Securing Federated Sensitive Topic Classification against Poisoning Attacks

Abstract

We present a Federated Learning (FL) based solution for building a distributed classifier capable of detecting URLs containing GDPR-sensitive content related to categories such as health, sexual preference, political beliefs, etc. Although such a classifier addresses the limitations of previous offline/centralised classifiers,it is still vulnerable to poisoning attacks from malicious users that may attempt to reduce the accuracy for benign users by disseminating faulty model updates. To guard against this, we develop a robust aggregation scheme based on subjective logic and residual-based attack detection. Employing a combination of theoretical analysis, trace-driven simulation, as well as experimental validation with a prototype and real users, we show that our classifier can detect sensitive content with high accuracy, learn new labels fast, and remain robust in view of poisoning attacks from malicious users, as well as imperfect input from non-malicious ones.

References in corpus (13)

• Federated Learning for Mobile Keyboard Prediction
• How To Backdoor Federated Learning
• Learning Differentially Private Recurrent Language Models
• Byzantine-Robust Distributed Learning: Towards Optimal Statistical Rates
• The Hidden Vulnerability of Distributed Learning in Byzantium
• Analyzing Federated Learning through an Adversarial Lens
• Can You Really Backdoor Federated Learning?
• Mitigating Sybils in Federated Learning Poisoning
• Geometric median and robust estimation in Banach spaces
• DeepSight: Mitigating Backdoor Attacks in Federated Learning Through Deep Model Inspection
• Federated Learning: Opportunities and Challenges
• Attack-Resistant Federated Learning with Residual-based Reweighting
• Beyond content analysis: Detecting targeted ads via distributed counting

Cited by in corpus (1)

• Location Leakage in Federated Signal Maps