English

Unsupervised Pretraining for Fact Verification by Language Model Distillation

Computation and Language 2024-03-08 v3 Machine Learning Machine Learning

Abstract

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.

Keywords

Cite

@article{arxiv.2309.16540,
  title  = {Unsupervised Pretraining for Fact Verification by Language Model Distillation},
  author = {Adrián Bazaga and Pietro Liò and Gos Micklem},
  journal= {arXiv preprint arXiv:2309.16540},
  year   = {2024}
}

Comments

ICLR 2024 Camera Ready

R2 v1 2026-06-28T12:35:05.065Z