Domain2Vec: Domain Embedding for Unsupervised Domain Adaptation

TL;DR

This paper introduces Domain2Vec, a novel method for representing and understanding relationships between numerous visual domains, enhancing unsupervised domain adaptation by predicting domain similarities.

Contribution

The paper proposes Domain2Vec, a new model that learns vector representations of domains using feature disentanglement and Gram matrices, and introduces large-scale benchmarks for domain adaptation.

Findings

Domain2Vec accurately predicts domain similarities.

The benchmarks facilitate evaluation of multi-source domain adaptation methods.

Domain2Vec improves understanding of domain relationships.

Abstract

Conventional unsupervised domain adaptation (UDA) studies the knowledge transfer between a limited number of domains. This neglects the more practical scenario where data are distributed in numerous different domains in the real world. The domain similarity between those domains is critical for domain adaptation performance. To describe and learn relations between different domains, we propose a novel Domain2Vec model to provide vectorial representations of visual domains based on joint learning of feature disentanglement and Gram matrix. To evaluate the effectiveness of our Domain2Vec model, we create two large-scale cross-domain benchmarks. The first one is TinyDA, which contains 54 domains and about one million MNIST-style images. The second benchmark is DomainBank, which is collected from 56 existing vision datasets. We demonstrate that our embedding is capable of predicting domain similarities that match our intuition about visual relations between different domains. Extensive experiments are conducted to demonstrate the power of our new datasets in benchmarking state-of-the-art multi-source domain adaptation methods, as well as the advantage of our proposed model.