Adversarial Feature Learning
Jeff Donahue Computer Science Division University of California, Berkeley、Philipp Krähenbühl Department of Computer Science University of Texas, Austin、Trevor Darrell Computer Science Division University of California, Berkeley
Abstract
The ability of the Generative Adversarial Networks (GANs) framework to learn generative models mapping from simple latent distributions to arbitrarily complex data distributions has been demonstrated empirically, with compelling results showing that the latent space of such generators captures semantic variation in the data distribution. Intuitively, models trained to predict these semantic latent representations given data may serve as useful feature representations for auxiliary problems where semantics are relevant. However, in their existing form, GANs have no means of learning the inverse mapping – projecting data back into the latent space. We propose Bidirectional Generative Adversarial Networks (BiGANs) as a means of learning this inverse mapping, and demonstrate that the resulting learned feature representation is useful for auxiliary supervised discrimination tasks, competitive with contemporary approaches to unsupervised and self-supervised feature learning.
中文速览
生成对抗网络（GAN）能把随机噪声映射成逼真图像，其潜在空间（latent space）还隐含着语义信息，但标准GAN只有"从噪声到图像"的单向映射，无法把真实数据反向编码回潜在空间，因此难以直接用作通用特征提取器。本文提出双向生成对抗网络（BiGAN），在原有生成器之外引入一个编码器，让判别器同时在"数据＋潜在编码"的联合空间中工作，从而迫使编码器和生成器在对抗训练中自发地互为逆映射——论文从理论上严格证明了这一性质。在ImageNet等基准上的实验表明，BiGAN编码器学到的特征在图像分类等下游任务中与当时最先进的无监督和自监督方法相当，验证了该框架无需任何标注即可学习有意义语义表示的能力，为GAN在无监督特征学习领域的应用开辟了新路径。
原文 arXiv:1605.09782;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1605.09782v7