Proof Search

Proof search, the automated discovery of mathematical proofs, aims to leverage computational power to solve complex mathematical problems and advance our understanding of mathematical reasoning. Current research focuses on integrating deep learning models, particularly large language models and graph neural networks, with established methods like Monte Carlo Tree Search and sequent calculi to improve proof generation and selection of relevant lemmas. These advancements are enhancing automated theorem proving capabilities, impacting fields like knowledge representation and potentially leading to breakthroughs in solving currently intractable mathematical problems.

Papers