Causal Generative Domain Adaptation Networks

TL;DR

This paper introduces G-DAN and CG-DAN, generative models that explicitly capture distribution changes across domains using latent variables, improving domain adaptation and prediction efficiency.

Contribution

It proposes a novel generative domain adaptation framework with causal decomposition, enhancing learning efficiency and ability to generate data in new domains.

Findings

G-DAN effectively models feature distribution changes across domains.

CG-DAN improves statistical and computational efficiency through causal decomposition.

Both models demonstrate successful domain generation and cross-domain prediction on synthetic and real data.

Abstract

An essential problem in domain adaptation is to understand and make use of distribution changes across domains. For this purpose, we first propose a flexible Generative Domain Adaptation Network (G-DAN) with specific latent variables to capture changes in the generating process of features across domains. By explicitly modeling the changes, one can even generate data in new domains using the generating process with new values for the latent variables in G-DAN. In practice, the process to generate all features together may involve high-dimensional latent variables, requiring dealing with distributions in high dimensions and making it difficult to learn domain changes from few source domains. Interestingly, by further making use of the causal representation of joint distributions, we then decompose the joint distribution into separate modules, each of which involves different low-dimensional latent variables and can be learned separately, leading to a Causal G-DAN (CG-DAN). This improves both statistical and computational efficiency of the learning procedure. Finally, by matching the feature distribution in the target domain, we can recover the target-domain joint distribution and derive the learning machine for the target domain. We demonstrate the efficacy of both G-DAN and CG-DAN in domain generation and cross-domain prediction on both synthetic and real data experiments.