Timezone: »

 
Hidden Poison: Machine Unlearning Enables Camouflaged Poisoning Attacks
Jimmy Di · Jack Douglas · Jayadev Acharya · Gautam Kamath · Ayush Sekhari

@
in Trustworthy and Socially Responsible Machine Learning »
Event URL: https://openreview.net/forum?id=BrgzlYUUT3n »

We introduce camouflaged data poisoning attacks, a new attack vector that arises in the context of machine unlearning and other settings when model retraining may be induced. An adversary first adds a few carefully crafted points to the training dataset such that the impact on the model's predictions is minimal. The adversary subsequently triggers a request to remove a subset of the introduced points at which point the attack is unleashed and the model's predictions are negatively affected. In particular, we consider clean-label targeted attacks (in which the goal is to cause the model to misclassify a specific test point) on datasets including CIFAR-10, Imagenette, and Imagewoof. This attack is realized by constructing camouflage datapoints that mask the effect of a poisoned dataset.

[ Poster]  [ OpenReview]  [ Topia] 

Author Information

Jimmy Di (University of Waterloo)
Jack Douglas (University of Waterloo)
Jayadev Acharya (Cornell University)
Gautam Kamath (University of Waterloo)
Ayush Sekhari (Cornell University)

More from the Same Authors