Privacy Risks of Securing Machine Learning Models against Adversarial Examples

Abstract

The arms race between attacks and defenses for machine learning models has come to a forefront in recent years, in both the security community and the privacy community. However, one big limitation of previous research is that the security domain and the privacy domain have typically been considered separately. It is thus unclear whether the defense methods in one domain will have any unexpected impact on the other domain. In this paper, we take a step towards resolving this limitation by combining the two domains. In particular, we measure the success of membership inference attacks against six state-of-the-art defense methods that mitigate the risk of adversarial examples (i.e., evasion attacks). Membership inference attacks determine whether or not an individual data record has been part of a model's training set. The accuracy of such attacks reflects the information leakage of training algorithms about individual members of the training set. Adversarial defense methods against adversarial examples influence the model's decision boundaries such that model predictions remain unchanged for a small area around each input. However, this objective is optimized on training data. Thus, individual data records in the training set have a significant influence on robust models. This makes the models more vulnerable to inference attacks. To perform the membership inference attacks, we leverage the existing inference methods that exploit model predictions. We also propose two new inference methods that exploit structural properties of robust models on adversarially perturbed data. Our experimental evaluation demonstrates that compared with the natural training (undefended) approach, adversarial defense methods can indeed increase the target model's risk against membership inference attacks.

arXiv commentACM CCS 2019, code is available at https://github.com/inspire-group/privacy-vs-robustness

References in corpus (4)

• Deep Learning with Differential Privacy
• Comprehensive Privacy Analysis of Deep Learning: Passive and Active White-box Inference Attacks against Centralized and Federated Learning
• Wild Patterns: Ten Years After the Rise of Adversarial Machine Learning
• Evasion Attacks against Machine Learning at Test Time

Cited by in corpus (16)

• A Survey of Privacy Attacks in Machine Learning
• Defensive Approximation: Securing CNNs using Approximate Computing
• Data Privacy and Trustworthy Machine Learning
• Audit to Forget: A Unified Method to Revoke Patients' Private Data in Intelligent Healthcare
• Membership Inference Attacks and Generalization: A Causal Perspective
• Knowledge Cross-Distillation for Membership Privacy
• TransMIA: Membership Inference Attacks Using Transfer Shadow Training
• Bounding Information Leakage in Machine Learning
• Fault Tolerance of Neural Networks in Adversarial Settings
• NeuGuard: Lightweight Neuron-Guided Defense against Membership Inference Attacks
• Is My Data in Your AI? Membership Inference Test (MINT) applied to Face Biometrics
• Careful What You Wish For: on the Extraction of Adversarially Trained Models
• GECKO: Reconciling Privacy, Accuracy and Efficiency in Embedded Deep Learning
• Benchmarking Federated Learning in Edge Computing Environments: A Systematic Review and Performance Evaluation
• Split Unlearning
• MUSO: Achieving Exact Machine Unlearning in Over-Parameterized Regimes