Unsupervised Pretraining for Fact Verification by Language Model Distillation
This addresses the challenge of verifying claims using knowledge bases without annotated data, offering a novel approach for fact-checking applications.
The paper tackles the problem of fact verification by proposing an unsupervised pretraining framework that distills self-supervised features from language models to align claims with evidence, achieving state-of-the-art results with a +5.3% Hits@1 improvement on FB15k-237 and +8% accuracy on FEVER.
Fact verification aims to verify a claim using evidence from a trustworthy knowledge base. To address this challenge, algorithms must produce features for every claim that are both semantically meaningful, and compact enough to find a semantic alignment with the source information. In contrast to previous work, which tackled the alignment problem by learning over annotated corpora of claims and their corresponding labels, we propose SFAVEL (Self-supervised Fact Verification via Language Model Distillation), a novel unsupervised pretraining framework that leverages pre-trained language models to distil self-supervised features into high-quality claim-fact alignments without the need for annotations. This is enabled by a novel contrastive loss function that encourages features to attain high-quality claim and evidence alignments whilst preserving the semantic relationships across the corpora. Notably, we present results that achieve a new state-of-the-art on FB15k-237 (+5.3% Hits@1) and FEVER (+8% accuracy) with linear evaluation.