Model Will Tell: Training Membership Inference for Diffusion Models

arXiv:2403.08487v1 Announce Type: new
Abstract: Diffusion models pose risks of privacy breaches and copyright disputes, primarily stemming from the potential utilization of unauthorized data during the training phase. The Training Membership Inference (TMI) task aims to determine whether a specific sample has been used in the training process of a target model, representing a critical tool for privacy violation verification. However, the increased stochasticity inherent in diffusion renders traditional shadow-model-based or metric-based methods ineffective when applied to diffusion models. Moreover, …

abstract arxiv breaches copyright cs.cv data diffusion diffusion models inference privacy process risks sample stemming training type will