Brisbane–Udine–Melbourne (BUM)@CLEF2022 CheckThat! Lab: A Composite Deep Learning Approach to Fake News Detection using Evidence Retrieval

Abstract

We detail a deep learning approach based on the transformer architecture for performing fake news detection. The proposed approach is composed of a deep learning network which receives as input the claim to be verified, a series of predictions made by other models, and supporting evidence in the form of ranked passages. We validate our approach on the data from the CLEF2022-CheckThat! Lab (Task 3: Fake News Detection), where we achieve an F1-score of $0.275$, ranking $10^textth$ out of $25$ participants.

Publication
Working Notes of CLEF 2022—Conference and Labs of the Evaluation Forum