Securing Federated Sensitive Topic Classification against Poisoning Attacks

TL;DR

This paper introduces a federated learning approach for sensitive URL classification that is robust against poisoning attacks, using novel aggregation and attack detection methods validated through simulations and real-world testing.

Contribution

It proposes a robust aggregation scheme based on subjective logic and residual detection to defend against poisoning attacks in federated sensitive content classification.

Findings

High accuracy in detecting sensitive URLs

Fast learning of new labels

Robustness against malicious 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.