TL;DR该论文提出了一种名为 rGAN 的新型 GAN 模型,通过加入噪声转移模型,在训练标签出现噪声的情况下,可以学习到一个干净的带条件的生成分布。该模型在不同的噪声设置和多种评估指标下得到了有效的表现。
Abstract
generative adversarial networks (GANs) are a framework that learns a
generative distribution through adversarial training. Recently, their
class-conditional extensions (e.g., conditional gan (cGAN) and auxiliary