👍 202
09/27 20:00
Latent visual reasoning (LVR) enables multimodal large language models (MLLMs) to perform intermediate computation in continuous latent tokens rather than expressing every reasoning step in words. However, unlike textual CoT, latent reasoning is not directly observable, making it difficult to superv
中文介绍 本文探讨了潜在视觉推理(LVR),为多模态大型语言模型(MLLMs)在连续潜在标记中执行中间计算提供了基础。研究显示,与文本中的链式推理(CoT)不同,潜在推理难以直接观察,这使得监督学习成困难。此项研究可能对改进视觉和语言推理的应用具有重要意义。
👍 187
09/29 20:00
Self-evolving search agents build their own training curricula by jointly optimizing a proposer that generates questions and a solver that answers them. This closed loop introduces a failure mode we call co-cheating: the proposer and solver increasingly agree on shared errors, so internal reward imp
中文介绍 提出了一种针对自演化搜索代理中共同作弊的诊断与缓解方法。自演化的搜索代理通过优化提问者和解答者来构建训练课程,但这种闭环系统可能导致两者在共享错误上达成共识,影响内部奖励效果。此研究对提升自适应学习模块的可靠性具有重要意义。
👍 98
09/29 20:00
Terminal agents act through stochastic model generations, yet the ability to generate a useful action does not ensure its reliable execution. A poor command (e.g., wrong package install) can change the environment in ways that hinder subsequent progress, even when the model could generate a better a
中文介绍 研究了终端代理中模型与执行环境之间的动作扩展能力。通过生成随机模型,本文揭示了命令生成与可靠执行之间的差距,使得执行不佳的命令可能影响后续进展。此成果对提高代理的复杂任务处理能力具有重要意义。
👍 63
09/28 20:00
Large language models (LLMs) are increasingly deployed as agents for multi-step decision-making, yet transfer poorly to unseen environments. World-model methods address this by training agents to predict future observations, at the cost of additional training and errors that compound when prediction
中文介绍 针对大型语言模型(LLMs)在未知环境中决策转移能力差的问题,提出了EVOKE方法,通过训练代理预测未来观测来增强世界模型能力。尽管引入额外的训练成本和预测错误,该方法显著提升了多步骤决策的有效性,对智能代理的实际应用具有重要影响。
👍 31
09/28 20:00
Long-form subtitle translation requires reasoning over discourse and cultural context spanning episodes or entire series, while maintaining consistent terminology and style. Existing single-LLM methods are largely sentence-level, and multi-agent systems often use static workflows that do not adapt t
中文介绍 本文提出了一种自演化的多代理系统用于长篇字幕翻译,能够在跨剧集或系列的语篇和文化背景下进行推理,并保持术语和风格的一致性。这种动态适应能力大幅提升了翻译的准确性与连贯性,具有广泛的应用潜力。
👍 30
09/28 20:00
Recent advances in musculoskeletal modeling and reinforcement learning have enabled muscle-actuated agents to reproduce increasingly complex human motions. Yet these capabilities remain largely confined to flat ground, in part because motion datasets rarely include aligned terrain geometry and becau
中文介绍 研究了肌肉骨骼运动的地形感知重建、重定向和控制方法,结合近期的肌肉建模与强化学习进展,提高了复杂人类动作的再现能力。在现实世界复杂地形中的应用意义重大,有助于提升机器人运动控制的灵活性与适应性。
👍 24
09/28 20:00
We introduce LoopVL to study whether Loop Transformers can be effectively extended to vision- language models. LoopVL combines Module-Loop and Model-Loop computation to iteratively update a unified vision-language state through shared modules. We train LoopVL from scratch through language pre-traini
中文介绍 提出LoopVL,旨在研究是否可以将Loop Transformers有效扩展至视觉-语言模型。LoopVL结合了模块循环与模型循环计算,通过共享模块迭代更新统一的视觉-语言状态。该方法从语言预训练开始训练,展现出新的模型构建视角,具有较强的研究价值。
👍 24
09/29 20:00
Scientific software presents a demanding test for computer-using agents based on visual language models (VLMs): completing a research workflow requires interpreting specialized interfaces, manipulating scientific objects, and producing verifiable results. We thus introduce OSWorld-Science, a benchma
中文介绍 OSWorld-Science是一个针对学习和使用科学软件的计算机使用代理基准,旨在评估基于视觉语言模型(VLMs)的代理在复杂研究工作流中的表现。该基准的提出对推动科学分析自动化具有重要的参考价值。
👍 17
09/25 20:00
Reward models score responses from large language models (LLMs) and guide LLM training toward human preferences. However, reward models can favor superficial attributes such as length or confidence, leading LLMs to produce higher-scoring but not more correct responses. Existing mitigation methods ei
中文介绍 BiasReducer方法旨在通过自适应方式减轻奖励模型中的偏见,解决大型语言模型(LLMs)训练中对表面特征的偏好问题。研究表明,这一方法能够有效提升模型的推理准确性,具有显著的应用前景。
👍 15
09/28 20:00
Modern agentic systems combine an AI model with a harness that controls execution and environmental interactions. Harness design strongly affects long-horizon performance, yet its combinatorial search space demands substantial human effort that must be repeated as models change. Existing automated m
中文介绍 提出MILO,通过协调多代理演化实现自动化的执行框架发现,解决了传统代理系统中执行控制对长期表现的重要影响。这一方法显著降低了人类设计干预的需求,对智能代理设计具有深远影响。
👍 15
09/29 20:00
Looped transformers and Mixture-of-Experts (MoE) offer complementary routes to efficient scaling: recurrence increases computational depth at fixed parameters, while MoE sparsity expands total capacity at fixed active compute. Yet existing scaling laws model recurrence or sparsity in isolation. In t
中文介绍 研究了循环混合专家(MoE)和循环变换器的扩展法则,认为两者的结合提供了一种有效的扩展途径。该研究打破了以往对循环与稀疏计算的单独建模,推动了大型语言模型在高效计算方面的研究进展。
👍 13
09/25 20:00
Software development requires more than editing code: developers repeatedly run software, interact with its interfaces, visually inspect its behavior, and use these observations to decide what to change next and whether a change works. Existing coding agents and computer-use agents are largely studi
中文介绍 CUA-SWE探讨计算机使用代理在视觉软件工程中的应用,强调开发过程中的交互与反馈。该研究有助于提升编码代理与计算机使用代理的综合性能,推动软件开发的智能化进程。
👍 8
09/28 20:00
TTS systems with autoregressive semantic modeling have demonstrated strong zero-shot voice cloning performance and rich expressive variation, but their sequential decoding incurs substantial latency. Non-autoregressive alternatives offer much faster generation, yet often rely on more restrictive ref
中文介绍 Tacit-TTS通过从自回归解码转向掩蔽预测,大幅提升了无转录语音克隆的效率。该方法在保留语音表现力的同时显著降低了生成延迟,为现实应用中的语音合成技术提供了新的可能性。
👍 6
09/29 20:00
Latent communication enables multi-agent systems to exchange information directly in internal representation space, reducing the token, computation, and latency overhead of text-based communication. To this end, lightweight trainable links are introduced to map the sender's representations into the
中文介绍 本文研究了多代理系统中潜在通信的安全性,探讨了直接在内部表示空间交换信息的方法,减少了基于文本的通信的延迟和计算开销。该研究为多代理协作提供了新的高效信号传递机制,具有重要应用前景。
👍 6
09/29 20:00
Self-attention gives LLMs fine-grained, query-dependent access to context, but dense token interactions incur quadratic prefill cost and a key--value cache growing with context length. Research thus spans explicit-memory compression, sparse access, recurrent state construction, structured state dyna
中文介绍 分析了大规模语言模型中注意力机制的演变,指出了自注意力在上下文访问的精细化与计算成本之间的权衡。研究提出了多种策略,以提升模型的性能与效率,对于未来的 LLM 研究方向具有启发意义。
👍 5
09/29 20:00
We study training LLM judges from natural language feedback, especially for subjective tasks where the verdict depends strongly on which evaluation criteria the judge invokes and how it weighs them. The dominant approach, outcome-supervised RL (e.g., GRPO), credits every token in the rollout with a
中文介绍 研究了如何通过位置选择自蒸馏训练大型语言模型(LLM)评判者,尤其在主观任务中,评判标准和权重的选择对结果影响显著。这一方法为提高评判的准确性提供了新思路,具有重要的后续应用价值。
👍 5
09/27 20:00
Large language model (LLM) routing aims to select the most suitable model for each incoming query. Most existing routers learn this decision directly from query embeddings, model representations, preference data, or clusters of similar examples. Such approaches can be effective, yet the representati
中文介绍 提出SeLMRoute,为大型语言模型(LLM)的路由选择提供概率语义证据,以优化查询响应的模型选择。该方法超越了传统的基于嵌入的方法,可能在多任务处理和适应性学习方面产生重要影响。
👍 4
09/27 20:00
Artificial-intelligence (AI) agents hold promise for automating bioimage analysis, yet no benchmark evaluates whether they can carry out real-world analyses end to end. Such analyses are hard for agents because 2D images, 3D volumes and time-lapse sequences are often too large to read as context, so
中文介绍 BIABench建立了一个评估AI代理在现实生物图像分析任务中的表现的基准,填补了学术界对AI代理能否执行完整生物分析流程的评估空白。这一研究将促进AI在生物医学领域的应用与发展。
👍 4
09/29 20:00
We find that major reported improvements in decoding words from non-invasive brain recordings are largely reproducible without any brain data. In the influential work of d'Ascoli et al. (2025), time series of brain activity from subjects perceiving continuous speech are segmented into fixed-length w
中文介绍 研究发现,从非侵入性脑电记录中解码语言的主要改进在没有脑数据的情况下也是可再现的。本文对现有技术进行了系统分析,为未来的脑-文本解码研究提供了新的视角,推动了神经信号解读领域的发展。
👍 4
09/28 20:00
Skills equip LLM agents with professional knowledge and guidance to complete long-horizon and complex tasks. Although skills have been widely adopted in recent agent paradigms and harnesses, how to synthesize reliable training data and how to train agents for skill use remain underexplored. In this
中文介绍 SkillGym研究如何通过自动生成可验证的环境来训练具备技能的代理,以完成复杂任务。该方法扩展了技能在最新代理系统中的应用,解决了训练数据合成与模型适应性的问题,具有显著的实践意义。