Semi-Supervised Domain Adaptation via Minimax Entropy

TLDR

Domain adaptation methods align source and target feature distributions without target supervision. The study proposes a Minimax Entropy approach to address semi‑supervised domain adaptation. The method employs a feature encoder and prototype‑based classifier, alternating entropy maximization on unlabeled target data with entropy minimization on the encoder. The approach outperforms baselines and achieves state‑of‑the‑art performance in SSDA. Code is available at http://cs-people.bu.edu/keisaito/research/MME.html.

Abstract

Contemporary domain adaptation methods are very effective at aligning feature distributions of source and target domains without any target supervision. However, we show that these techniques perform poorly when even a few labeled examples are available in the target domain. To address this semi-supervised domain adaptation (SSDA) setting, we propose a novel Minimax Entropy (MME) approach that adversarially optimizes an adaptive few-shot model. Our base model consists of a feature encoding network, followed by a classification layer that computes the features' similarity to estimated prototypes (representatives of each class). Adaptation is achieved by alternately maximizing the conditional entropy of unlabeled target data with respect to the classifier and minimizing it with respect to the feature encoder. We empirically demonstrate the superiority of our method over many baselines, including conventional feature alignment and few-shot methods, setting a new state of the art for SSDA. Our code is available at http://cs-people.bu.edu/keisaito/research/MME.html.