Adversarial Discriminative Domain Adaptation

TL;DR

This paper introduces ADDA, a novel adversarial discriminative domain adaptation method that effectively handles large domain shifts, outperforming previous approaches in digit and object classification tasks.

Contribution

The paper proposes a new generalized framework for adversarial domain adaptation and introduces ADDA, combining discriminative modeling, untied weights, and GAN loss for improved performance.

Findings

ADDA surpasses state-of-the-art in digit classification tasks.

ADDA achieves better results on cross-modality object classification.

The proposed method is simpler and more effective than existing approaches.

Abstract

Adversarial learning methods are a promising approach to training robust deep networks, and can generate complex samples across diverse domains. They also can improve recognition despite the presence of domain shift or dataset bias: several adversarial approaches to unsupervised domain adaptation have recently been introduced, which reduce the difference between the training and test domain distributions and thus improve generalization performance. Prior generative approaches show compelling visualizations, but are not optimal on discriminative tasks and can be limited to smaller shifts. Prior discriminative approaches could handle larger domain shifts, but imposed tied weights on the model and did not exploit a GAN-based loss. We first outline a novel generalized framework for adversarial adaptation, which subsumes recent state-of-the-art approaches as special cases, and we use this generalized view to better relate the prior approaches. We propose a previously unexplored instance of our general framework which combines discriminative modeling, untied weight sharing, and a GAN loss, which we call Adversarial Discriminative Domain Adaptation (ADDA). We show that ADDA is more effective yet considerably simpler than competing domain-adversarial methods, and demonstrate the promise of our approach by exceeding state-of-the-art unsupervised adaptation results on standard cross-domain digit classification tasks and a new more difficult cross-modality object classification task.