Unveiling the Superior Paradigm: A Comparative Study of Source-Free Domain Adaptation and Unsupervised Domain Adaptation

TL;DR

This paper compares source-free and unsupervised domain adaptation, demonstrating SFDA's advantages in efficiency, robustness, and negative transfer mitigation, and introduces a novel method to improve multi-source SFDA in data-sharing scenarios.

Contribution

The study provides a comprehensive comparison of UDA and SFDA, and proposes a new weight estimation method to enhance multi-source SFDA performance in practical data-sharing contexts.

Findings

SFDA generally outperforms UDA in real-world scenarios.

SFDA reduces negative transfer and overfitting risks.

The proposed method improves multi-source SFDA effectiveness.

Abstract

In domain adaptation, there are two popular paradigms: Unsupervised Domain Adaptation (UDA), which aligns distributions using source data, and Source-Free Domain Adaptation (SFDA), which leverages pre-trained source models without accessing source data. Evaluating the superiority of UDA versus SFDA is an open and timely question with significant implications for deploying adaptive algorithms in practical applications. In this study, we demonstrate through predictive coding theory and extensive experiments on multiple benchmark datasets that SFDA generally outperforms UDA in real-world scenarios. Specifically, SFDA offers advantages in time efficiency, storage requirements, targeted learning objectives, reduced risk of negative transfer, and increased robustness against overfitting. Notably, SFDA is particularly effective in mitigating negative transfer when there are substantial distribution discrepancies between source and target domains. Additionally, we introduce a novel data-model fusion scenario, where data sharing among stakeholders varies (e.g., some provide raw data while others provide only models), and reveal that traditional UDA and SFDA methods do not fully exploit their potential in this context. To address this limitation and capitalize on the strengths of SFDA, we propose a novel weight estimation method that effectively integrates available source data into multi-SFDA (MSFDA) approaches, thereby enhancing model performance within this scenario. This work provides a thorough analysis of UDA versus SFDA and advances a practical approach to model adaptation across diverse real-world environments.