ChainGAN: A sequential approach to GANs

Abstract

We propose a new architecture and training methodology for generative adversarial networks. Current approaches attempt to learn the transformation from a noise sample to a generated data sample in one shot. Our proposed generator architecture, called $\text{ChainGAN}$, uses a two-step process. It first attempts to transform a noise vector into a crude sample, similar to a traditional generator. Next, a chain of networks, called $\text{editors}$, attempt to sequentially enhance this sample. We train each of these units independently, instead of with end-to-end backpropagation on the entire chain. Our model is robust, efficient, and flexible as we can apply it to various network architectures. We provide rationale for our choices and experimentally evaluate our model, achieving competitive results on several datasets.

References in corpus (11)

• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• Deep Residual Learning for Image Recognition
• InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets
• BEGAN: Boundary Equilibrium Generative Adversarial Networks
• C-RNN-GAN: Continuous recurrent neural networks with adversarial training
• A Note on the Inception Score
• Generalization and Equilibrium in Generative Adversarial Nets (GANs)
• Adversarial Feature Matching for Text Generation
• Multi-Generator Generative Adversarial Nets
• Learning Deep ResNet Blocks Sequentially using Boosting Theory
• Aesthetic-Driven Image Enhancement by Adversarial Learning

Cited by in corpus (1)

• Adaptive Divergence for Rapid Adversarial Optimization