Testing the Generalization of Neural Language Models for COVID-19 Misinformation Detection

TL;DR

This study evaluates the generalization ability of fifteen Transformer-based models across diverse COVID-19 misinformation datasets, revealing that specialized tokenizers do not significantly outperform general-purpose ones, thus providing a realistic assessment for future research.

Contribution

It offers a comprehensive evaluation of models across multiple data types, highlighting the limited advantage of COVID-19-specific tokenizers and informing future misinformation detection efforts.

Findings

Tokenizers tailored to COVID-19 data do not significantly outperform general-purpose tokenizers.

Models show limited generalization across different content types and platforms.

The study provides a realistic benchmark for future misinformation detection research.

Abstract

A drastic rise in potentially life-threatening misinformation has been a by-product of the COVID-19 pandemic. Computational support to identify false information within the massive body of data on the topic is crucial to prevent harm. Researchers proposed many methods for flagging online misinformation related to COVID-19. However, these methods predominantly target specific content types (e.g., news) or platforms (e.g., Twitter). The methods' capabilities to generalize were largely unclear so far. We evaluate fifteen Transformer-based models on five COVID-19 misinformation datasets that include social media posts, news articles, and scientific papers to fill this gap. We show tokenizers and models tailored to COVID-19 data do not provide a significant advantage over general-purpose ones. Our study provides a realistic assessment of models for detecting COVID-19 misinformation. We expect that evaluating a broad spectrum of datasets and models will benefit future research in developing misinformation detection systems.