Bridging Theory and Algorithm for Domain Adaptation

TL;DR

This paper extends domain adaptation theories to multiclass classification, introduces Margin Disparity Discrepancy for better distribution comparison, and develops an adversarial algorithm that achieves state-of-the-art results.

Contribution

It introduces Margin Disparity Discrepancy and bridges the gap between theoretical domain adaptation bounds and practical adversarial algorithms.

Findings

Achieves state-of-the-art accuracy on domain adaptation benchmarks.

Provides rigorous generalization bounds for the proposed measurement.

Successfully connects theory with practical adversarial training methods.

Abstract

This paper addresses the problem of unsupervised domain adaption from theoretical and algorithmic perspectives. Existing domain adaptation theories naturally imply minimax optimization algorithms, which connect well with the domain adaptation methods based on adversarial learning. However, several disconnections still exist and form the gap between theory and algorithm. We extend previous theories (Mansour et al., 2009c; Ben-David et al., 2010) to multiclass classification in domain adaptation, where classifiers based on the scoring functions and margin loss are standard choices in algorithm design. We introduce Margin Disparity Discrepancy, a novel measurement with rigorous generalization bounds, tailored to the distribution comparison with the asymmetric margin loss, and to the minimax optimization for easier training. Our theory can be seamlessly transformed into an adversarial learning algorithm for domain adaptation, successfully bridging the gap between theory and algorithm. A series of empirical studies show that our algorithm achieves the state of the art accuracies on challenging domain adaptation tasks.