URVFL: Undetectable Data Reconstruction Attack on Vertical Federated Learning
This addresses a critical security vulnerability in VFL systems, where privacy is paramount, by exposing a stealthy attack method that could compromise client data without detection.
The paper tackles the problem of launching undetectable data reconstruction attacks in vertical federated learning (VFL) by developing URVFL, a novel attack strategy that uses label information to generate malicious gradients indistinguishable from honest ones, significantly outperforming existing attacks and circumventing state-of-the-art detection methods.
Launching effective malicious attacks in VFL presents unique challenges: 1) Firstly, given the distributed nature of clients' data features and models, each client rigorously guards its privacy and prohibits direct querying, complicating any attempts to steal data; 2) Existing malicious attacks alter the underlying VFL training task, and are hence easily detected by comparing the received gradients with the ones received in honest training. To overcome these challenges, we develop URVFL, a novel attack strategy that evades current detection mechanisms. The key idea is to integrate a discriminator with auxiliary classifier that takes a full advantage of the label information and generates malicious gradients to the victim clients: on one hand, label information helps to better characterize embeddings of samples from distinct classes, yielding an improved reconstruction performance; on the other hand, computing malicious gradients with label information better mimics the honest training, making the malicious gradients indistinguishable from the honest ones, and the attack much more stealthy. Our comprehensive experiments demonstrate that URVFL significantly outperforms existing attacks, and successfully circumvents SOTA detection methods for malicious attacks. Additional ablation studies and evaluations on defenses further underscore the robustness and effectiveness of URVFL. Our code will be available at https://github.com/duanyiyao/URVFL.