ReTrack: Data Unlearning in Diffusion Models through Redirecting the Denoising Trajectory
This addresses privacy and safety issues in diffusion models for users and developers, though it is an incremental improvement over existing unlearning methods.
The paper tackles the problem of data memorization in diffusion models, which raises privacy and safety concerns, by proposing ReTrack, a fast and effective data unlearning method that redirects denoising trajectories toward k-nearest neighbors, achieving state-of-the-art performance with the best trade-off between unlearning strength and generation quality preservation.
Diffusion models excel at generating high-quality, diverse images but suffer from training data memorization, raising critical privacy and safety concerns. Data unlearning has emerged to mitigate this issue by removing the influence of specific data without retraining from scratch. We propose ReTrack, a fast and effective data unlearning method for diffusion models. ReTrack employs importance sampling to construct a more efficient fine-tuning loss, which we approximate by retaining only dominant terms. This yields an interpretable objective that redirects denoising trajectories toward the $k$-nearest neighbors, enabling efficient unlearning while preserving generative quality. Experiments on MNIST T-Shirt, CelebA-HQ, CIFAR-10, and Stable Diffusion show that ReTrack achieves state-of-the-art performance, striking the best trade-off between unlearning strength and generation quality preservation.