2019 · 106 citations · 14 references
ComputingComputer ScienceComputer SecurityData MiningMachine LearningData PrivacyData ProtectionData SecurityDifferential PrivacyData Poisoning AttacksDifferentially Private Machine LearningOutput PerturbationPrivacy-preserving LearningPrivacy-preserving Machine LearningPrivacy-preserving Model TrainingPrivate Machine Learning
Data poisoning attacks aim to manipulate the model produced by a learning algorithm by adversarially modifying the training set. We consider differential privacy as a defensive measure against this type of attack. We show that private learners are resistant to data poisoning attacks when the adversary is only able to poison a small number of items. However, this protection degrades as the adversary is allowed to poison more data. We emprically evaluate this protection by designing attack algorithms targeting objective and output perturbation learners, two standard approaches to differentially-private machine learning. Experiments show that our methods are effective when the attacker is allowed to poison sufficiently many training items.
14
Intriguing properties of neural networks
Christian Szegedy, Wojciech Zaremba, Ilya Sutskever et al. · arXiv (Cornell University) · 2013 · 5.7K citations · Full text