We propose Diffusion Adaptive Text Embedding (DATE), which improves text-to-image diffusion models by dynamically refining text embeddings throughout the diffusion sampling process.

Our paper introduces Semantic Surgery, a novel training-free method that dynamically manipulates text embeddings before diffusion to achieve concept erasure in text-to-image models while preserving image quality and locality.

HiFlow is a training-free and model-agnostic framework to unlock the resolution potential of pre-trained flow models.

We propose a pipeline for detecting visual inconsistencies using visual correspondences based on disentangled diffusion model features

We propose ADRPO, a method that dynamically adjusts divergence regularization strength based on advantage estimates, enabling more effective fine-tuning of generative models by automatically balancing exploration and exploitation at the sample level.

EraseFlow surgically removes unwanted concepts with reward-free alignment, preserving prior and setting a new standard for safe, efficient diffusion models.