An Adversarial Benchmark for Fake News Detection Models

TL;DR

This paper introduces an adversarial benchmark to evaluate fake news detection models' reasoning about real-world facts, revealing their vulnerabilities to compositional and lexical changes.

Contribution

It presents a novel adversarial benchmark targeting reasoning aspects of fake news detectors, highlighting their weaknesses and the need for combined fact-checking approaches.

Findings

Models fail to respond to compositional and lexical changes

BERT classifiers show significant vulnerabilities

Highlights the need for improved reasoning in detection models

Abstract

With the proliferation of online misinformation, fake news detection has gained importance in the artificial intelligence community. In this paper, we propose an adversarial benchmark that tests the ability of fake news detectors to reason about real-world facts. We formulate adversarial attacks that target three aspects of "understanding": compositional semantics, lexical relations, and sensitivity to modifiers. We test our benchmark using BERT classifiers fine-tuned on the LIAR arXiv:arch-ive/1705648 and Kaggle Fake-News datasets, and show that both models fail to respond to changes in compositional and lexical meaning. Our results strengthen the need for such models to be used in conjunction with other fact checking methods.