Timezone: »

 
Draft, Sketch, and Prove: Guiding Formal Theorem Provers with Informal Proofs
Albert Jiang · Sean Welleck · Jin Peng Zhou · Timothee Lacroix · Jiacheng Liu · Wenda Li · Mateja Jamnik · Guillaume Lample · Yuhuai Wu

@
in MATH-AI: Toward Human-Level Mathematical Reasoning »

The formalization of existing mathematical proofs is a notoriously difficult process. Despite decades of research on automation and proof assistants, writing formal proofs remains arduous and only accessible to a few experts. While previous studies to automate formalization focused on powerful search algorithms, no attempts were made to take advantage of available informal proofs. In this work, we introduce Draft, Sketch, and Prove (DSP), a method that maps informal proofs to formal proof sketches, and uses the sketches to guide an automated prover by directing its search to easier sub-problems. We investigate two relevant setups where informal proofs are either written by humans or generated by a language model. Our experiments and ablation studies show that large language models are able to produce well-structured formal sketches that follow the same reasoning steps as the informal proofs. Guiding an automated prover with these sketches enhances its performance from 20.9% to 39.3% on a collection of mathematical competition problems.

Author Information

Albert Jiang (University of Cambridge)
Sean Welleck (University of Washington)
Jin Peng Zhou (Department of Computer Science, University of Toronto)
Timothee Lacroix (FAIR)
Jiacheng Liu (Department of Computer Science, University of Washington)
Wenda Li (University of Cambridge)
Mateja Jamnik (University of Cambridge)
Guillaume Lample (Facebook AI Research)
Yuhuai Wu (Stanford University / Google)

More from the Same Authors