Yesterday's News: Benchmarking Multi-Dimensional Out-of-Distribution Generalization of Misinformation Detection Models

2024/10/12 by Ivo Verhoeven, Pushkar Mishra, Verhoeven, Ivo +3 · 1 citation
#cs.IR #cs.CL

paper · pdf · doi:10.48550/arxiv.2410.18122

Abstract

This article introduces misinfo-general, a benchmark dataset for evaluating misinformation models' ability to perform out-of-distribution generalization. Misinformation changes rapidly, much more quickly than moderators can annotate at scale, resulting in a shift between the training and inference data distributions. As a result, misinformation detectors need to be able to perform out-of-distribution generalization, an attribute they currently lack. Our benchmark uses distant labelling to enable simulating covariate shifts in misinformation content. We identify time, event, topic, publisher, political bias, misinformation type as important axes for generalization, and we evaluate a common class of baseline models on each. Using article metadata, we show how this model fails desiderata, which is not necessarily obvious from classification metrics. Finally, we analyze properties of the data to ensure limited presence of modelling shortcuts. We make the dataset and accompanying code publicly available: https://github.com/ioverho/misinfo-general

Cited by

• Benchmarking Debiasing Methods for LLM-based Parameter Estimates

Related

• Continuously Reliable Detection of New-Normal Misinformation: Semantic Masking and Contrastive Smoothing in High-Density Latent Regions
• On Unifying Misinformation Detection
• A Guide to Misinformation Detection Data and Evaluation
• Defining Misinformation and Understanding its Bounded Nature: Using Expertise and Evidence for Describing Misinformation
• Multimodal Misinformation Detection by Learning from Synthetic Data with Multimodal LLMs