T
traeai
Sign in

产品

Lean

别名:Lean theorem prover

形式化验证与数学证明的交互式定理证明器,Axiom用于将非正式证明转化为可验证形式。

已跟踪 8 条高相关材料

TraeAI 观察

相关材料

已收录 8 条与 Lean 相关的内容,按评分排序。

Latent Space 图标

Scaling Past Informal AI - Carina Hong, Axiom Math

Latent Space1535 字 (约 7 分钟)
87

Axiom Math advances Verified AI to scale brilliance and compound it through formal proofs with Lean, achieving 12/12 on Putnam and 99% (187/189) on Verina Codegen, far exceeding OpenAI o3’s 4.9%, providing critical capability verification and knowledge propagation for AGI.

入选理由:Axiom在Putnam考试中取得12/12,优于顶尖本科生与当时最接近的AI系统DeepSeek(103/120)。

FeaturedArticle#Verified AI#Formal Verification#Lean#AGI#Putnam Exam英文
5篇AI生成的数学论文被接收!00后创始人洪乐潼融资14个亿

Axiom Math’s AI system AxiomProver generated and formally verified 8 math papers, 5 of which have passed peer review; its core is a ‘natural language → Lean formalization → machine verification’ loop. Founder Hong Letong (born 2001) led a $1.4B RMB funding round, with valuation reaching $1.6B USD.

入选理由:AxiomProver在24小时内可生成完整、机器验证的数学证明,已解决6个Ballantine等提出的猜想并发现1个反例

FeaturedArticle#AI Math#Formal Proof#Lean#Axiom Math#Hong Letong中文
mimalloc: A new, high-performance, scalable memory allocator for the modern era

mimalloc: A new, high-performance, scalable memory allocator for the modern era

Microsoft Research Blog2098 字 (约 9 分钟)
85

mimalloc 是一个高效、可扩展的内存分配器,适用于现代高并发、大内存规模的应用场景,显著提升了 Bing 等服务的响应时间。

入选理由:mimalloc 提供了有界最坏情况分配时间,空间开销最小,内部碎片率低。

FeaturedArticle#内存分配器#NoGIL CPython#Unreal Engine中英混合
What mathematicians figured out about AI that most enterprises haven’t

What mathematicians figured out about AI that most enterprises haven’t

Gradient Flow1272 字 (约 6 分钟)
85

数学家发现AI在验证、迭代流程和人类监督下的优势,企业应借鉴这些条件来加速AI进展。

入选理由:AI在数学中的应用加速了验证和迭代流程。

FeaturedArticle#AI#数学#验证#迭代英文
Test-time verification for AI agents: New from Microsoft Research #ai #agenticai #verification

Test-time Verification for AI Agents: New from Microsoft Research

Microsoft Research200 字 (约 1 分钟)
72

Microsoft Research introduces Intervene, a real-time framework that uses LLM-driven projection to decompose agent outputs into verifiable properties and automatically generates formal specs and verifiers (Python/Lean) for runtime intervention.

入选理由:Intervene 是微软研究院提出的实时 AI agent 验证框架,支持对部分响应进行即时验证。

FeaturedVideo#AI Agent#Formal Verification#Microsoft Research#Intervene#Agentic AI英文
Introduction to Lean for Programmers

Introduction to Lean for Programmers

Towards Data Science2997 字 (约 12 分钟)
52

The article claims to introduce Lean to programmers but actually focuses entirely on interactive theorem proving with Lean and AI, misrepresenting its topic with low practical value and incomplete content.

入选理由:Lean 4 配合 GitHub Copilot 可用于交互式定理证明,Terence Tao 已在 YouTube 上直播使用。

FeaturedArticle#Lean#proof assistant#AI theorem proving#AlphaProof#Curry-Howard英文
Gary Marcus(@GaryMarcus) 图标

Is the new math result neurosymbolic with Lean, harnesses etc or a pure LLM?

Gary Marcus(@GaryMarcus)56 字 (约 1 分钟)
30

Gary Marcus questions whether the new math achievement uses neurosymbolic approaches with Lean tools or relies purely on LLMs, but provides no technical details.

入选理由:该推文仅提出技术路线疑问,未展示任何实证分析或架构细节

FeaturedTweet#Neurosymbolic AI#LLM#Lean Theorem Prover英文

跨材料问答 · Lean

回答基于:Lean 相关 8 条材料
    0 / 500

    AI may generate inaccurate information. Please verify important content.