1 paper
Shaomeng Wang, He Wang, Xiaolu Wei +2
Diffusion models have achieved remarkable success in conditional image generation, yet their outputs often remain misaligned with human preferences. To address this, recent work ha…