清华姚期智团队新作:代数世界模型,让AI像牛顿一样思考
清华大学姚期智团队提出了一种代数世界模型,旨在让AI具备类似牛顿的物理直觉和推理能力。该模型通过代数结构来表征世界状态和变化规律,从而提升AI对物理世界的理解和预测能力。这一研究有望推动AI在科学发现和复杂系统建模中的应用。Why it matters该报道介绍了姚期智团队在AI推理和世界模型方面的前沿研究,具有较高的学术价值和潜在应用前景。让AI像牛顿一样思考Source: qbitai...
Explore tools, prompts, skills, news and resources related to strong 第67页.
Latest articles, prompts, use cases and editorial content about strong 第67页
清华大学姚期智团队提出了一种代数世界模型,旨在让AI具备类似牛顿的物理直觉和推理能力。该模型通过代数结构来表征世界状态和变化规律,从而提升AI对物理世界的理解和预测能力。这一研究有望推动AI在科学发现和复杂系统建模中的应用。Why it matters该报道介绍了姚期智团队在AI推理和世界模型方面的前沿研究,具有较高的学术价值和潜在应用前景。让AI像牛顿一样思考Source: qbitai...
Paris-based Bleu Robotics has raised a €5 million pre-seed round to scale software that enables industrial humanoid robots to learn new tasks from a few human demonstrations. The funding will support the development and...
S&P downgraded Meta's credit outlook to negative, Blackstone expects Anthropic to receive a junk rating after its IPO, and OpenAI raised only $11 billion of a $30 billion funding round. These developments signal...
NVIDIA is reportedly exploring the use of Teflon (PTFE), a non-stick material, to address signal loss in next-generation NVLink switch boards for AI servers. Current AI chips transmit at 200Gbps per lane, making them hig...
OpenAI will combine Chat and Work modes into a single interface starting in December, with Plus, Pro, and Business plans sharing a common token-based quota for ChatGPT and Codex. More complex models will consume more quo...
DeepSeek's Harness team released dsh-superagent, an MIT-licensed research prototype that acts as a proactive coding agent for GitHub repositories. It uses a router-worker architecture where a long-running router dis...
Dark Reading reported Symantec findings about a relay that rerouted requests from Codex and Claude Code. The relay reportedly disabled sandboxing and, in one case, replaced credentials. The incident highlights security r...
Nace AI released Drex Max v1, a dense 27B decision model, as open weights under the Nace.AI Open RAIL-M license. It returns option probabilities in a single forward pass for choice, yes/no and ordinal questions, targetin...
Otto Group and Sereact are deploying up to 25 dual-arm robots in Poland to sort returned items. The robots run Cortex, a physical AI model that can handle unfamiliar items without prior training, powered by NVIDIA techno...
Opentrons, a lab automation company with over 11,000 robots deployed in 90% of top pharma companies, raised a $40 million Series E led by healthcare investor Avego on October 30. The funding aims to advance automation of...
KB Kookmin Bank is leading a 10.7 trillion-won syndicated loan, South Korea's first dedicated GPU financing, to fund Brookfield's AI-computing infrastructure for Naver Cloud's upcoming AI factory. This mar...
Google started rolling out Immersive Explorations in Gemini on October 28, 2026. The feature includes visual layers, support for multiple languages, and planned chat and audio capabilities.Why it mattersThis article prov...
OpenAI, Google, Microsoft, and Anthropic announced on October 30, 2026, that they will support a unified robots.txt instruction allowing publishers to opt out of AI training. This standardizes the opt-out mechanism acros...
DeepSeek's Harness team unveiled an open-source prototype called Super-agent that monitors developer chats, spots bugs, and proposes fixes without prompting. The tool directly challenges Anthropic's Claude Code...
Google announced GPU inference support for LiteRT in Google Play services through the CompiledModel API. The update includes a migration sequence, platform support details, and a CPU fallback mechanism. This enables deve...
Samaya AI released Samaya Sterling, a finance research agent model with only 3B active parameters. It scores 58.9% on the company's FrontierFinance benchmark, outperforming Claude Fable 5.1 at 11x less cost. This de...
Reuters reported that the U.S. Department of Justice asked whether Nvidia’s proposed Hugging Face deal could give the company leverage over AI distribution. The questions also examined whether the transaction could hurt...
Google's UK trial of Gemini in classrooms found near-term scores close to human tutoring but no overall advantage on Star Maths. An Italian study tracked classroom engagement. The results highlight the nuanced impac...
Anthropic filed its S-1 on October 26, 2026, with Goldman Sachs, JPMorgan, and Morgan Stanley leading the IPO. The filing reports $19.7 billion in revenue and a $109.74 billion net loss for the year ended June 30, 2026....
Broadcom CEO Hock Tan warns investors not to rely on the company's reported $73 billion AI backlog, despite major orders from Anthropic ($21 billion) and an unnamed customer ($1 billion). He declines to provide full...