English

Adding Additional Control to One-Step Diffusion with Joint Distribution Matching

Computer Vision and Pattern Recognition 2025-03-13 v2

Abstract

While diffusion distillation has enabled one-step generation through methods like Variational Score Distillation, adapting distilled models to emerging new controls -- such as novel structural constraints or latest user preferences -- remains challenging. Conventional approaches typically requires modifying the base diffusion model and redistilling it -- a process that is both computationally intensive and time-consuming. To address these challenges, we introduce Joint Distribution Matching (JDM), a novel approach that minimizes the reverse KL divergence between image-condition joint distributions. By deriving a tractable upper bound, JDM decouples fidelity learning from condition learning. This asymmetric distillation scheme enables our one-step student to handle controls unknown to the teacher model and facilitates improved classifier-free guidance (CFG) usage and seamless integration of human feedback learning (HFL). Experimental results demonstrate that JDM surpasses baseline methods such as multi-step ControlNet by mere one-step in most cases, while achieving state-of-the-art performance in one-step text-to-image synthesis through improved usage of CFG or HFL integration.

Keywords

Cite

@article{arxiv.2503.06652,
  title  = {Adding Additional Control to One-Step Diffusion with Joint Distribution Matching},
  author = {Yihong Luo and Tianyang Hu and Yifan Song and Jiacheng Sun and Zhenguo Li and Jing Tang},
  journal= {arXiv preprint arXiv:2503.06652},
  year   = {2025}
}
R2 v1 2026-06-28T22:12:57.347Z