UofA-Truth at Factify 2022 : Transformer And Transfer Learning Based Multi-Modal Fact-Checking

TL;DR

This paper presents a multi-modal fact-checking approach using transformer and transfer learning techniques, achieving competitive results in the De-Factify@AAAI2022 shared task for fake news detection.

Contribution

The paper introduces a novel multi-modal fact-checking method leveraging transformer models and transfer learning, specifically designed for text and image misinformation detection.

Findings

Achieved an F1-weighted score of 74.807% in the shared task.

Ranked fourth among all submissions in the competition.

Demonstrated effectiveness of transformer-based transfer learning for multi-modal fake news detection.

Abstract

Identifying fake news is a very difficult task, especially when considering the multiple modes of conveying information through text, image, video and/or audio. We attempted to tackle the problem of automated misinformation/disinformation detection in multi-modal news sources (including text and images) through our simple, yet effective, approach in the FACTIFY shared task at De-Factify@AAAI2022. Our model produced an F1-weighted score of 74.807%, which was the fourth best out of all the submissions. In this paper we will explain our approach to undertake the shared task.