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). Conventional HumanGAN aims to model the human-acceptable distribution wider than the real-data distribution by training a neural network-based generator with human-based discriminators. However, HumanGAN training tends to converge in a meaningless distribution due to the gradient vanishing or mode collapse and requires careful heuristics. In contrast, our HumanDiffusion learns the human-acceptable distribution through Langevin dynamics based on gradients of human perceptual evaluations. Our training iterates a process to diffuse real data to cover a wider human-acceptable distribution and can avoid the issues in the HumanGAN training. The evaluation results demonstrate that our HumanDiffusion can successfully represent the human-acceptable distribution without any heuristics for the training.

翻译：我们提出HumanDiffusion，一种利用人类感知梯度训练的扩散模型，旨在学习人类可接受的数据范围（即人类可接受分布）。传统HumanGAN通过基于人类判别器训练神经网络生成器，试图建模比真实数据分布更宽泛的人类可接受分布。然而，HumanGAN训练常因梯度消失或模式崩溃而收敛于无意义的分布，且需要精心设计的启发式策略。相比之下，我们的HumanDiffusion基于人类感知评估的梯度，通过朗之万动力学学习人类可接受分布。其训练过程迭代扩散真实数据以覆盖更广泛的人类可接受分布，从而避免HumanGAN训练中的问题。评估结果表明，HumanDiffusion能够无需任何训练启发式策略，成功表征人类可接受分布。