PRover: Proof Generation for Interpretable Reasoning over Rules
Preprint
Abstract
Recent work by Clark et al. (2020) shows that transformers can act as 'soft theorem provers' by answering questions over explicitly provided knowledge in natural language. In our work, we take a step closer to emulating formal theorem provers, by proposing PROVER, an interpretable transformer-based model that jointly answers binary questions over rule-bases and generates the corresponding proofs. Our model learns to predict nodes and edges corresponding to proof graphs in an efficient constrained training paradigm. During inference, a valid proof, satisfying a set of global constraints is generated. We conduct experiments on synthetic, hand-authored, and human-paraphrased rule-bases to show promising results for QA and proof generation, with strong generalization performance. First, PROVER generates proofs with an accuracy of 87%, while retaining or improving performance on the QA task, compared to RuleTakers (up to 6% improvement on zero-shot evaluation). Second, when trained on questions requiring lower depths of reasoning, it generalizes significantly better to higher depths (up to 15% improvement). Third, PROVER obtains near perfect QA accuracy of 98% using only 40% of the training data. However, generating proofs for questions requiring higher depths of reasoning becomes challenging, and the accuracy drops to 65% for 'depth 5', indicating significant scope for future work. Our code and models are publicly available at https://github.com/swarnaHub/PRoverKeywords
All Related Versions
This publication has 14 references indexed in Scilit:
- oLMpics-On What Language Model Pre-training CapturesTransactions of the Association for Computational Linguistics, 2020
- Stop explaining black box machine learning models for high stakes decisions and use interpretable models insteadNature Machine Intelligence, 2019
- QuaRTz: An Open-Domain Dataset of Qualitative Relationship Questions2019
- Reasoning Over Paragraph Effects in Situations2019
- Learning Deep Transformer Models for Machine Translation2019
- "Why Should I Trust You?"2016
- Natural Logic and Natural Language Inference2014
- Multicommodity max-flow min-cut theorems and their use in designing approximation algorithmsJournal of the ACM, 1999
- Network Flow and Testing Graph ConnectivitySIAM Journal on Computing, 1975
- The logic theory machine--A complex information processing systemIEEE Transactions on Information Theory, 1956