Adversarial machine learning for protecting against online manipulation
This work addresses the challenge of online manipulation for users and platforms by improving detection systems, though it appears incremental as it applies existing adversarial techniques to specific tasks.
The paper tackles the problem of adversarial examples in machine learning, which can cause incorrect outputs and severe consequences like misclassifying traffic signs, and explores using them to build stronger models for fake news and social bot detection.
Adversarial examples are inputs to a machine learning system that result in an incorrect output from that system. Attacks launched through this type of input can cause severe consequences: for example, in the field of image recognition, a stop signal can be misclassified as a speed limit indication.However, adversarial examples also represent the fuel for a flurry of research directions in different domains and applications. Here, we give an overview of how they can be profitably exploited as powerful tools to build stronger learning models, capable of better-withstanding attacks, for two crucial tasks: fake news and social bot detection.