How To Backdoor Federated Learning

TL;DR

This paper reveals that federated learning systems are vulnerable to backdoor attacks where malicious participants can embed hidden functionalities into the global model, bypassing defenses and achieving perfect backdoor accuracy.

Contribution

The authors introduce a novel model-poisoning attack method based on model replacement that can immediately implant effective backdoors in federated learning models, outperforming existing data poisoning techniques.

Findings

Attack can reach 100% backdoor accuracy in a single round.

Outperforms data poisoning in effectiveness.

Evades anomaly detection defenses through a new constraint-and-scale technique.

Abstract

Federated learning enables thousands of participants to construct a deep learning model without sharing their private training data with each other. For example, multiple smartphones can jointly train a next-word predictor for keyboards without revealing what individual users type. We demonstrate that any participant in federated learning can introduce hidden backdoor functionality into the joint global model, e.g., to ensure that an image classifier assigns an attacker-chosen label to images with certain features, or that a word predictor completes certain sentences with an attacker-chosen word. We design and evaluate a new model-poisoning methodology based on model replacement. An attacker selected in a single round of federated learning can cause the global model to immediately reach 100% accuracy on the backdoor task. We evaluate the attack under different assumptions for the standard federated-learning tasks and show that it greatly outperforms data poisoning. Our generic constrain-and-scale technique also evades anomaly detection-based defenses by incorporating the evasion into the attacker's loss function during training.