Provably effective detection of effective data poisoning attacks
Main:30 Pages
6 Figures
Bibliography:5 Pages
4 Tables
Abstract
This paper establishes a mathematically precise definition of dataset poisoning attack and proves that the very act of effectively poisoning a dataset ensures that the attack can be effectively detected. On top of a mathematical guarantee that dataset poisoning is identifiable by a new statistical test that we call the Conformal Separability Test, we provide experimental evidence that we can adequately detect poisoning attempts in the real world.
View on arXivComments on this paper
