Verdict Inference with Claim and Retrieved Elements Using RoBERTa
In-Zu Gi, Ting-Yu Fang, Richard Tzong‐Han Tsai
Abstract
Automatic fact verification has attracted recent research attention as the increasing dissemination of disinformation on social media platforms. The FEVEROUS shared task introduces a benchmark for fact verification, in which a system is challenged to verify the given claim using the extracted evidential elements from Wikipedia documents. In this paper, we propose our 3 rd place three-stage system consisting of document retrieval, element retrieval, and verdict inference for the FEVER-OUS shared task. By considering the context relevance in the fact extraction and verification task, our system achieves 0.29 FEVER-OUS score on the development set and 0.25 FEVEROUS score on the blind test set, both outperforming the FEVEROUS baseline.