Membership Inference over Diffusion-models-based Synthetic Tabular Data
This work addresses privacy concerns for users of synthetic data generation in domains like healthcare or finance, but it is incremental as it builds on existing attack methods and models.
The study investigated privacy risks of diffusion-based synthetic tabular data generation by applying Membership Inference Attacks to TabDDPM and TabSyn, finding that TabDDPM is more vulnerable while TabSyn is resilient.
This study investigates the privacy risks associated with diffusion-based synthetic tabular data generation methods, focusing on their susceptibility to Membership Inference Attacks (MIAs). We examine two recent models, TabDDPM and TabSyn, by developing query-based MIAs based on the step-wise error comparison method. Our findings reveal that TabDDPM is more vulnerable to these attacks. TabSyn exhibits resilience against our attack models. Our work underscores the importance of evaluating the privacy implications of diffusion models and encourages further research into robust privacy-preserving mechanisms for synthetic data generation.