论文标题
SPI-GAN：用直径插值的剥离扩散gan
SPI-GAN: Denoising Diffusion GANs with Straight-Path Interpolations
论文作者
论文摘要
基于得分的生成模型（SGM）显示了最新的采样质量和多样性。但是，由于高度复杂的前进/反向过程，他们的训练/抽样复杂性众所周知，因此它们不适合资源有限的设置。要解决这个问题，学习一个更简单的过程目前正在吸引很多关注。我们使用我们提出的直线插值定义提出了一种增强的基于GAN的denoising方法，称为SpiGAN。为此，我们提出了一个gan架构i）通过直接路径进行降解，而ii）以连续映射神经网络模仿来模仿denoising路径的特征。这种方法大大减少了采样时间，同时以高度采样质量和多样性与SGM相同。结果，SPI-GAN是CIFAR-10和Celeba-HQ-256的采样质量，多样性和时间中最优秀的模型之一。
Score-based generative models (SGMs) show the state-of-the-art sampling quality and diversity. However, their training/sampling complexity is notoriously high due to the highly complicated forward/reverse processes, so they are not suitable for resource-limited settings. To solving this problem, learning a simpler process is gathering much attention currently. We present an enhanced GAN-based denoising method, called SPI-GAN, using our proposed straight-path interpolation definition. To this end, we propose a GAN architecture i) denoising through the straight-path and ii) characterized by a continuous mapping neural network for imitating the denoising path. This approach drastically reduces the sampling time while achieving as high sampling quality and diversity as SGMs. As a result, SPI-GAN is one of the best-balanced models among the sampling quality, diversity, and time for CIFAR-10, and CelebA-HQ-256.
