Full Professor, National Yang Ming Chiao Tung University
1 paper at NeurIPS 2025
We present a learning framework that aligns text-to-image diffusion models with human preferences through inverse reinforcement learning and a balance of offline and online training.
