2026-08-11

Interpretable MEG Decoding of Perceived Speech: Cortical Sources and the Stimulus Features That Drive Retrieval

👍 68

08/02 08:00

Short segments of perceived speech can be retrieved from non-invasive magnetoencephalographic (MEG) recordings by deep networks trained with a CLIP-style objective against wav2vec 2.0 audio embeddings. Yet their weights do not map onto electrophysiological quantities, and it remains unclear which sp

中文介绍 针对深度网络从非侵入性脑磁图（MEG）中解码言语时权重难以对应电生理量、刺激特征不明的问题，提出可解释解码方法。利用CLIP式目标联合wav2vec 2.0嵌入训练，并分析皮层来源及驱动检索的刺激特征，揭示模型依赖的关键神经与声学信号。

From Economic Agents to Agentic Economies: A Systems Blueprint for Economic World Models

👍 33

08/06 08:00

Economic World Models (EWMs) are generative economic models that simulate how economies evolve from within by modeling heterogeneous agents, their beliefs and actions, and the market and institutional mechanisms through which their interactions produce aggregate outcomes. This paper develops an impl

中文介绍 提出经济世界模型（EWM）的系统蓝图，用于从内部模拟经济演化。模型刻画异质代理的信念与行动，以及市场与制度机制如何将交互聚合为宏观结果。文章给出实现方案并探讨该范式在政策分析与经济模拟中的潜力。

SFT Conflicts, RL Coexists: A Theoretical and Empirical Analysis of Multi-Task Learning for LLMs

👍 31

08/06 08:00

Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) exhibit fundamentally different behaviors in enhancing multi-task reasoning for large language models (LLMs). Our preliminary experiments revealed a phenomenon: SFT suffers from severe task conflicts under multi-stage training, whereas RL

中文介绍 发现多任务训练中大语言模型的行为差异：SFT在多阶段训练下存在严重任务冲突，而RL能更好共存。通过理论与实证分析两者机制，解释为何RL在提升多任务推理时更稳健，并为如何组合SFT与RL提供指导。

DataSpace: Benchmarking Data Agents for Verifiable Analytics over Heterogeneous Workspaces

👍 30

08/04 08:00

Data agents enable natural-language analytics over organizational workspaces, where relevant evidence may be scattered across databases, structured files, long documents, and multimedia. Existing benchmarks largely isolate structured querying, retrieval, or open-ended analysis, leaving heterogeneous

中文介绍 现有基准将结构化查询、检索与开放式分析割裂，难以评测真实异构工作区中的自然语言数据分析。提出DataSpace基准，涵盖数据库、结构化文件、长文档与多媒体混杂的证据场景，用于验证数据代理在可核查分析中的能力，并给出评测结果。

Teaching Nemotron Greek: Mining a Corpus, Adapting Retrieval, and Grounding Generation for Modern Greek across Specialist Domains

👍 29

08/05 08:00

Modern Greek is absent from NVIDIA's Nemotron retrieval models and from major multilingual retrieval benchmarks, despite being important for retrieval-augmented generation (RAG) in legal, energy, financial, and medical applications. We present an end-to-end adaptation of the Nemotron retrieval stack

中文介绍 现代希腊语在NVIDIA Nemotron检索模型及主流多语言基准中缺失，制约法律、能源、金融、医疗领域的RAG应用。提出端到端适配方案：挖掘语料、调整检索、结合领域知识约束生成，使Nemotron检索栈支持现代希腊语并提升专业领域效果。

Beyond Simply Environment Scaling: Designing Effective Environment Distributions for Multimodal Agent Learning

👍 28

08/06 08:00

Recent works train agents by constructing large-scale multimodal environment pools. However, we find that simply increasing the number of multimodal environments does not always benefit. We further analyze the limitations in current multimodal environment distributions through a series of experiment

中文介绍 研究发现单纯增加多模态环境数量并不总能提升智能体学习效果。通过系列实验分析当前环境分布的局限，提出设计有效环境分布的策略而非盲目扩展规模，从而改善多模态智能体的训练效率与泛化能力。

Activity Frames: Deterministic Screen-Activity Compilation for Agent Memory and Replay

👍 24

08/06 08:00

Computer-use agents pay full frontier inference to re-derive routines their user has already performed, because an agent's memory today records what the user said, not what the user did. We compile passively captured screen activity into agent memory with a deterministic, zero-model pipeline: it seg

中文介绍 针对电脑使用智能体记忆只记录用户说过什么而非做过什么，导致重复推理的问题，提出Activity Frames：用确定性零模型管道将被动捕获的屏幕活动编译为智能体记忆。该系统分割屏幕活动并结构化存储，支持记忆回放，减少冗余推理成本。

EffectLearner: World-Aware Object-Effect Reasoning for Real-World Video Object Removal

👍 20

08/06 08:00

Video object removal must eliminate not only the target object but also its induced effects while maintaining high-fidelity and spatiotemporally coherent restoration. Existing methods mainly learn object-effect correspondences implicitly from predefined effect categories and fixed data distributions

中文介绍 视频物体移除不仅要消除目标物体，还需清除其引发的影响并保持高保真时空一致性。现有方法隐式学习对象-效果对应，受限于预定义效果类别。提出EffectLearner，以世界感知方式显式推理对象效果，提升真实场景移除的泛化性与恢复质量。

YOLO-PEFT: Parameter-Efficient Fine-Tuning on YOLO Family

👍 15

08/07 08:00

Generic parameter-efficient fine-tuning (PEFT) methods transferred from language models can fail silently on real-time detectors, whose heterogeneous operators and detection-specific components impose placement constraints absent from regular Transformer stacks. We propose YOLO-PEFT, a structure-awa

中文介绍 通用PEFT方法直接迁移到实时检测器可能静默失效，因其异构算子与检测专属组件施加了Transformer塔中不存在的约束。提出结构感知的YOLO-PEFT，针对YOLO家族设计参数高效微调策略，在保持检测精度的同时降低可训练参数。

MatrAIx: Simulating the World with 8.3 Billion Persona Agents

👍 14

08/04 08:00

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 针对AI系统与数字产品人工评估成本高、难以规模化的问题，提出MatrAIx，一个83亿人格代理的人口规模模拟用户评估基础设施。它通过模拟多样化用户交互行为，在离线环境下大规模测试系统，兼顾可扩展性与人类多样性。

StreamArena: Toward Continuous, Interactive, and Long-Horizon Agentic Streaming Video Understanding

👍 14

08/06 08:00

Deploying autonomous multimodal agents in continuous, real-world environments requires them to ingest unbounded audio-visual streams and maintain hour-scale memory. However, current evaluations predominantly rely on brief clips and multiple-choice formats. This design allows minimal baselines that p

中文介绍 现有视频理解评测多基于短视频片段和选择题，使简单基线即可获得高分，无法反映真实连续环境需求。提出StreamArena，面向持续、交互、长时程智能体流式视频理解的新评测框架，要求智能体处理无限音视频流并维持小时级记忆。

MASS: Multiplayer World Models with Authoritative Shared State

👍 14

08/06 08:00

Current video world models struggle in multiplayer environments because they entangle world state with view-dependent visual latents, leading to redundant compute, view inconsistencies, and poor scalability. We propose MAS (Multiplayer world models with Authoritative Shared State) to resolve this li

中文介绍 多玩家视频世界模型常将世界状态与视角相关的视觉潜变量纠缠，导致冗余计算、视角不一致与扩展性差。提出MASS（Multiplayer world models with Authoritative Shared State），分离权威共享状态与视角渲染，提升多玩家环境中的效率与一致性。

ContextMaster: Interactive Multi-Shot Video Creation via Fixed-Budget Sparse Context Routing

👍 14

08/05 08:00

Recent video models increasingly support generation, reference conditioning, and editing within a single model, yet typically expose them as separate operations over fixed inputs. Practical creation unfolds across multiple shots, requiring one model to generate from text, follow a reference, or edit

中文介绍 现有多模态视频模型虽能生成、参考条件化与编辑，却将其作为固定输入的独立操作，难以支撑多镜头创作。提出ContextMaster，通过固定预算稀疏上下文路由实现交互式多镜头视频创建，让同一模型灵活地根据文本、参考或编辑需求生成下一镜头。

DCAS: Decoupling CLI Agent Scaffolding to Internalize Planning across Scaffolds

👍 12

08/06 08:00

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 开源CLI软件工程智能体生态的训练轨迹几乎全部来自OpenHands，导致微调模型在该环境下得分高但在其他脚手架下性能下降。提出的DCAS解耦CLI智能体脚手架，将规划能力内化到各脚手架中，提升跨环境泛化性与鲁棒性。

GaussianSelector: Lightweight Human-Guided Object Selection in 3D Gaussian Splatting with Graph Optimization

👍 10

08/02 08:00

Selecting a complete 3D object from a reconstructed scene with minimal user effort is essential for practical scene editing and embodied interaction. Existing 3DGS-based methods either retrain the Gaussian representation to embed per-object labels, or build dense multi-view SAM observations, both re

中文介绍 现有3DGS物体选择方法要么重训练高斯以嵌入每对象标签，要么构建密集多视图SAM观测，均开销巨大。提出GaussianSelector，通过图优化在3D高斯泼溅中实现轻量级人工引导的物体选择，仅需少量用户交互即可完整选中物体，无需重训练。

Weights or Skills? A Survey of Robot-Learning Techniques: from Action-Predicting Weights to Robots that Write their Own Skills

👍 10

08/03 08:00

Robot learning is splitting into two bets: policies that bake competence into frozen weights (vision-language-action, or VLA, models), and agents that write and refine their own executable skills as code. This survey organises the field around that axis of weights versus skills. Its central analytic

中文介绍 机器人学习正分化为两条路线：将技能固化在权重中的VLA模型，以及以代码形式自写自改技能的程序化智能体。本综述围绕“权重vs技能”轴组织该领域，分析两条路线的理论基础、能力边界与融合趋势，为后续研究提供结构化视角。

FactorJEPA: Factorizing Monolithic Futures into Layout-Agent-Interaction Channels for Crowded and Chaotic Global South Urban Worlds

👍 10

08/02 08:00

World models have attracted significant attention for their ability to capture and predict the structure and dynamics of the physical world. In this emerging landscape, Joint Embedding Predictive Architectures (JEPA) offer a particularly compelling direction. We study a largely unexplored regime:

中文介绍 研究JEPA在拥挤混乱的全球南方城市这一少被探索场景中的应用。提出FactorJEPA，将整体未来预测分解为布局-智能体-交互通道，分别建模静态布局、动态代理及相互作用，从而提升对高密度、非结构化城市环境的预测能力。

Efficient Knowledge Distillation for LLMs: Offline Top-K Logits and a Fused Chunked KL Loss

👍 9

08/04 08:00

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 针对小模型知识蒸馏效率低的问题，提出离线Top-K Logits方法预先筛选关键类别，并设计融合的分块KL损失，避免在线计算完整分布。在保持蒸馏质量的同时显著降低训练开销，使资源受限场景下的小模型压缩更高效。

Reinforcement Learning with Evolving Rubrics as Rewards for Audio Reasoning

👍 9

08/04 03:46

Audio reasoning is essential for machine understanding of the acoustic world. Reinforcement learning with verifiable rewards can elicit such reasoning, yet existing reward designs are complementary in their limitations: outcome-based rewards supervise only the final answer and let the model reach it

中文介绍 现有音频推理的强化学习奖励设计各有局限：基于结果的奖励只监督最终答案，过程奖励则难以覆盖完整推理。提出使用演化规则（Evolving Rubrics）作为奖励，动态调整评估标准，结合结果与过程信号，有效提升音频推理能力。

Douyin Multimodal Embedding Model Technical Report

👍 8

08/03 08:00

Multimodal representation learning is a cornerstone of modern AI. By encoding multimodal queries and targets into vectors, it powers industrial search and recommendation and underpins modern agents. Real-world platforms with complex modalities and massive-scale content, such as Douyin, Xiaohongshu,

中文介绍 该技术报告介绍抖音多模态嵌入模型，面向复杂模态与海量内容的工业搜索与推荐场景。模型统一编码多模态查询与目标为向量，支撑智能体与检索系统，报告详述其架构、训练策略及在大规模平台上的部署经验与效果。

AI Settles a 25 Year-old Problem We Left Behind

@DimitrisPapail · 28.5K 粉丝 · 142.1K 阅 · 522 赞 · 86 转

08/09 02:30

Last week, GPT-5.6 and Claude Fable appear to have settled an open theoretical question in wireless communications that was intensely studied between 2000 till the 2010s and that I briefly worked on

中文介绍 GPT-5.6 与 Claude Fable 最近似乎解决了一个无线通信领域的开放性理论问题，该问题在 2000-2010 年代受到密集研究，博主本人也曾短暂涉足。这展示了前沿模型在工程理论问题上的潜力，属于 AI 科研应用的新案例。

Grok Build's most important features you should know

@XFreeze · 253.7K 粉丝 · 60.9K 阅 · 515 赞 · 67 转

08/08 16:03

In under 12 weeks, Grok Build has evolved from a terminal coding assistant into a full agentic software engineering platform....becoming one of the most powerful agent harnesses in the world. And

中文介绍 Grok Build 在不到 12 周内从终端编码助手成长为完整的 agentic 软件工程平台，博主称其已是全球最强大的 agent harness 之一，并介绍其关键功能。属于工具更新/测评类内容，突出其进化速度和能力。

Profiling LLM Inference with SGLang and Torch Profiler

@jino_rohit · 6.9K 粉丝 · 20.3K 阅 · 519 赞 · 59 转

08/08 12:35

This blog is my attempt to learn how to profile an llm served using sglang and its in built torch profiler integration that comes out of box. The idea here would be try and understand when I hit

中文介绍 一篇 LLM 推理性能剖析教程：作者记录学习用 SGLang 内置的 Torch Profiler 集成来分析服务性能，重点理解何时会遇到瓶颈。属于实操型技术分享，适合用 SGLang 部署的朋友参考。

AI professors are negotiating the new realities of academic research

MIT Tech Review AI · 08/11 04:00

This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here. Last week, I headed 30 miles south of San Francisco to a hotel in Mountain View, California, to join some of the most accomplished, and some of the most promising

中文介绍 MIT科技评论报道，AI教授们正协商适应学术研究的新现实。作者前往加州山景城，与一些最有成就的学者共同探讨AI对学术的影响。

What building an AI-native finance function taught me

OpenAI News · 08/11 01:00

OpenAI CFO Sarah Friar shares five lessons for building an AI-native finance function, from automated forecasting to stronger controls and AI ROI.

中文介绍 OpenAI首席财务官Sarah Friar分享打造AI原生财务部门的五点经验，包括自动化预测、强化管控和衡量AI投资回报率。

Build Low-Latency Multilingual Voice Agents: Open Weights & Full Deployment Control with NVIDIA Magpie TTS

Hugging Face Blog · 08/11 00:25

中文介绍 NVIDIA发布Magpie TTS，支持构建低延迟多语言语音代理，提供开放权重和完整部署控制，便于开发者自主部署。

OpenAI’s letter to Governor Abbott on responsible AI infrastructure in Texas

OpenAI News · 08/10 22:00

OpenAI sent Governor Greg Abbott a letter outlining its commitment to responsible AI infrastructure in Texas. The letter supports reliable, transparent growth that benefits Texans.

中文介绍 OpenAI致信得克萨斯州州长Greg Abbott，承诺在得州建设负责任的人工智能基础设施，支持可靠、透明的增长，使得州居民受益。

Model ML completes finance work more efficiently with GPT-5.6 Sol

OpenAI News · 08/10 20:00

Model ML uses GPT-5.6 Sol to carry finance work from research and analysis through editable, traceable PowerPoint decks and Excel workbooks.

中文介绍 Model ML使用GPT-5.6 Sol更高效完成金融工作，从研究和分析到生成可编辑、可追溯的PowerPoint演示文稿和Excel工作簿。

Making Knowledge Distillation Cheap Enough to Run at Scale

Hugging Face Blog · 08/10 18:05

中文介绍 Hugging Face博客刊文，探讨如何降低知识蒸馏的成本，使其能够大规模运行，为高效模型蒸馏技术提供可行路径。

Expanding Daybreak as the Cyber Defense Window Narrows

OpenAI News · 08/10 18:00

Meet GPT-5.6-Cyber, OpenAI’s cybersecurity-specific model available through Daybreak Red for authorized vulnerability research, exploit validation, and security testing.

中文介绍 OpenAI推出网络安全专用模型GPT-5.6-Cyber，可通过Daybreak Red用于授权漏洞研究、漏洞验证和安全测试，以应对日益收窄的网络防御窗口。

Putting frontier cyber models in more trusted hands

OpenAI News · 08/10 18:00

Approved Daybreak partners can use OpenAI’s frontier cyber models to deliver authorized, governed cybersecurity services to customers.

中文介绍 OpenAI表示，经批准的Daybreak合作伙伴可使用其前沿网络模型，向客户提供经授权且受治理的网络安全服务，旨在将前沿网络模型交到更可信的机构手中。

AI for science needs reasoning, not just data

MIT Tech Review AI · 08/10 17:00

Every few decades, someone announces that science has reached its end. In 1903, the revered physicist Albert Michelson wrote that the “facts of physical science have all been discovered.” In the 1980s, Stephen Hawking predicted that theoretical physics might be finished by the end of the century. Wi

中文介绍 MIT科技评论文章认为，AI用于科学需要推理能力，而不仅是数据。文章回顾物理学家阿尔伯特·迈克耳孙1903年的断言及霍金对理论物理学的预测，警示科学终结论。

These startups are chasing the next big thing in LLMs

MIT Tech Review AI · 08/10 17:00

MIT Technology Review’s What’s Next series looks across industries, trends, and technologies to give you a first look at the future. You can read the rest of them here. Way back in the summer of 2017, AI researchers at Google put out a paper called “Attention Is All You Need,” in which they describe

中文介绍 MIT科技评论「What's Next」系列报道，多家初创公司正追逐大语言模型的下一个重大突破。文章回溯2017年谷歌AI研究员发表的论文《Attention Is All You Need》。

Premium seats are coming to ChatGPT Business

OpenAI News · 08/10 08:00

Premium seats are coming to ChatGPT Business. Sign up by August 20 to get $100 in workspace credits and unlock higher usage for your team's most demanding work.

中文介绍 OpenAI宣布ChatGPT Business将推出高级席位，8月20日前注册可获100美元工作区积分，并解锁更高使用额度，满足团队最繁重工作需求。

How Zapier transformed core marketing processes with ChatGPT Work

OpenAI News · 08/10 08:00

The enterprise marketing team at Zapier uses ChatGPT Work to reduce the number of drop-offs in its lead funnel, build campaign assets, and automate reporting.

中文介绍 Zapier企业营销团队使用ChatGPT Work减少潜在客户漏斗的流失，构建营销活动素材并自动化报告流程，提升营销效率。

Virgin Atlantic sharpens customer journeys with ChatGPT Work

OpenAI News · 08/10 08:00

Virgin Atlantic is accelerating research, product planning, and decision-making with ChatGPT Work, helping teams connect signals across the customer journey.

中文介绍 维珍大西洋航空使用ChatGPT Work加速研究、产品规划和决策，帮助团队整合客户旅程中的各类信号，优化客户体验。

Meta is back with Muse Glimmer: local, agentic, multimodal, and open source

Hugging Face Blog · 08/10 08:00

中文介绍 Hugging Face博客报道，Meta推出Muse Glimmer模型，具备本地化、智能体、多模态和开源特性，为开发者提供新的选择。

OpenAI Astra pause 🚨, Claude Code cross-session 🤖, how Cursor Router works 🔀

TLDR AI · 08/10 08:00

中文介绍 TLDR AI新闻汇总：OpenAI暂停Astra项目，Claude Code实现跨会话功能，以及Cursor Router工作原理的解析。