Coupled Generative Adversarial Networks

Abstract

We propose coupled generative adversarial network (CoGAN) for learning a joint distribution of multi-domain images. In contrast to the existing approaches, which require tuples of corresponding images in different domains in the training set, CoGAN can learn a joint distribution without any tuple of corresponding images. It can learn a joint distribution with just samples drawn from the marginal distributions. This is achieved by enforcing a weight-sharing constraint that limits the network capacity and favors a joint distribution solution over a product of marginal distributions one. We apply CoGAN to several joint distribution learning tasks, including learning a joint distribution of color and depth images, and learning a joint distribution of face images with different attributes. For each task it successfully learns the joint distribution without any tuple of corresponding images. We also demonstrate its applications to domain adaptation and image transformation.

To be published in NIPS 2016

References in corpus (7)

• Conditional Generative Adversarial Nets
• Deep Domain Confusion: Maximizing for Domain Invariance
• Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks
• Unifying Visual-Semantic Embeddings with Multimodal Neural Language Models
• Beyond Sharing Weights for Deep Domain Adaptation
• Generative Moment Matching Networks
• Attribute2Image: Conditional Image Generation from Visual Attributes

Cited by in corpus (15)

• One-Shot Unsupervised Cross Domain Translation
• Discovering state-parameter mappings in subsurface models using generative adversarial networks
• Unsupervised Pixel-Level Domain Adaptation with Generative Adversarial Networks
• An Unsupervised Domain Adaptation Model based on Dual-module Adversarial Training
• Adversarial Message Passing For Graphical Models
• Message Passing Multi-Agent GANs
• MixerGAN: An MLP-Based Architecture for Unpaired Image-to-Image Translation
• Galaxy Image Translation with Semi-supervised Noise-reconstructed Generative Adversarial Networks
• Adversarial learning for product recommendation
• DW-GAN: A Discrete Wavelet Transform GAN for NonHomogeneous Dehazing
• SDA-GAN: Unsupervised Image Translation Using Spectral Domain Attention-Guided Generative Adversarial Network
• Holdout-Based Fidelity and Privacy Assessment of Mixed-Type Synthetic Data
• Learning Invariant Representation with Consistency and Diversity for Semi-supervised Source Hypothesis Transfer
• Unsupervised Many-to-Many Image-to-Image Translation Across Multiple Domains
• Synthesizing Photorealistic Images with Deep Generative Learning