Stay Fair! Ensuring Group Fairness in Diffusion Models Across Guidance Scales 事件

Stay Fair! Ensuring Group Fairness in Diffusion Models Across Guidance Scales arXiv:2605.28036v1 Announce Type: new Abstract: Diffusion models steer conditional generation with a tunable guidance scale to trade off prompt alignment and diversity. However, existing debiasing techniques are optimized for a single scale, degrading fairness when users adjust this parameter. We trace this behavior to a previously overlooked source by decomposing total bias into two components: a model bias and a gui