Leveraging Label Information for Stealthy Data Stealing in Vertical Federated Learning

arXiv:2404.19582v1 Announce Type: new
Abstract: We develop DMAVFL, 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 (which was completely ignored in previous attacks): 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 …

abstract arxiv attacks classifier cs.cr cs.lg current data detection federated learning information key novel stealing strategy the key type