Blog — Latest in Math Tech & AI
News, studies, deep-dives, and trends at the intersection of mathematics and AI.
AI-assisted Lean project maps a formal route through the Poincaré conjecture
An October 6 preprint describes how mathematicians and agents organized a large Lean formalization around reviewed milestones.
Study finds LLMs can silently change a theorem’s scope when translating mathematics
An October 2 preprint tests whether AI preserves the assumptions hidden inside mathematical language.
Cogentic reports AI-assisted results on open problems in learning and auctions
A new preprint describes a multi-agent proof workflow whose reported results were checked by domain experts.
Study separates AI mathematical answers from discovery of the key idea
An October 1 preprint tests whether AI can identify the structural idea that makes a mathematics solution work.
AI-designed proof interfaces show gains in Rocq, tradeoffs in Lean
A September 30 preprint tests whether better tools can make AI mathematical proofs more efficient.
GenLimitLib tests reusable Lean libraries for AI mathematics
A September 29 preprint explores how organized formal mathematics can help AI develop checked proofs.
Wolfram outlines language plans for AI-assisted pure mathematics
A September 28 essay describes work toward mathematical representations that humans and AI can inspect.
MathAgent study finds mixed benefits from knowledge graphs in proof search
A September 28 preprint tests when extra mathematical context helps AI theorem provers.
ProofGap tests AI on individual steps in mathematical analysis
A September 24 preprint offers a closer look at where formal reasoning succeeds or fails.
Showing page 3 of 10 · 86 total