JointGAN: Multi-Domain Joint Distribution Learning with Generative Adversarial Nets

TL;DR

JointGAN introduces a novel generative adversarial network that learns the joint distribution of multiple domains, enabling flexible sampling of marginals, conditionals, and full joint distributions for multi-domain data analysis.

Contribution

It proposes a new multi-domain GAN framework that jointly learns the entire distribution across domains, unlike prior models focusing only on conditional distributions.

Findings

Able to generate samples from marginals, conditionals, and joint distributions

Effective for two and three domain data analysis

Demonstrates versatility in multi-domain generative modeling

Abstract

A new generative adversarial network is developed for joint distribution matching. Distinct from most existing approaches, that only learn conditional distributions, the proposed model aims to learn a joint distribution of multiple random variables (domains). This is achieved by learning to sample from conditional distributions between the domains, while simultaneously learning to sample from the marginals of each individual domain. The proposed framework consists of multiple generators and a single softmax-based critic, all jointly trained via adversarial learning. From a simple noise source, the proposed framework allows synthesis of draws from the marginals, conditional draws given observations from a subset of random variables, or complete draws from the full joint distribution. Most examples considered are for joint analysis of two domains, with examples for three domains also presented.