Unleashing the Tiger: Inference Attacks on Split Learning
This work exposes critical security flaws in Split Learning, a collaborative machine learning framework, impacting users and organizations relying on its privacy claims.
This paper investigates the security of Split Learning, demonstrating vulnerabilities that allow a malicious server to reconstruct clients' private training data and hijack the learning process. The proposed attacks are effective against various datasets and realistic threat scenarios, even overcoming recent defensive techniques.
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.