论文精选 82
Context-Aware Interleaved Batching for WhisperX· WhisperX的上下文感知交错批处理
While WhisperX accelerates speech transcription via intra-audio batching, it isolates audi…
领域：cs.CL作者：Carlos Bain、Max Bain
📰 AI 研究简报 · 9月2日
今日头条：Zeroth 元点机器人预告发布 OpenBridge 生态：具身智能的「安卓时刻」来了 —— <section data-mpa-powered-by="yiban.io" data-pm-slice="0 0 []" style="-webkit-tap-highlight-color: transparent;margin: 0px 0px 24px;padding: 0px;outline: 0px;max-width: 100%;box-sizing: border-box !important;overflow-wrap: break-word !important;color: rgba(0, 0, 0, 0.9);font-family: "PingFang SC", system-ui, -apple-system, BlinkMacSystemFont, "Helvetica Neue", "Hiragino Sans GB", "Microsoft YaHei
共 68 条预计阅读 41 分钟Token 消耗 122,400
arXiv 最新论文
论文精选 82
Configurable Semantic Chunking for Biomedical Information Extraction in Retrieval-Augmented Generation· 生物医学信息抽取中的可配置语义分块
BioMedRAG introduced retrieval-augmented generation with a learned chunk scorer for biomed…
领域：cs.CL作者：Riya Ahuja、Tim Kacprowski、Roya Shiasi Sardoabi
论文精选 65已解读
OntoAligner-Ensemble: Voting-Based Fusion across Heterogeneous Ontology Alignment Techniques· OntoAligner-Ensemble: 跨异构本体对齐技术的投票融合
Ontology alignment (OA) has evolved through several methodological paradigms, ranging from…
领域：cs.AI作者：Hamed Babaei Giglou、Sören Auer、Peio Popov
论文精选 85
DIASENTINEL: An Auditable Multi-Agent System for Guideline-Grounded Diabetes Risk Screening· DIASENTINEL：用于糖尿病风险筛查的多代理系统
Large language models (LLMs) offer promising clinical decision support but remain vulnerab…
领域：cs.CL作者：Yung Wei Shueh、Zhi-Jie Chen、Chia-Hsuan Hsu
论文精选 82
PaperGym: Rubric-Centered Evolution for Research-Plan Generation· PaperGym: 以评分标准为中心的研究计划生成进化
Research planning is the decisive capability of AI scientists. Yet a research plan admits …
领域：cs.CL作者：Yuhan Wang、Zhengxi Lu、Yuchen Yan
论文精选 65已解读
When Does Bigger Help? A Controlled Study of LLM Scale for Ontology Learning· 大模型规模何时有益？本体学习的控制研究
The effect of Large Language Model (LLM) scale on ontology learning (OL) performance remai…
领域：cs.AI作者：Hamed Babaei Giglou、Sören Auer、Jennifer D'Souza
论文精选 85
Aspire: Can Models Self-Evolve from Vague Goals?· AI模型能否从模糊目标自我进化？
Many important forms of human learning begin with a vague goal, such as "become a better p…
领域：cs.CL作者：Yuhao Wu、Jingyuan Zhang、Jiajun Shi
论文精选 65已解读
BLOOM-WILT: Logit Tilting for Behaviour Elicitation in Automated LLM Auditing· BLOOM-WILT：大模型审计中的逻辑倾斜技术
Users of a deployed language model routinely encounter behaviours that testing almost neve…
领域：cs.AI作者：Adrians Skapars、Edoardo Manino
论文精选 65已解读
Cross-Regional Grapevine Cold Hardiness Prediction via Learned Multimodal Latent Representations· 跨区域葡萄藤抗寒性预测
Accurate daily predictions of cold hardiness in woody plants are critical in regions where…
领域：cs.AI作者：William Solow、Paola Pesantez-Cabrera、Markus Keller
论文精选 65已解读
Token-Efficient Data Reasoning Agents via Adaptive Structuring of Unstructured Data· 通过自适应结构化提高数据推理效率
Valuable data remains embedded in unstructured sources: web pages, reports, contracts, fil…
领域：cs.AI作者：Milad Rezaei Hajidehi、Qitong Wang、Stratos Idreos
论文精选 65已解读
Reconciling Process Supervision with Outcome-Based Credit in Agentic Policy Optimization· 代理策略优化中过程监督与基于结果的信用的协调
Outcome-based reinforcement learning provides verified feedback for language-model agents,…
领域：cs.AI作者：Jingxiao Yang、Wangjie Gan、Yingxuan Zhuang
论文精选 65已解读
Scaling Large Reasoning Models beyond Human Supervision: A Path toward Superintelligence· 大型推理模型超越人类监督：迈向超智能的路径
Recent advances in large reasoning models (LRMs) have shown that reinforcement learning wi…
领域：cs.AI作者：Zhiqin Yang、Jingwen Fu、Yuhan Liu
论文精选 65已解读
Wrong Prediction, Right Answer: Recovering Evidence from Collapsed LLM Sequence Scores· 错误预测，正确答案：从崩溃的 LLM 序列分数中恢复证据
When a large language model fails a reasoning task, it is often assumed to lack the underl…
领域：cs.AI作者：Qiyao Yan、Chenpeng Wang、Liangming Pan
论文精选 65已解读
Measure Before You Manage: Evaluating Agent Working Memory in Coding Agents· 先测量再管理：编码代理工作内存评估
Agent working memory is heterogeneous. Objects such as instructions, artifacts, tool outpu…
领域：cs.AI作者：Le Chen、Zishen Wan、Baixi Sun
论文精选 65已解读
MNIST-PRO: MNIST is Back as a Partially Observable World for AI Agents· MNIST-PRO：AI代理的部分可观测世界
AI agents in partially observable environments need to coordinate active sensing with work…
领域：cs.AI作者：Vernon Toh、Navonil Majumder、Zhengyuan Liu
论文精选 65已解读
Learning Action Models with Conditional and Quantified Effects via Uncertainty-Guided Exploration· 通过不确定性引导探索学习条件和量化效果的动作模型
Accurate action models are critical for effective planning. Existing action-model learning…
领域：cs.AI作者：Jeffrey Jewett、William Solow、Sandhya Saisubramanian
论文精选 65已解读
CARVE: Verified Expansion for Variable-Length Generation in Diffusion Language Models· CARVE：扩散语言模型中的可验证变长生成
Masked diffusion language models predict tokens from a partially observed response canvas,…
领域：cs.AI作者：Wail Bouhedja、Amr Mohamed、Guokan Shang
论文精选 82
Responsible Integration of AI in Cancer Genomics: Barriers, Risks, and Pathways to Trustworthy Clinical Translation· AI 在癌症基因组学中的责任集成
Artificial intelligence (AI) and natural language processing (NLP) are increasingly used t…
领域：cs.AI作者：Bahar İlgen、Yiannos Tolias、Denise Kühnert
论文精选 85
CAER: Causal Action Effect Reweighting for World Model Training
World models are becoming core infrastructure for embodied intelligence, with action-condi…
领域：cs.AI作者：Jianjie Fang、Xvyuan Liu、Ziyou Wang
论文精选 75
Predicting Residential Rents in Dakar Using Machine Learning· 使用机器学习预测达喀尔住宅租金
Dakar's residential rental market remains poorly documented despite its economic and socia…
领域：cs.AI作者：Amadou Tidiane Kassa Diallo
论文精选 85
VFR-Audit: Verdict-Level Reliability for Fairness Audits in Hospital Length-of-Stay Prediction· 医院住院预测公平性审计新方法
Fairness audits in clinical Artificial Intelligence convert continuous fairness metrics in…
领域：cs.AI作者：Md Jannatul Rakib Joy、Viet Vo、Caslon Chua
论文精选 85
HSRM: Hidden-State Reward Models for Test-Time Verification· HSRM: 隐藏状态奖励模型用于测试时验证
Large language models can often generate plausible mathematical reasoning traces, but reli…
领域：cs.AI作者：Xianzhi Li、Xiaodan Zhu
论文精选 82
SkillZip Pro: Execution-Aware Dynamic Compression of Progressively Loaded Skills for Self-Evolving Agents· 执行感知的动态压缩技术
Production agent skills are directory bundles, not isolated prompts. The root is loaded at…
领域：cs.AI作者：Xiaofan Bai、Chao Liu、Hongqiang Lin
论文精选 82
Which Rules Matter Now? Policy-Centroid Routing Before an Intelligent System Acts· 智能系统行动前的政策中心路由
Before an intelligent system can decide whether an action is allowed, it must first know w…
领域：cs.AI作者：Thomson D. Nguy
论文精选 82
Autoregressive Mosaics: Probing 2D Spatial Reasoning in Text-Only Language Models· 自动回归马赛克：测试纯文本模型的空间推理能力
Large language models (LLMs) trained only on text and code can sometimes generate programs…
领域：cs.AI作者：Ashwin Nedungadi、Stefan Oehmcke、Stefan Lüdtke
GitHub 热门开源项目
开源项目精选 75
ara-mkr/Wonder-Pill· 头脑风暴互动思维导图
A Claude skill that turns brainstorming into an interactive mind map of inverted assumptio…
⭐ 56 stars
精选 AI 行业资讯
行业资讯精选 60
OpenAI faces 30 more lawsuits tied to Tumbler Ridge shooting· OpenAI 面临更多与 Tumbler Ridge 枪击案相关的诉讼
Edelson PC is filing 30 new lawsuits against OpenAI over the Tumbler Ridge shooting, escal…
行业资讯精选 82
独家解读丨字节 Seed 大调整，分支背后谁在操盘？
大模型竞争开始拼组织，这一次轮到字节。 8 月 19 日，据晚点 Late Post 报道，Seed 基础模型团队新设 Pretrain Data（预训练数据）、Horizon R…
行业资讯精选 82
我所知道的李博杰
李博杰很不喜欢“天才少年”这个标签。 “还少年呢，我都中年了，而且它指的是一个人才项目”。他调侃道。 距离轰动全网的DeepSeek面试风波已过去30天左右。趁着这段时间，李博杰在…
行业资讯精选 65
勇夺 18 块金牌：机器人运动会「金牌王」走出赛场，还能派上用场吗？
<section data-mpa-powered-by="yiban.io" data-pm-slice="0 0 []" style="-webkit-tap-highligh…
行业资讯精选 85
Zeroth 元点机器人预告发布 OpenBridge 生态：具身智能的「安卓时刻」来了· 元点机器人预告发布 OpenBridge 生态
<section data-mpa-powered-by="yiban.io" data-pm-slice="0 0 []" style="-webkit-tap-highligh…
行业资讯精选 82
SkyProduction天工工作台：从剧本到成片，一套工作台把精品短剧创作真正跑起来· 天工工作台：AI助推短剧创作
SkyProduction（天工工作台）全新版本于8月31日正式上线
行业资讯精选 82
腾讯WorkBuddy开放平台上线：首批超百家伙伴入局 国内首个同时打通软硬件、开发者三层生态的Agent基座
9月2日，腾讯WorkBuddy宣布开放平台（open.WorkBuddy.cn）正式上线，首批引入超百家生态伙伴，面向智能硬件、行业应用与开发者等开放底层Agent能力。这是国内…
行业资讯精选 82
还在为大模型洗数据熬夜？蚂蚁拿下VLDB工业最佳论文，一套宽表搞定35PB语料，效率狂飙5.6倍
蚂蚁集团推出统一宽表系统OmniTable，论文获评VLDB 2026工业赛道最佳论文
行业资讯精选 82
AIROBO联手中科院力学所，机器人运营技术注入国家科研力量
当行业将大量注意力投向人形机器人、本体制造和基础模型时，机器人进入现实世界后的“运行问题”，正在成为另一个关键技术战场。9月1日，港股上市公司弘毅文化集团（00419.HK）发布自…
行业资讯精选 82
前字节强化学习专家孙鹏博士加盟星尘智能，完善Physical AI全栈技术布局
9月2日，前字节跳动强化学习专家、前腾讯Robotics X智能体中⼼负责⼈孙鹏博⼠正式加入星尘智能
行业资讯精选 85
阿里云企业级Agent协作平台万有无界开启公测
9月2日，阿里云旗下企业级人与Agent协作平台“万有无界”开启公测。企业用户和个人用户均可通过官网(https://www.qianwenai.com/agents/wanyou…
行业资讯精选 85
建立AI时代“度量衡”：OPPO联合OpenKG推出首个端侧AI记忆评测基准MobileMem
随着终端AI从“回答问题”走向长期理解用户和主动任务执行，记忆能力正成为决定AI个性化与主动服务水平的关键基础能力。 近日，OPPO联合中文开放知识图谱（OpenKG）社区，推出行…
行业资讯精选 83
阿里更新旗舰模型Qwen3.8-Max，前端编程能力跃居全球第一
9月2日，阿里更新旗舰模型Qwen3.8-Max，性能较旧版本显著提升。在针对编程（Coding）和专业办公（Cowork）进行专项后训练后，Qwen3.8-Max新版本整体性能增…
行业资讯精选 82
企业级Agent落地样板间！百融硅基员工批量上岗，按结果领工资
AI客服日扛1.5万通电话
行业资讯精选 83
Claude最强Fable 5.1发布！8项屠榜，最高降价45%，反蒸馏机制上线· Claude最强Fable 5.1发布
无论你此前让Claude处理什么任务，Fable 5.1都能做更多，并且把最难的部分完成得更出色
行业资讯精选 65
香港首个真实开放场景服务机器人落地兰桂坊
在香港兰桂坊，一位特殊的“酒保”正式上岗
行业资讯精选 75
巨简单，更懂家！海信JUOS正式发布：行业首个家庭智能伴侣级AIOS
8月31日，海信正式发布行业首个家庭智能伴侣级AIOS——海信JUOS。
行业资讯精选 85
李飞飞发布：全球首个多模态世界模型
一张图补全3D世界，还能给机器人造训练场
行业资讯精选 82
AfterQuery reportedly becomes Y Combinator’s fastest-ever unicorn, now valued at $3.2B· AfterQuery 成为 Y Combinator 最快独角兽，估值32亿美元
AI model-training startup AfterQuery has reportedly raised a round that valued it at $3.2 …
行业资讯精选 83
NVIDIA and CrowdStrike Strengthen Agentic Cybersecurity Frontier· NVIDIA 与 CrowdStrike 加强智能网络安全前沿
“We’re at an inflection point in cybersecurity,” Jensen Huang told a sold-out crowd at Cro…
行业资讯精选 85
Open AI’s Astra model is on the way—and very good at breaking into computer systems· OpenAI 新模型 Astra 即将发布，擅入侵系统
OpenAI previewed the precautions it is taking as it prepares to release Astra, its newest,…
行业资讯精选 80
The latest AI news we announced in August 2026· 2026 年 8 月 AI 新闻
Here are Google’s latest AI updates from August 2026
行业资讯精选 80
Anthropic’s new Fable release is cheaper, less restrictive· Anthropic 的新 Fable 发布更便宜，限制更少
Fable 5.1 includes changes meant to reduce token cost and false-positive restrictions from…
行业资讯精选 82
Google’s answer to Canva is an AI tool where you prompt instead of design· 谷歌的回答是Canva的AI工具
With Google Pics, Google is pushing deeper into the creative software market dominated by …
行业资讯精选 75
ChatGPT Health adds Epic integration for clinicians to import patient data· ChatGPT Health添加Epic集成
OpenAI said that the integration provides read-only access to health records for clinician…
行业资讯精选 82
How AI-native companies turn workflows into operating capability· AI 原生公司的工作流转型
Basis, Clay, and Exa Labs use AI agents to improve onboarding, account management, and dev…
行业资讯精选 75
Sequoia-incubated Empirik launches with $21M to predict outages before they happen· 红杉孵化的 Empirik 融资 2100 万美元用于预测故障
The startup wants to do for IT infrastructure what Cursor did for software engineering.
行业资讯精选 75
Try Google Pics: Easy image creation and editing in Google Workspace· 尝试 Google Pics：谷歌工作区中的图像创作与编辑
Built on our latest Nano Banana model, Google Pics — our image creation and editing tool —…
行业资讯精选 68
Amazon Alexa can now alert you when something new might tempt you to shop· 亚马逊 Alexa 可提醒你新购物诱因
Amazon is adding a new Alexa-powered feature called “Update Me When” that can send persona…
行业资讯精选 75
AIR raises $50M to help companies vet the skills and add-ons AI agents use· AIR 获 5000 万美元融资，帮助企业审核 AI 代理技能
AIR's platform can discover agents running at a company, continuously vets any skills and …
行业资讯精选 82
Fambot introduces an ‘AI chief of staff’ for families· Fambot 引入家庭‘AI 首席幕僚’
Fambot is building an AI “chief of staff” to help families manage the emails, calendars, s…
行业资讯精选 82
Path to Astra: critical capabilities and frontier safeguards· Astra 之路：关键能力和前沿保障
Astra is the first OpenAI model to meet the Critical cybersecurity capability threshold un…
行业资讯精选 82
Healthcare organizations can now connect EHR and additional industry data to ChatGPT· ChatGPT 连接医疗数据
ChatGPT can now connect to trusted healthcare data, helping clinicians securely access pat…
行业资讯精选 85
追觅旗下户外庭院板块将亮相IFA 2026，AI与机器人重构户外多场景应用
全球消费电子产业的“风向标”——德国柏林国际消费电子展（IFA 2026）即将于9月4日启幕。追觅旗下户外庭院板块将亮相IFA 2026，携全新产品阵容与全球化战略布局登上国际舞台…
行业资讯精选 80
How law firm Gilbert + Tobin governs and scales AI with OpenAI· 吉尔伯特+托宾律师事务所如何治理和扩展AI
See how Gilbert + Tobin combines CEO-led commitment, rigorous governance, and human accoun…
YouTube 热门 AI 视频
AI 视频精选 75
Local AI is just getting started, and there’s room for you to shape it.· 本地AI：方兴未艾
YouTube · DeepLearningAI
📺 DeepLearningAI▶ 118 次观看
AI 视频精选 83
Agentic approaches to processing long videos with Gemini· Gemini 处理长视频的代理方法
YouTube · Google for Developers
📺 Google for Developers▶ 80 次观看
AI 视频精选 65
Tokenomics 101: What Are Tokens & Why They Matter | AI Factory Insider Ep. 4· 代币经济学101：代币为何重要
YouTube · NVIDIA
📺 NVIDIA▶ 965 次观看
AI 视频精选 82
NVIDIA and CrowdStrike are advancing agentic cyber defense with SafeMind· 英伟达和 CrowdStrike 推进网络安全智能防御
YouTube · NVIDIA
📺 NVIDIA▶ 359 次观看
AI 视频精选 65
AI Dev | San Francisco 2026 | 2,599 AI Developers· 旧金山 2026 年 AI 开发者大会
YouTube · DeepLearningAI
📺 DeepLearningAI▶ 278 次观看
AI 视频精选 82
Agentic video understanding in Gemini· Gemini中的代理视频理解
YouTube · Google for Developers
📺 Google for Developers▶ 51 次观看
AI 视频精选 85
Koray Kavukcuoglu on frontier models, coding agents, and building AGI· Koray Kavukcuoglu 谈前沿 AI 模型
YouTube · Google for Developers
📺 Google for Developers▶ 941 次观看