Spider GAN：利用友好邻居加速GAN训练
Spider GAN: Leveraging Friendly Neighbors to Accelerate GAN Training
AI总结：
本文提出Spider GAN，利用图像输入和友好邻域度量（SID）加速GAN训练，通过级联迁移实现快速收敛，在多个数据集上以五分之一迭代达到最先进FID。
摘要：
稳定地训练生成对抗网络（GAN）是一项具有挑战性的任务。GAN中的生成器将噪声向量（通常为高斯分布）转换为逼真的数据（如图像）。在本文中，我们提出了一种以图像为输入训练GAN的新方法，但无需施加任何成对约束。其直觉是，图像比噪声更具结构性，生成器可以利用这一点来学习更稳健的变换。通过识别紧密相关的数据集，即目标分布的“友好邻域”，可以使该过程高效，这启发了Spider GAN这一名称。为了利用数据集之间的邻近性来定义友好邻域，我们提出了一种新的度量，称为符号初始距离（SID），其灵感来源于多调和核。我们表明，Spider GAN公式能实现更快的收敛，因为生成器甚至可以发现看似不相关数据集之间的对应关系，例如Tiny-ImageNet和CelebA人脸之间。此外，我们展示了级联Spider GAN，其中预训练GAN生成器的输出分布用作后续网络的输入。有效地，以级联方式将一个分布传输到另一个分布，直到学习到目标分布——这是一种新的迁移学习形式。我们在DCGAN、条件GAN、PGGAN、StyleGAN2和StyleGAN3上展示了Spider方法的有效性。所提出的方法在MetFaces、Ukiyo-E Faces和AFHQ-Cats等高分辨率小数据集上，以五分之一的训练迭代次数，达到了最先进的弗雷歇初始距离（FID）值，与基线方法相比。
英文摘要：
Training Generative adversarial networks (GANs) stably is a challenging task. The generator in GANs transform noise vectors, typically Gaussian distributed, into realistic data such as images. In this paper, we propose a novel approach for training GANs with images as inputs, but without enforcing any pairwise constraints. The intuition is that images are more structured than noise, which the generator can leverage to learn a more robust transformation. The process can be made efficient by identifying closely related datasets, or a ``friendly neighborhood'' of the target distribution, inspiring the moniker, Spider GAN. To define friendly neighborhoods leveraging proximity between datasets, we propose a new measure called the signed inception distance (SID), inspired by the polyharmonic kernel. We show that the Spider GAN formulation results in faster convergence, as the generator can discover correspondence even between seemingly unrelated datasets, for instance, between Tiny-ImageNet and CelebA faces. Further, we demonstrate cascading Spider GAN, where the output distribution from a pre-trained GAN generator is used as the input to the subsequent network. Effectively, transporting one distribution to another in a cascaded fashion until the target is learnt -- a new flavor of transfer learning. We demonstrate the efficacy of the Spider approach on DCGAN, conditional GAN, PGGAN, StyleGAN2 and StyleGAN3. The proposed approach achieves state-of-the-art Frechet inception distance (FID) values, with one-fifth of the training iterations, in comparison to their baseline counterparts on high-resolution small datasets such as MetFaces, Ukiyo-E Faces and AFHQ-Cats.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
