Decision Boundary Optimization-Informed Domain Adaptation

TL;DR

This paper introduces a novel Decision Boundary optimization-informed MMD (DB-MMD) that enhances domain adaptation by jointly optimizing distribution alignment and decision boundaries, leading to improved performance across multiple datasets.

Contribution

It proposes a new MMD variant that incorporates decision boundary information, enabling more effective domain adaptation with theoretical guarantees.

Findings

DB-MMD improves baseline DA methods significantly, up to 9.5 margin.

Embedding DB-MMD into existing methods enhances their effectiveness.

The approach is validated across 8 standard DA datasets.

Abstract

Maximum Mean Discrepancy (MMD) is widely used in a number of domain adaptation (DA) methods and shows its effectiveness in aligning data distributions across domains. However, in previous DA research, MMD-based DA methods focus mostly on distribution alignment, and ignore to optimize the decision boundary for classification-aware DA, thereby falling short in reducing the DA upper error bound. In this paper, we propose a strengthened MMD measurement, namely, Decision Boundary optimization-informed MMD (DB-MMD), which enables MMD to carefully take into account the decision boundaries, thereby simultaneously optimizing the distribution alignment and cross-domain classifier within a hybrid framework, and leading to a theoretical bound guided DA. We further seamlessly embed the proposed DB-MMD measurement into several popular DA methods, e.g., MEDA, DGA-DA, to demonstrate its effectiveness w.r.t different experimental settings. We carry out comprehensive experiments using 8 standard DA datasets. The experimental results show that the DB-MMD enforced DA methods improve their baseline models using plain vanilla MMD, with a margin that can be as high as 9.5.