Reasoning Over Semantic-Level Graph for Fact Checking

TL;DR

This paper introduces a novel fact-checking method that leverages semantic-level graph structures and advanced neural networks to improve reasoning over evidence, achieving state-of-the-art accuracy on the FEVER dataset.

Contribution

It proposes a new approach using semantic role labeling and graph neural networks to enhance evidence reasoning in fact checking, surpassing previous methods.

Findings

Graph-based mechanisms improve fact checking accuracy.

Semantic structure utilization enhances evidence reasoning.

Achieves state-of-the-art results on FEVER dataset.

Abstract

Fact checking is a challenging task because verifying the truthfulness of a claim requires reasoning about multiple retrievable evidence. In this work, we present a method suitable for reasoning about the semantic-level structure of evidence. Unlike most previous works, which typically represent evidence sentences with either string concatenation or fusing the features of isolated evidence sentences, our approach operates on rich semantic structures of evidence obtained by semantic role labeling. We propose two mechanisms to exploit the structure of evidence while leveraging the advances of pre-trained models like BERT, GPT or XLNet. Specifically, using XLNet as the backbone, we first utilize the graph structure to re-define the relative distances of words, with the intuition that semantically related words should have short distances. Then, we adopt graph convolutional network and graph attention network to propagate and aggregate information from neighboring nodes on the graph. We evaluate our system on FEVER, a benchmark dataset for fact checking, and find that rich structural information is helpful and both our graph-based mechanisms improve the accuracy. Our model is the state-of-the-art system in terms of both official evaluation metrics, namely claim verification accuracy and FEVER score.