REFACTOR: Learning to Extract Theorems from Proofs
This addresses the challenge of automating theorem discovery in mathematical reasoning, offering incremental improvements for formal theorem proving systems.
The paper tackles the problem of automatically extracting reusable theorems from proofs in formal mathematics, achieving a 19.6% extraction rate on unseen proofs and generating 16 new theorems that shorten proofs and improve prover performance.
Human mathematicians are often good at recognizing modular and reusable theorems that make complex mathematical results within reach. In this paper, we propose a novel method called theoREm-from-prooF extrACTOR (REFACTOR) for training neural networks to mimic this ability in formal mathematical theorem proving. We show on a set of unseen proofs, REFACTOR is able to extract 19.6% of the theorems that humans would use to write the proofs. When applying the model to the existing Metamath library, REFACTOR extracted 16 new theorems. With newly extracted theorems, we show that the existing proofs in the MetaMath database can be refactored. The new theorems are used very frequently after refactoring, with an average usage of 733.5 times, and help shorten the proof lengths. Lastly, we demonstrate that the prover trained on the new-theorem refactored dataset proves more test theorems and outperforms state-of-the-art baselines by frequently leveraging a diverse set of newly extracted theorems. Code can be found at https://github.com/jinpz/refactor.