TL;DR
This work treats diffusion-based image generation as a sequential decision process and uses Diffusion Denoising Policy Optimization (DDPO) to learn a policy that steers generation away from a chosen target concept - in this case, "teddy bear."
Text-to-image diffusion models have achieved remarkable progress in image synthesis, but their potential misuse for generating unauthorized or harmful content has raised growing safety concerns. This has created an urgent need for safe diffusion-based image generation methods that can selectively suppress sensitive con...
The results reveal that attention-space regulation offers a considerably more promising path to safer diffusion transformer based image generation than the existing concept erasing mechanism.
This paper learns the guidance schedule as a function of diffusion time, conditioning and the current noisy sample, in order to better align sampled images with the text prompt.
A scalable and flexible solution for safe generative deployment that allows weight-preserving detection, keeps non-violating inference latency nearly unchanged as the maximum budget increases, and offers flexibility in safety performance via test-time scaling.
This paper forms few-step generation as a controlled base generative process, and shows that self-consistency loss can be understood through the lens of optimal control, and draws a connection between this approach and reinforcement learning, potentially opening the door to a new set of approaches for few-step generati...
Uni-AdaVD is presented, a universal inference-time concept erasure framework for visual generation that treats the value space of multimodal attention as a unified intervention space and introduces encoder-aware target representation construction to localize target semantics across heterogeneous text encoders.
We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.
