4 citations · 15 across the 24 of their papers we have counts for
1 paper · 1 filter
Yota Ueda, Shinnosuke Takamichi, Yuki Saito +2
We propose {\it HumanDiffusion,} a diffusion model trained from humans' perceptual gradients to learn an acceptable range of data for humans (i.e., human-acceptable distribution).…