Reinforcement Learning from human feedback (RLHF) has been shown a promising direction for aligning generative models with human intent and has also been explored in recent works for alignment of diffusion generative models. In this work, we provide a rigorous treatment by formulating the task of fine-tuning diffusion models, with reward functions learned from human feedback, as an exploratory continuous-time stochastic control problem. Our key idea lies in treating the score-matching functions as controls/actions, and upon this, we develop a unified framework from a continuous-time perspective, to employ reinforcement learning (RL) algorithms in terms of improving the generation quality of diffusion models. We also develop the corresponding continuous-time RL theory for policy optimization and regularization under assumptions of stochastic different equations driven environment. Experiments on the text-to-image (T2I) generation will be reported in the accompanied paper.
@article{arxiv.2409.08400,
title = {Scores as Actions: a framework of fine-tuning diffusion models by continuous-time reinforcement learning},
author = {Hanyang Zhao and Haoxian Chen and Ji Zhang and David D. Yao and Wenpin Tang},
journal= {arXiv preprint arXiv:2409.08400},
year = {2024}
}