展的代理智能设计的模型系列,旨在实现低延迟响应和强大的推理能力,同时保持训练、服务和部署的实用性。通过引入新的架构,模型在多项测评中表现出色,尤其在响应时间和推理准确度上显著优于传统方法。意义:这项研究推动了大规模语言模型的应用,尤其是即时智能助手等领域。
👍 39
06/14 20:00
Vision-Language-Action (VLA) models benefit from large-scale and diverse embodied data, yet scaling robot trajectory collection is costly and labor-intensive. Recent advances show that large-scale egocentric human videos provide complementary real-world supervision in pretraining. However, joint tra
中文介绍 ACE-Ego-0 针对 Vision-Language-Action 模型提出了一种统一的策略,利用大型自我中心人类视频数据进行预训练,以弥补机器人轨迹收集的高成本与劳动强度。通过结合人类的数据,该方法在预训练阶段提高了模型的效果和适应性,从而促进了 VLA 模型的发展。意义:此研究在机器人学习和人机交互领域引入了新思路,具有重要的实际应用价值。
👍 38
06/15 20:00
Knowledge distillation transfers a teacher's competence to a small student but is brittle in the small-student regime: forcing the student to imitate logits from a much larger teacher concentrates it on the teacher's sharpest modes, hurting generalization on benchmark families beyond the training co
中文介绍 Zone of Proximal Policy Optimization 提出了将教师知识通过提示而非梯度进行蒸馏的新方法,解决了在小学生中的模型泛化能力不足的问题。通过对教师模型的非参数性约束,学生模型能更好地在更广泛的基准上进行迁移学习。该方法在多个数据集上展现了优越的泛化性能。意义:该研究推动了知识蒸馏的研究,尤其在增强模型泛化能力方面具有重要意义。
👍 35
06/15 20:00
Game generation is an emerging application of coding agents, requiring models to transform natural-language specifications into playable interactive systems. Unlike traditional coding tasks, game generation takes place within a game engine, where scripts, scenes, assets, rendering, and runtime inter
中文介绍 GameCraft-Bench 旨在测试代理智能在游戏引擎中从自然语言规格生成可玩游戏的能力。与传统编码任务不同,该方法强调在游戏引擎内进行生成,整合了脚本、场景和实时性能。初步结果表明,模型在生成游戏的各个方面均表现良好,具有实用性。意义:此研究为智能代理在游戏设计中的应用提供了新方向,可能改变游戏开发的格局。
👍 30
06/14 20:00
Effective personalized AI-assisted learning demands systems that can not only generate accurate learner-specific educational materials, but also dynamically adapt their instruction to diverse learners. However, existing educational agents have primarily focused on lecture content automation and simu
中文介绍 LectūraAgents 提出了一个多代理框架,旨在适应个性化 AI 辅助学习。该系统可以根据不同学习者的需求动态调整教学内容,突破了传统教育代理主要集中于讲座内容自动化的局限。实验表明,该方法在学习者的参与度和满意度方面有显著提升。意义:该研究推动了个性化教育技术的发展,对未来的教育方式具有深远影响。
👍 25
06/07 20:00
Clinical early warning systems built on electronic health records, in which clinical observations are recorded as irregularly sampled medical time series (ISMTS), must deliver both calibrated risk scores for patient triage and interpretable rationales that clinicians can verify. Large Language Model
中文介绍 TRIAGE 旨在通过对电子健康记录中的不规则采样医疗时间序列进行辩证推理,提高临床早期预警系统的可解释性。该系统能够提供可信的风险评分和可验证的推理依据,以便临床医生进行更加精准的患者分流,提升诊断效果。意义:这项研究在医疗人工智能领域具有重要价值,对提高临床决策支持系统的应用准确性和效率具有深远影响。
👍 23
06/15 20:00
On-policy self-distillation (OPSD) has proven effective for post-training large language models (LLMs), yet its application to diffusion LLMs (dLLMs) remains unexplored. Existing OPSD methods are inherently autoregressive-centric. They inject privileged information via left-to-right prefix condition
中文介绍 On-policy Self-distillation (OPSD) 在大型语言模型(LLMs)中的应用显示出有效性,但在扩散 LLMs(dLLMs)中的探索尚属首次。研究者提出了一种结合自回归特性的 OPSD 方法,通过注入特权信息以提升模型生成的质量,实验表明,该方法在多个任务上表现优越。意义:此研究为 LLM 和无监督学习的结合提供了新思路,对自我提升模型的研究具有重要意义。
👍 20
06/15 20:00
Memory has become a standard substrate for self-evolving agents, yet retaining experience is not the same as learning how to evolve through it. Existing memory agents can store trajectories, retrieve reflections, or accumulate skills, but often lack the holistic competence to select useful experienc
中文介绍 OPD-Evolver 提出了通过在政策蒸馏中融入记忆机制,以培养全面的自演进代理。虽然现有的记忆代理能存储经验,但缺乏识别和利用有价值经验的能力。该研究通过优化策略选择,提高了代理的演进能力。意义:这项研究开辟了智能体进化研究的新领域,对未来的自适应系统和智能体有重要应用潜力。
👍 11
04/07 20:00
Web agents act through long interaction sequences, yet existing benchmarks evaluate only terminal success, discarding all process information and offering little guidance on improvement. In this work, we conduct a process-level analysis of web agents. We introduce WebStep, a benchmark of 1,800 task
中文介绍 Where Did It Go Wrong? 引入了 WebStep 基准,进行 Web 代理的过程级评估,旨在通过 1,800 个任务序列来分析代理的实际表现。研究者强调,当前评估方法仅关注最终成功率,忽略了过程信息,提出的基准能够提供更细致的改进引用途径。意义:此研究提高了对 Web 代理的评估标准,推动了智能代理系统的可解释性研究。
👍 10
06/11 20:00
We introduce the Massive Video Embedding Benchmark (MVEB), a 23-task benchmark for video embeddings spanning classification, zero-shot classification, clustering, pair classification, retrieval, and video-centric question answering. We evaluate 33 models and find that no single model dominates: MLLM
中文介绍 MVEB 提出了一个涵盖 23 项视频嵌入任务的大规模基准,涉及分类、零样本分类、聚类、配对分类等多方面。该基准评估了 33 个模型,结果显示没有单一模型能够显著优于其他模型,为视频嵌入研究提供了多维度的比较视角。意义:该研究丰富了视频嵌入领域的标准,为后续研究提供了参考和方向。
👍 10
06/14 20:00
As LLMs advance, post-training reinforcement learning (RL) increasingly relies on multi-dimensional rewards to cultivate comprehensive capabilities. This shift demands new algorithms capable of optimizing diverse and potentially competing objectives simultaneously. To address this, existing methods
中文介绍 GD^2PO 提出了一个新的政策优化算法,针对 LLMs 的多重奖励冲突问题,能够同时优化多维度目标。通过解耦奖励结构,模型在面对复杂任务时表现出更强的协调性和稳定性,实验结果证明此方法在各种奖励设置下优于现有的优化方案。意义:此研究推动了强化学习领域的进展,尤其在多目标优化和复杂环境中的应用具有重要价值。
👍 9
06/13 20:00
Re-rendering an existing video from a novel camera viewpoint requires the output to follow the prescribed camera trajectory while preserving the appearance and dynamics of the original scene across every frame. Existing methods rely on per-frame pose embeddings, noisy point-cloud renderings, or impl
中文介绍 Track2View 提出了 4D 一致的相机控制视频生成方法,旨在从新视点重渲染现有视频,保持画面的一致性和动态特征。该方法借助配对的三维点轨迹,实现了视频的重渲染,并在多项任务中展现了出色的性能。意义:此研究为视频生成技术的应用提供了新的思路,能够在电影制作与虚拟现实等领域产生深远影响。
👍 8
06/14 20:00
Existing image editing methods can be generally categorized into textual instruction-based and visual prompt-based ones. Textual instructions are semantically expressive, but are limited by the coarse granularity of spatial control of the editing results. In contrast, visual prompts such as drag and
中文介绍 Text-Vision Co-Instructed Image Editing 探讨了文本和视觉提示在图像编辑中的结合,旨在克服现有方法中对空间控制的局限性。该方法展示了文本指令的语义表达能力,加上视觉提示的灵活性,提升了编辑的效果与用户体验。意义:此研究推动了图像编辑技术的进步,对设计和艺术创作领域具有重要影响。
👍 7
06/15 20:00
The shift from video generation to interactive world modeling places new demands on data: beyond captioned videos, world models require temporally aligned video-action-language trajectories grounded in the actions, camera motion, states, and events that drive future scene changes. However, such data
中文介绍 EgoCS-400K 提供了一个以自我视角为核心的游戏数据集,专为世界模型的构建设计,能够支持同步视频、动作和语言的轨迹生成。该数据集为模型开发提供了丰富的训练材料,提升了互动性和环境响应能力。意义:此研究为交互式世界建模提供了重要的数据支持,为游戏和仿真技术的发展奠定基础。
👍 7
06/13 20:00
Progress in AI has largely been driven by methods that assume less. As compute and data increase, approaches with weaker inductive biases generally outperform those with stronger assumptions. This is particularly characteristic of the field of Visual Representation Learning, where approaches have go
中文介绍 You Don't Need Strong Assumptions 提出了以时间差为基础的视觉表征学习方法,强调在计算和数据增加的背景下,弱假设方法通常能获得更好的效果。该研究通过实验验证了弱结论偏置在视觉学习中的有效性,具有广泛的应用潜力。意义:此研究推动了计算机视觉领域的发展,为新颖的学习策略提供了方向。
👍 6
06/15 20:00
Large language models perform increasingly well on standardized logical reasoning benchmarks, but whether this ability remains robust beyond English is unclear. We introduce ChLogic, an English--Chinese aligned benchmark that tests whether models preserve logical reasoning performance when the same
中文介绍 ChLogic 引入了中英对齐的基准,测试大型语言模型在中文表达中的逻辑推理能力,以确定其在非英语语境下的鲁棒性。实验证明,部分模型在中文逻辑推理上表现不佳,揭示了跨语言一致性的挑战。意义:此研究为多语言推理模型的研究提供了依据,对跨文化智能体的开发具有重要参考价值。
👍 6
06/14 20:00
Reinforcement learning with verifiable rewards (RLVR) improves language-model reasoning, but GRPO-style optimization remains prone to collapse. We analyse this instability through token-level gradient dynamics, deriving a taxonomy that predicts how updates affect next-token probabilities and entropy
中文介绍 A Gradient Perspective 研究了可验证奖励强化学习(RLVR)中的优化稳定性,分析了 token 层面的梯度动态,以确定更新如何影响下一 token 的概率及熵。此研究在复杂任务环境中提出了一种系统性的框架,有助于提高 RLVR 的优化效果。意义:该研究为强化学习算法的稳定性和鲁棒性改进提供了新思路,推动了 RLVR 的技术进步。