On the Sample Complexity of Adversarial Multi-Source PAC Learning
arXiv:2002.10384
Abstract
We study the problem of learning from multiple untrusted data sources, a scenario of increasing practical relevance given the recent emergence of crowdsourcing and collaborative learning paradigms. Specifically, we analyze the situation in which a learning system obtains datasets from multiple sources, some of which might be biased or even adversarially perturbed. It is known that in the single-source case, an adversary with the power to corrupt a fixed fraction of the training data can prevent PAC-learnability, that is, even in the limit of infinitely much training data, no learning system can approach the optimal test error. In this work we show that, surprisingly, the same is not true in the multi-source setting, where the adversary can arbitrarily corrupt a fixed fraction of the data sources. Our main results are a generalization bound that provides finite-sample guarantees for this learning setting, as well as corresponding lower bounds. Besides establishing PAC-learnability our results also show that in a cooperative learning setting sharing data with other parties has provable benefits, even if some participants are malicious.
arXiv commentInternational Conference on Machine Learning (ICML) 2020: Camera-ready. Strengthened the definition of adversarial PAC-learnability, added explicit bounds on sample complexity
References in corpus (13)
- A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networks
- A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attacks
- Certified Adversarial Robustness via Randomized Smoothing
- Byzantine-Robust Distributed Learning: Towards Optimal Statistical Rates
- Certified Defenses against Adversarial Examples
- Mitigating Sybils in Federated Learning Poisoning
- Byzantine Stochastic Gradient Descent
- Distributed Robust Learning
- Robust Learning from Untrusted Sources
- Universal Multi-Party Poisoning Attacks
- Efficient PAC Learning from the Crowd
- Optimal Robust Learning of Discrete Distributions from Batches
- On Fundamental Limits of Robust Learning