Inverting the Generator of a Generative Adversarial Network
Inverting the Generator of a Generative Adversarial Network
复制标题
生成对抗网络生成器的逆向操作
DOI：
10.1109/tnnls.2018.2875194
复制
发表时间：
2019-07-01
影响因子：
10.4
通讯作者：
Bharath, Anil Anthony
中科院分区：
文献类型：
--
作者：
Creswell, Antonia;Bharath, Anil Anthony
生成对抗网络（GANs）学习一种深度生成模型，该模型能够合成新颖的高维数据样本。通过将从选定的先验分布中抽取的潜在样本输入生成模型来合成新的数据样本。一旦经过训练，潜在空间会呈现出一些有趣的特性，这些特性可能对诸如分类或检索等下游任务有用。不幸的是，GANs没有提供一个“逆模型”，即从数据空间到潜在空间的映射，这使得推断给定数据样本的潜在表示变得困难。在本文中，我们引入一种技术——反转，使用预训练的GAN将数据样本（特别是图像）投影到潜在空间。利用我们提出的反转技术，我们能够确定一个经过训练的GAN能够对数据集的哪些属性进行建模，并根据重建损失量化GAN的性能。我们展示了我们提出的反转技术如何可用于定量比较在三个图像数据集上训练的各种GAN模型的性能。我们在网站（https://github.com/ToniCreswell/lnvertingGAN）上提供了所有实验的代码。
Generative adversarial networks (GANs) learn a deep generative model that is able to synthesize novel, high-dimensional data samples. New data samples are synthesized by passing latent samples, drawn from a chosen prior distribution, through the generative model. Once trained, the latent space exhibits interesting properties that may be useful for downstream tasks such as classification or retrieval. Unfortunately, GANs do not offer an "inverse model," a mapping from data space hack to latent space, making it difficult to infer a latent representation for a given data sample. In this paper, we introduce a technique, inversion, to project data samples, specifically images, to the latent space using a pretrained GAN. Using our proposed inversion technique, we are able to identify which attributes of a data set a trained GAN is able to model and quantify GAN performance, based on a reconstruction loss. We demonstrate how our proposed inversion technique may be used to quantitatively compare the performance of various GAN models trained on three image data sets. We provide codes for all of our experiments in the website (https://github.com/ToniCreswell/lnvertingGAN).