Generate to Adapt: Aligning Domains Using Generative Adversarial Networks

TLDR

Domain adaptation is a central problem in computer vision, distinct from methods that use adversarial frameworks for data generation and model retraining. This work proposes an unsupervised approach that aligns source and target distributions within a learned joint feature space. The method induces a symbiotic relationship between a learned embedding and a generative adversarial network, and is evaluated on digit classification, object recognition, and synthetic‑to‑real domain adaptation tasks. It achieves state‑of‑the‑art performance across these tasks, becoming the only GAN‑based method shown to work well on datasets such as OFFICE and DIGITS.

Abstract

Domain Adaptation is an actively researched problem in Computer Vision. In this work, we propose an approach that leverages unsupervised data to bring the source and target distributions closer in a learned joint feature space. We accomplish this by inducing a symbiotic relationship between the learned embedding and a generative adversarial network. This is in contrast to methods which use the adversarial framework for realistic data generation and retraining deep models with such data. We demonstrate the strength and generality of our approach by performing experiments on three different tasks with varying levels of difficulty: (1) Digit classification (MNIST, SVHN and USPS datasets) (2) Object recognition using OFFICE dataset and (3) Domain adaptation from synthetic to real data. Our method achieves state-of-the art performance in most experimental settings and by far the only GAN-based method that has been shown to work well across different datasets such as OFFICE and DIGITS.