Poisoning Attacks and Subpopulation Susceptibility



Evan Rose, David Evans, Fnu Suya

 View presentation:2022-10-17T21:15:00ZGMT-0600Change your timezone on the schedule page
2022-10-17T21:15:00Z

Exemplar figure, described by caption below — Machine learning is susceptible to poisoning attacks, in which adversaries inject maliciously crafted training data into the training set to induce specific model behavior. What happens when attackers focus their efforts on specific subpopulations of the input space?

The live footage of the talk, including the Q&A, can be viewed on the session page, VISxAI: Session 2, Session 3, Closing.

Abstract