Can Fake News Detection be Accountable? The Adversarial Examples Challenge

Bogaert, Jérémie;Carbonnelle, Quentin;Descampe, Antonin;Standaert, François-Xavier
(2021) 41st WIC Symposium on Information Theory in the Benelux — Location: Online (20.May.2021)

Files

265.pdf
  • Open Access
  • Adobe PDF
  • 796.33 KB
Download

Details

Authors
Abstract
Automated fake news detection is an important challenge in view of the increasing ability of statistical language models to generate large amounts of (possibly fake) articles, so that recognizing them manually becomes unrealistic. Yet, the reliable deployment of such automated detection tools would require ensuring that they are accountable. Algorithmic accountability is known to be difficult to reach, especially when adversarial behaviors aim to make algorithms deviate from their expected mode of operation. In this paper, we illustrate with a case study that this challenge is further amplified in contexts where the labeling of the articles is prone to errors, which is the case of fake news detection
Affiliations

Citations

  • APA
  • Chicago
  • FWB

Bogaert, J., Carbonnelle, Q., Descampe, A., & Standaert, F.-X. (2021). Can Fake News Detection be Accountable? The Adversarial Examples Challenge. Proceedings of the 2021 Symposium on Information Theory and Signal Processing in the Benelux, 25-32. https://hdl.handle.net/2078.5/232489