Logic-Guided Data Augmentation and Regularization for Consistent Question Answering

TL;DR

This paper introduces a logic-guided data augmentation and regularization method to enhance the accuracy and consistency of question answering systems, especially for comparison questions involving logical and linguistic reasoning.

Contribution

It presents a novel approach that combines logic rules with neural models to improve data quality and model consistency in QA tasks, advancing state-of-the-art results.

Findings

Significant performance improvements on WIQA and QuaRel datasets.

58% reduction in consistency violations on HotpotQA.

Effective learning from limited training data.

Abstract

Many natural language questions require qualitative, quantitative or logical comparisons between two entities or events. This paper addresses the problem of improving the accuracy and consistency of responses to comparison questions by integrating logic rules and neural models. Our method leverages logical and linguistic knowledge to augment labeled training data and then uses a consistency-based regularizer to train the model. Improving the global consistency of predictions, our approach achieves large improvements over previous methods in a variety of question answering (QA) tasks including multiple-choice qualitative reasoning, cause-effect reasoning, and extractive machine reading comprehension. In particular, our method significantly improves the performance of RoBERTa-based models by 1-5% across datasets. We advance the state of the art by around 5-8% on WIQA and QuaRel and reduce consistency violations by 58% on HotpotQA. We further demonstrate that our approach can learn effectively from limited data.