How to Train Your Fact Verifier: Knowledge Transfer with Multimodal Open Models

TL;DR

This paper explores how large multimodal models can be trained for fact verification and misinformation detection without continual updates, using knowledge transfer techniques evaluated on multiple benchmarks.

Contribution

It demonstrates that knowledge transfer strategies can enhance multimodal fact-checking performance on recent benchmarks without ongoing model updates.

Findings

Knowledge transfer improves Fakeddit accuracy by up to 1.7%.

Knowledge transfer improves Mocheg accuracy by up to 2.9%.

Models can benefit from existing benchmarks and explanations without continual retraining.

Abstract

Given the growing influx of misinformation across news and social media, there is a critical need for systems that can provide effective real-time verification of news claims. Large language or multimodal model based verification has been proposed to scale up online policing mechanisms for mitigating spread of false and harmful content. While these can potentially reduce burden on human fact-checkers, such efforts may be hampered by foundation model training data becoming outdated. In this work, we test the limits of improving foundation model performance without continual updating through an initial study of knowledge transfer using either existing intra- and inter- domain benchmarks or explanations generated from large language models (LLMs). We evaluate on 12 public benchmarks for fact-checking and misinformation detection as well as two other tasks relevant to content moderation -- toxicity and stance detection. Our results on two recent multi-modal fact-checking benchmarks, Mocheg and Fakeddit, indicate that knowledge transfer strategies can improve Fakeddit performance over the state-of-the-art by up to 1.7% and Mocheg performance by up to 2.9%.