On Unifying Misinformation Detection

TL;DR

UnifiedM2 is a comprehensive misinformation detection model that jointly addresses multiple misinformation tasks, improving performance and generalizability across domains and enabling effective few-shot learning for unseen misinformation types.

Contribution

The paper introduces UnifiedM2, a unified model that jointly learns multiple misinformation detection tasks, enhancing representation richness and generalization capabilities.

Findings

Achieves state-of-the-art or comparable results across four misinformation tasks.

Demonstrates improved few-shot learning for unseen misinformation datasets.

Shows enhanced generalizability to unseen events.

Abstract

In this paper, we introduce UnifiedM2, a general-purpose misinformation model that jointly models multiple domains of misinformation with a single, unified setup. The model is trained to handle four tasks: detecting news bias, clickbait, fake news, and verifying rumors. By grouping these tasks together, UnifiedM2learns a richer representation of misinformation, which leads to state-of-the-art or comparable performance across all tasks. Furthermore, we demonstrate that UnifiedM2's learned representation is helpful for few-shot learning of unseen misinformation tasks/datasets and model's generalizability to unseen events.