AI discovers formally verified quantum programs
Original title: Discovering formally verified quantum programs with AI
LLM agents synthesize novel quantum programs in Lean and verify them, improving baselines by up to 290x.
Large language model agents build a quantum program library in Lean against formal specifications. They collaborate via wikis and conferences to share results. The method yields order-of-magnitude improvements on quantum arithmetic and scientific simulations.
- Agents use Lean theorem prover for synthesis and verification
- Team collaboration via wikis and conferences
- Improves quantum arithmetic by 290x
- Fusion simulations improved 11x
- Shortens path to practical quantum computation
From the original article
1 min read at arxiv.orgThe timeline by which quantum computers can solve breakthrough problems in science depends on improved quantum algorithms as much as on better hardware. Such algorithmic advances are typically achieved through many years of expert human collaboration. In mathematics and computer science, large language models have accelerated progress through automated discovery loops.
By Sergey Shirobokov · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.