Harbin
China
Heilongjiang University
Compositional Generation, Diffusion Models, Feedback Learning, Semantic Alignment, Reward Function