Generating Natural Language Proofs with Verifier-Guided Search
share
Deductive reasoning (drawing conclusions from assumptions) is a challenging problem in NLP. In this work, we focus on proof generation: given a hypothesis and a set of supporting facts in natural language, the model generates a proof tree indicating how to deduce the hypothesis from supporting facts. Instead of generating the entire proof in one shot, prior work has demonstrated the promise of stepwise generation but achieved limited success on real-world data. Existing stepwise methods struggle to generate proof steps that are both valid and relevant. In this paper, we present a novel stepwise method NLProofS (Natural Language Proof Search), which learns to generate relevant steps conditioning on the hypothesis. At the core of our approach, we train an independent verifier to check the validity of proof steps. Instead of generating steps greedily, we search for proofs maximizing a global proof score judged by the verifier. NLProofS achieves state-of-the-art performance on EntailmentBank and RuleTaker. For example, it improves the percentage of correctly predicted proofs from 20.9 EntailmentBank. This is the first time stepwise methods have led to better generation of challenging human-authored proofs.
READ FULL TEXT