A Framework of Randomized Selection Based Certified Defenses Against Data Poisoning Attacks
arXiv:2009.08739
Abstract
Neural network classifiers are vulnerable to data poisoning attacks, as attackers can degrade or even manipulate their predictions thorough poisoning only a few training samples. However, the robustness of heuristic defenses is hard to measure. Random selection based defenses can achieve certified robustness by averaging the classifiers' predictions on the sub-datasets sampled from the training set. This paper proposes a framework of random selection based certified defenses against data poisoning attacks. Specifically, we prove that the random selection schemes that satisfy certain conditions are robust against data poisoning attacks. We also derive the analytical form of the certified radius for the qualified random selection schemes. The certified radius of bagging derived by our framework is tighter than the previous work. Our framework allows users to improve robustness by leveraging prior knowledge about the training set and the poisoning model. Given higher level of prior knowledge, we can achieve higher certified accuracy both theoretically and practically. According to the experiments on three benchmark datasets: MNIST 1/7, MNIST, and CIFAR-10, our method outperforms the state-of-the-art.
References in corpus (15)
- Targeted Backdoor Attacks on Deep Learning Systems Using Data Poisoning
- Poisoning Attacks against Support Vector Machines
- Spectral Signatures in Backdoor Attacks
- Is feature selection secure against training data poisoning?
- Detecting Backdoor Attacks on Deep Neural Networks by Activation Clustering
- MetaPoison: Practical General-purpose Clean-label Data Poisoning
- Defending Neural Backdoors via Generative Distribution Modeling
- On Certifying Robustness against Backdoor Attacks via Randomized Smoothing
- RAB: Provable Robustness Against Backdoor Attacks
- Label-Consistent Backdoor Attacks
- Curse of Dimensionality on Randomized Smoothing for Certifiable Robustness
- Certified Robustness for Top-k Predictions against Adversarial Perturbations via Randomized Smoothing
- Data Poisoning against Differentially-Private Learners: Attacks and Defenses
- Certified Robustness to Label-Flipping Attacks via Randomized Smoothing
- Regula Sub-rosa: Latent Backdoor Attacks on Deep Neural Networks