Proof Artifact Co-training for Theorem Proving with Language Models
arXiv:2102.06203
Abstract
Labeled data for imitation learning of theorem proving in large libraries of formalized mathematics is scarce as such libraries require years of concentrated effort by human specialists to be built. This is particularly challenging when applying large Transformer language models to tactic prediction, because the scaling of performance with respect to model size is quickly disrupted in the data-scarce, easily-overfitted regime. We propose PACT ({\bf P}roof {\bf A}rtifact {\bf C}o-{\bf T}raining), a general methodology for extracting abundant self-supervised data from kernel-level proof terms for co-training alongside the usual tactic prediction objective. We apply this methodology to Lean, an interactive proof assistant which hosts some of the most sophisticated formalized mathematics to date. We instrument Lean with a neural theorem prover driven by a Transformer language model and show that PACT improves theorem proving success rate on a held-out suite of test theorems from 32\% to 48\%.
References in corpus (10)
- Learning Transferable Visual Models From Natural Language Supervision
- Language Models are Few-Shot Learners
- Cross-lingual Language Model Pretraining
- MASS: Masked Sequence to Sequence Pre-training for Language Generation
- Scaling Laws for Autoregressive Generative Modeling
- Generative Language Modeling for Automated Theorem Proving
- HolStep: A Machine Learning Dataset for Higher-order Logic Theorem Proving
- Holophrasm: a neural Automated Theorem Prover for higher-order logic
- Hammering Mizar by Learning Clause Guidance
- Universal Policies for Software-Defined MDPs
Cited by in corpus (5)
- On the Opportunities and Risks of Foundation Models
- MiniF2F: a cross-system benchmark for formal Olympiad-level mathematics
- TacticZero: Learning to Prove Theorems from Scratch with Deep Reinforcement Learning
- Training a First-Order Theorem Prover from Synthetic Data
- BAIT: Benchmarking (Embedding) Architectures for Interactive Theorem-Proving