👍 241
08/09 20:00
We introduce BDH-CQ, a reasoning model that combines in-context learning with recurrent latent reasoning. Inputs presented at inference time continuously update the model's recurrent memory; the model then solves a query through iterative computation in a high-dimensional latent space, without verba
中文介绍 提出了 BDH-CQ 模型,将上下文学习与递归潜在推理相结合。该模型通过迭代计算在高维潜在空间中解决查询,并在推理时持续更新模型的递归记忆。核心创新在于不依赖于传统的监督方法,以实现更灵活的推理能力。意义:推动了在复杂推理任务中 Agent 能力的提升与应用。
👍 200
08/09 20:00
Macaron-V1 is an open agent-model family for experiential intelligence: learning from experience in real environments and continuing to learn after deployment. It is organized around two system goals. Adaptation is pursued through recursive improvement of versioned model-harness pairs, where experie
中文介绍 Macaron-V1 是一个开放的智能体模型家族,旨在通过在真实环境中的经验学习实现持续自我改进。该方法通过对版本化模型与工具组合的递归改进进行适应,推动智能体在部署后不断学习。意义:引领持续学习和适应性智能体的发展,提升智能体在现实环境中的表现。
👍 116
08/09 20:00
As AI coding agents take on increasingly complex, long-horizon software engineering tasks, existing benchmarks are rapidly saturating and their evaluation quality has come under serious scrutiny: a recent audit found that nearly 60% of unsolved SWE-bench Verified instances contain flawed tests -- ei
中文介绍 SWE-Bench ProMax 旨在评估在大型多语言代码重构任务中 AI 编码代理的表现。该基准检测了现有基准的不足,特别是存在的测试缺陷,提供更可靠的评估标准。意义:为软件工程中的 AI 应用提供标准化评估,提高了任务复杂性的应对能力。
👍 66
08/07 20:00
We present Ouroboros, a self-developing agent harness whose tools, prompts, context assembly, and core implementation improve through reviewed commits that become the runtime for later work. Core evolution proceeds in two modes. In recursive free evolution, improvement is itself a task, and completi
中文介绍 Ouroboros 是一个自我发展的编码代理,通过审查的提交来改进其工具、提示和核心实现,进而提升效率。该模型采用递归自由演化与任务驱动的改进方式,对智能体能力的不断增强提供了新的视角。意义:有效推动智能体自我优化技术的发展,适用于复杂编码任务。
👍 52
08/08 20:00
On-policy (Self-)Distillation (OPD / OPSD) has shown strong potential for post-training large language models (LLMs). However, existing methods still rely heavily on external supervision, including ground-truth signals, environmental feedback, or guidance from larger models, and therefore fall short
中文介绍 提出了一种无监督的 On-Policy Self-Distillation (OPSD) 方法,显示出对后训练大语言模型的强大潜力。该方法突破了依赖外部监督的限制,旨在提高模型性能,从而提升应用效果。意义:推动了无监督学习与模型自我优化方向的探索,可能影响多种智能应用。
👍 33
08/09 20:00
We introduce Motif 3, a decoder-only Mixture-of-Experts language model with 314 billion total parameters and 13.2 billion activated per token. Each sparse MoE layer contains 384 routed experts, with eight selected per token. This fine-grained sparsity provides substantial expert capacity while limit
中文介绍 Motif 3 是一个解码器-仅 Mixture-of-Experts 语言模型,拥有 3140 亿个参数,并在每个 token 上激活 132 亿个参数。该模型通过细粒度的稀疏性增强了专家容量,并提高了计算效率。意义:为大规模语言模型的技术发展提供新的范例,改善模型的推理与生成能力。
👍 31
08/09 20:00
Leading large language model providers now conceal their models' step-by-step reasoning, or chain-of-thought, to protect intellectual property and limit information leakage. Rather than storing these traces server-side, providers return them to the client as blocks of encrypted text, which the clien
中文介绍 研究了从专有 LLM API 中提取推理痕迹的方法,挑战了当前模型提供商的知识产权保护。该方法强调了通过加密文本形式返回推理过程,突破了传统束缚。意义:推动机器学习模型透明度及开放研究方向,影响智能应用和知识产权管理的策略。
👍 31
08/06 20:00
Memory systems have shown promise for improving agent performance, but their potential remains largely unexplored for small language models, which struggle to generate sufficient successful trajectories on their own. We propose Agent Memory Distillation (AMD), a training-free framework that transfer
中文介绍 提出了 Agent Memory Distillation (AMD) 框架,用于改善小型语言模型的代理性能。AMD 在不需要额外训练的情况下,通过传递层次化教师记忆来提高智能体生成有效轨迹的能力。意义:为小型模型的性能提升提供新思路,有助于增强智能体在复杂任务中的适应能力。
👍 27
08/03 20:00
Human evaluation of AI systems and digital products is costly, slow, and difficult to scale. Offline evaluations are more scalable but often abstract away human diversity and interactive behavior. We therefore introduce MatrAIx, a population-scale simulated-user evaluation infrastructure for testing
中文介绍 MatrAIx 是一个用于测试 AI 系统与数字产品的人类评估基础架构,解决了传统人类评估的规模与成本问题。通过模拟用户的方法,该系统可以更好地反映人类行为的多样性。意义:为模拟人类评估提供可扩展的解决方案,适用于 AI 和人机交互领域。
👍 23
08/09 20:00
We introduce Sci-VBench, a comprehensive benchmark for evaluating knowledge- and reasoning-intensive video generation across scientific domains. It contains 1,253 expert-annotated examples spanning 60 subjects across four core disciplines: Natural Science, Healthcare, Humanities & Social Sciences, a
中文介绍 Sci-VBench 是一个评估知识与推理密集型科学视频生成的综合基准,涵盖了多个学科的专家注释示例。其目标是提升科学领域视频生成模型的评估标准与应用潜力。意义:推动科学领域智能生成技术的发展,助力新兴的视频生成应用。
👍 22
08/02 20:00
Conversational assistants increasingly recommend follow-up edits to help users continue a task. Existing systems primarily target text-only interactions, leaving image-creation conversations underexplored. In image-creation tasks, useful follow-up edit suggestions must reflect user preferences, offe
中文介绍 探讨了在图像创建任务中,如何根据用户偏好提供有用的后续编辑建议。该系统旨在提升现有对话助手在图像编辑中的效果,以适应用户的实时需求。意义:促进图像生成任务中的个性化指南的发展,影响用户体验的提升。
👍 16
08/07 20:00
Large language model (LLM) inference serving is increasingly constrained by memory rather than compute. As long-context and long-form reasoning workloads become more prevalent, the key-value (KV) cache dominates both memory footprint and memory traffic during LLM token generation, i.e., decode. In p
中文介绍 OasisKV 提出了利用前瞻性稀疏预取方法来扩展关键值缓存的能力,以解决长上下文推理中的内存限制问题。该方法优化了在 LLM 解码过程中记忆流量的管理,从而提升了效率。意义:推动长上下文模型性能的提升,有助于复杂推理任务的实现。
👍 16
08/04 20:00
On-policy distillation (OPD) provides dense teacher supervision on student-generated trajectories, but standard reverse-KL training can assign insufficient probability to other plausible continuations. Teacher entropy alone does not reveal whether uncertainty is concentrated among a few plausible ne
中文介绍 SPOT 评估了 On-Policy Distillation (OPD) 中的稀疏探测与结果校准问题,以提高学生生成轨迹的监督质量。该研究识别出标准方法中的不足之处,提供了改进方向。意义:为强化学习中的优化方法提供新思路,有助于提高智能体的学习能力。
👍 16
08/05 20:00
CLI-based software-engineering agents have matured rapidly, yet the open ecosystem has converged on a single training environment: trajectory datasets used to fine-tune open models are collected almost exclusively under OpenHands. Models fine-tuned on this data score well under OpenHands but degrade
中文介绍 DCAS 通过解耦 CLI 代理支撑结构,提升了 Planning 跨支撑的内化能力。尽管 CLI 基础的智能体在快速发展,然而对训练环境的局限性却使模型适应性降低。意义:为软件工程代理技术的灵活性提供新路径,促进开放生态的进步。
👍 11
08/03 20:00
Small language models are often the only option for deployment under tight latency, cost, and on-premises constraints, but they are rarely trained from scratch: a compressed model is usually recovered through knowledge distillation (KD). This recovery step largely decides the final quality, yet it i
中文介绍 研究了高效的知识蒸馏方法,针对 LLM 在严格的延迟、成本和本地化约束下的应用,提出 Offline Top-K Logits 和融合分块 KL 损失。新方法改善了压缩模型的质量,推动模型在资源受限环境中的部署。意义:为低资源环境中的大语言模型应用提供支持,拓宽了 LLM 的应用场景。
👍 10
08/09 20:00
Large Language Models (LLMs) have driven rapid progress in autonomous agents, yet standard evaluations remain confined to static task solving. An emerging frontier is harness evolution---the agent's capacity to autonomously optimize its own operating harness. However, systematically benchmarking thi
中文介绍 Evo-Bench 关注于大型语言模型 (LLM) 如何改善智能体创建与优化的能力,探索代理自身操作环境的自动优化。该研究提出了系统性的基准评估智能体的进化能力。意义:促进智能体自我优化的研究,推动任务复杂性应对能力的提升。
👍 9
08/09 20:00
Learning-based memory systems for self-evolving LLM agents face two tightly coupled challenges. First, trajectory-indexed utilities grow with the interaction history, thereby dispersing limited feedback over an ever-expanding state space. Second, because trajectory-level rewards are jointly assigned
中文介绍 RoMeRL 解决了有关学习基础的记忆系统的两大挑战,提出通过降阶效用状态来平衡反馈覆盖与记忆奖励的困境。该方法有助于提升智能体在复杂状态空间中的学习效率。意义:为自我演化智能体的学习机制提供理论基础,影响未来的智能体研发。
👍 9
08/07 20:00
Vision-Language Models (VLMs) should answer from concrete image evidence rather than language priors, dataset shortcuts, or irrelevant visual context. Existing perception-aware post-training methods encourage image use through global perturbations or attention proxies, but they do not test whether a
中文介绍 Evidence-RL 旨在增强视觉语言模型 (VLM) 基于具体图像证据的推理能力,从而提升模型对图像的理解与应用效果。该研究挑战现有依赖语言推理的流行方法,强调了基于证据的重要性。意义:推动视觉推理领域的进步,提高模型在处理实际应用中的能力。
👍 9
08/06 20:00
The malicious use of generative artificial intelligence to create highly realistic deepfake videos raises serious ethical concerns and poses substantial challenges to AI safety. However, existing deepfake video benchmarks provide limited coverage of recent synthesis methods and generally lack reliab
中文介绍 提出了多智能体取证推理方法,以应对深度伪造视频所带来的挑战。通过集成化的评估框架,该研究强调了检测新型合成方法的重要性与必要性。意义:提升对深度伪造检测方法的有效性与实用性,增强 AI 安全性研究的方向。
👍 9
08/04 20:00
Hard prompt compression reduces long-context inference cost by independently scoring tokens, sentences, or chunks and retaining the highest-scoring units under a budget. We identify a structural failure in this procedure: independent selection can split dependent evidence pairs, retaining one member
中文介绍 探讨了硬提示压缩的缺陷,揭示了在高分散的选择过程中可能导致的参考未完整问题。该文提出了改进的方法,以优化选择机制并增强信息保留的稳定性。意义:为提高提示压缩技术的性能指明方向,影响多模态学习的研究。