Fine-grained Fact Verification with Kernel Graph Attention Network

TL;DR

This paper introduces Kernel Graph Attention Network (KGAT), a novel model for fine-grained fact verification that uses kernel-based attentions to improve evidence importance measurement and evidence propagation, achieving state-of-the-art results.

Contribution

The paper proposes KGAT, which incorporates node and edge kernels into graph attention networks for more precise fact verification, outperforming existing models on FEVER.

Findings

KGAT achieves a 70.38% FEVER score.

Kernel-based attention focuses more on relevant evidence.

KGAT significantly outperforms existing models on FEVER.

Abstract

Fact Verification requires fine-grained natural language inference capability that finds subtle clues to identify the syntactical and semantically correct but not well-supported claims. This paper presents Kernel Graph Attention Network (KGAT), which conducts more fine-grained fact verification with kernel-based attentions. Given a claim and a set of potential evidence sentences that form an evidence graph, KGAT introduces node kernels, which better measure the importance of the evidence node, and edge kernels, which conduct fine-grained evidence propagation in the graph, into Graph Attention Networks for more accurate fact verification. KGAT achieves a 70.38% FEVER score and significantly outperforms existing fact verification models on FEVER, a large-scale benchmark for fact verification. Our analyses illustrate that, compared to dot-product attentions, the kernel-based attention concentrates more on relevant evidence sentences and meaningful clues in the evidence graph, which is the main source of KGAT's effectiveness.