Neural multi-task learning for stance prediction

Abstract

We present a multi-task learning model that leverages large amount of textual information from existing datasets to improve stance prediction. In particular, we utilize multiple NLP tasks under both unsupervised and supervised settings for the stance prediction task. Our model obtains state-of-the-art performance on a public benchmark dataset, Fake News Challenge, outperforming current approaches by a wide margin.

Publication
Proceedings of the Second Workshop on Fact Extraction and VERification (FEVER)