Learning causal representations for robust domain adaptation

TL;DR

This paper introduces a Causal AutoEncoder that learns causal representations from source domain data alone, enabling robust domain adaptation without target domain data during training.

Contribution

It proposes a novel CAE model combining autoencoder and causal structure learning to extract causal features for domain adaptation without target data.

Findings

CAE outperforms eleven state-of-the-art methods on real-world datasets.

Causal representations improve robustness in domain adaptation.

Model effectively separates causal and task-irrelevant features.

Abstract

Domain adaptation solves the learning problem in a target domain by leveraging the knowledge in a relevant source domain. While remarkable advances have been made, almost all existing domain adaptation methods heavily require large amounts of unlabeled target domain data for learning domain invariant representations to achieve good generalizability on the target domain. In fact, in many real-world applications, target domain data may not always be available. In this paper, we study the cases where at the training phase the target domain data is unavailable and only well-labeled source domain data is available, called robust domain adaptation. To tackle this problem, under the assumption that causal relationships between features and the class variable are robust across domains, we propose a novel Causal AutoEncoder (CAE), which integrates deep autoencoder and causal structure learning into a unified model to learn causal representations only using data from a single source domain. Specifically, a deep autoencoder model is adopted to learn low-dimensional representations, and a causal structure learning model is designed to separate the low-dimensional representations into two groups: causal representations and task-irrelevant representations. Using three real-world datasets the extensive experiments have validated the effectiveness of CAE compared to eleven state-of-the-art methods.