Unleashing the Tiger: Inference Attacks on Split Learning

TL;DR

This paper reveals significant security vulnerabilities in split learning, demonstrating that malicious actors can reconstruct private data and hijack the training process, even against recent defenses, through various attack strategies.

Contribution

It introduces new attack methods that compromise split learning security, showing the protocol's vulnerabilities against both malicious servers and clients, and evaluates their effectiveness.

Findings

Attacks can reconstruct private training data.

Malicious servers can hijack the training process.

Existing defenses are ineffective against these attacks.

Abstract

We investigate the security of Split Learning -- a novel collaborative machine learning framework that enables peak performance by requiring minimal resources consumption. In the present paper, we expose vulnerabilities of the protocol and demonstrate its inherent insecurity by introducing general attack strategies targeting the reconstruction of clients' private training sets. More prominently, we show that a malicious server can actively hijack the learning process of the distributed model and bring it into an insecure state that enables inference attacks on clients' data. We implement different adaptations of the attack and test them on various datasets as well as within realistic threat scenarios. We demonstrate that our attack is able to overcome recently proposed defensive techniques aimed at enhancing the security of the split learning protocol. Finally, we also illustrate the protocol's insecurity against malicious clients by extending previously devised attacks for Federated Learning. To make our results reproducible, we made our code available at https://github.com/pasquini-dario/SplitNN_FSHA.