Universal Semi-supervised Model Adaptation via Collaborative Consistency Training

TL;DR

This paper presents a novel universal semi-supervised domain adaptation approach that leverages collaborative consistency training between models to effectively adapt to new domains with limited labeled data and differing label sets.

Contribution

The paper introduces the USMA problem and proposes a collaborative consistency training framework that combines source and target models for improved domain adaptation.

Findings

Effective on multiple benchmark datasets

Outperforms existing domain adaptation methods

Utilizes minimal labeled target data

Abstract

In this paper, we introduce a realistic and challenging domain adaptation problem called Universal Semi-supervised Model Adaptation (USMA), which i) requires only a pre-trained source model, ii) allows the source and target domain to have different label sets, i.e., they share a common label set and hold their own private label set, and iii) requires only a few labeled samples in each class of the target domain. To address USMA, we propose a collaborative consistency training framework that regularizes the prediction consistency between two models, i.e., a pre-trained source model and its variant pre-trained with target data only, and combines their complementary strengths to learn a more powerful model. The rationale of our framework stems from the observation that the source model performs better on common categories than the target-only model, while on target-private categories, the target-only model performs better. We also propose a two-perspective, i.e., sample-wise and class-wise, consistency regularization to improve the training. Experimental results demonstrate the effectiveness of our method on several benchmark datasets.