Across continual erasure of celebrities, artistic styles, and instances, CEASE achieves the most consistent erase-preserve trade-off, while existing methods either degrade general generation or insufficiently erase targets.
Yong-Liang Wu, Haori Lu, Jin-Qi Luo et al.· 0 citations
Concept erasure aims to remove a target concept, such as a copyrighted style, a recognizable character, or unsafe content, from a pretrained text-to-image diffusion model while preserving its ability to generate other content. Existing activation steering methods build an erasure direction mainly from the target concep...
Yong-Liang Wu, Haori Lu, Yu-Lun Wu et al.· 1 citation
Precise control over camera and object motion is essential for professional video production. Existing methods control objects only coarsely, through image-plane cues that are ambiguous in depth and rotation or through 3D tracks and blobs that lack complete geometry and lose consistency across viewpoint changes. We int...
Wei Cao, Hao Zhang, Vikram Voleti et al.· 1 citation
This work proposes AlphaWiSE, a post-hoc weight-space interpolation method that composes two frozen source checkpoints that fits one scalar interpolation coefficient shared by all tensor entries for cross-modal retrieval.