2026-09-15 科技日报

扫描 552 篇候选内容 · 覆盖 650 个信息源 · 纳入 541 篇

📌 今日要点

• MoPA: Coordinated Mobile Manipulation via Subsystem-Specific Perception Alignment
• Uncertainty-Aware Conflict Detection Against Operator-Conditioned Weather Hazards
• RodForesight: A World Model Enhanced Diffusion Policy for Slender and Material Agnostic Rod Insertion
• Feature Recovery for Object Understanding After Irreversible Fire Damage
• Does Video Memory Use What It Retrieves? A Causal Audit of Memory Specificity
• Beyond Argmax: A Mechanistic Study of Semantic Retention in Frozen Foundation-Model Composition for Generalized Few-Shot 3D Segmentation

📝 科技简讯

• MoPA: Coordinated Mobile Manipulation via Subsystem-Specific Perception Alignment — ArXiv RO (cs.RO)
• Uncertainty-Aware Conflict Detection Against Operator-Conditioned Weather Hazards — ArXiv RO (cs.RO)
• RodForesight: A World Model Enhanced Diffusion Policy for Slender and Material Agnostic Rod Insertion — ArXiv RO (cs.RO)
• Feature Recovery for Object Understanding After Irreversible Fire Damage — ArXiv CV (cs.CV)
• Does Video Memory Use What It Retrieves? A Causal Audit of Memory Specificity — ArXiv CV (cs.CV)
• Beyond Argmax: A Mechanistic Study of Semantic Retention in Frozen Foundation-Model Composition for Generalized Few-Shot 3D Segmentation — ArXiv CV (cs.CV)
• HSI-Road Relabeled: Surface-Aware Road-Scene Segmentation — ArXiv CV (cs.CV)
• Single-Query Person-Centric Bimanual Hand-Object Interaction Detection — ArXiv CV (cs.CV)
• OpenAI has hundreds of contract workers reading your ChatGPT conversations — The Decoder
• Microsoft’s AI rulebook: readable thinking, no inner life, and definitely no rights — The Decoder
• Anthropic eyes Nasdaq listing as a second profitable quarter aims to win over investors ahead of a mega-IPO — The Decoder
• Clay Mathematics Institute says the Navier-Stokes Millennium Prize Problem has “apparently been settled” — The Decoder
• China fires back at U.S. AI safety warnings, calling them fearmongering to lock in American advantage — The Decoder
• Sam Altman calls for pacing AI development but promises rapid progress will continue — The Decoder
• Introducing The Campaign Agent That Turns Goals Into Growth — Salesforce AI Blog
• Introducing Security Mesh: Unify Your Security Data for Greater Visibility — Salesforce AI Blog
• Meet Your Digital Teammate, Slackbot: Spotlight on Productivity Agent — Salesforce AI Blog
• Meet Your Digital Teammate, Casey: Spotlight on Service — Salesforce AI Blog
• Meet Your Digital Teammate, Hunter: Spotlight on Sales Agent — Salesforce AI Blog
• 首届蚂蚁灵波具身大模型挑战赛正式启动 — 量子位
• 全国唯一！商汤大装置临港 AIDC 获“算效 + 算电”双 5A 认证 — 量子位
• 探索 RSI，生数新世界模型让机器人开始自我进化 — 量子位
• 分子之心 QuantaMind 登 Science Advances，让 AI 给分子世界”拍电影” — 量子位
• 端侧 AI 从「能跑」到「会进化」，元空智能跑进惠普预装 — 量子位
• 一年连融三轮，这家金融 AI 公司又拿下超 3 亿 B 轮 — 量子位
• Fundamental Dynamical Units for Physics-Informed Structural Inference from Perturbation Time-Series in Networked Systems — ArXiv ML (cs.LG)
• Physics-Informed Conformal Prediction: Embedding PDE Consistency into Distribution-Free Uncertainty Quantification for Neural Operators — ArXiv ML (cs.LG)
• Fed-Equilibrium Framework for Topological Pareto Control in Robust and Fair Clinical Federated Learning — ArXiv ML (cs.LG)
• Efficient AI Model Deployment Using Quantization Analysis Tool — ArXiv ML (cs.LG)
• Performance, Efficiency and Collapse — Advantages and Challenges in Offline Post-training of Code LLMs — ArXiv ML (cs.LG)
• Look Before You Leap: Pre-Action Verification for LLM Agents — ArXiv ML (cs.LG)
• Decoding Mixture Perception through Computational Modeling of Component Interactions — ArXiv ML (cs.LG)
• Model Weight Exfiltration Seems Overrated — LessWrong
• Improving Psychiatric Medicine Development with AI — LessWrong
• Alignment Problem Redux — LessWrong
• Self Inoculation — LessWrong
• University of Cambridge – College EA Meetups Everywhere Fall 2026 — LessWrong
• 7 Python Best Practices Senior Developers Follow (That Beginners Often Miss) — KDnuggets
• Why DeepSeek-V4.1-Flash Is Such an Exciting Open Model Release — KDnuggets
• Watch astronaut Christina Koch and Google’s James Manyika discuss space, technology, and discovery. — Google AI Blog

📊 数据概览