Abstract:S ource- F ree D omain A daptation (SFDA) task aims to transfer knowledge from a labeled source domain to a label-scarce target domain, in which the source data can not be accessed but only a pre-trained source model and unlabeled target data are available during adaptation. Previous methods for source model adaptation rely on hypothesis transfer learning that trains the feature extractor to learn target features aligned to the distribution of source features while freezing the source classifier. However, reusing only the source classifier without exploring the comprehensive knowledge of the source model can lead to biased feature alignment. To this end, we propose a novel method called Transformer-bAsed thorouGh Source HypOthesis Transfer (TagSHOT) framework to effectively unleash the thorough knowledge potential of pre-trained source hypothesis. Specifically, our approach delves into the correlation coefficient among CLS/patch tokens across different Transformer layers, uncovering the concealed insights within the pre-trained source model and constructing a comprehensive source hypothesis. By tailoring the target feature alignment to this thorough source hypothesis, our model facilitates the adaptation of a broader range of classification-related knowledge to the target domain. Furthermore, we introduce a Salient Token Extension (STE) module, designed to capture the target-specific discriminative information by propagating the salient information among tokens. This mechanism enriches our model's ability to understand and incorporate target-specific nuances. Extensive experiments have been conducted to validate the effectiveness of our method, which outperforms state-of-the-art approaches by a large margin.

Source-Free Domain Adaptation Via Target Prediction Distribution Searching

Target-Discriminability-Induced Multi-Source-Free Domain Adaptation

Source-Free Domain Adaptation Via Distribution Estimation

Generation, Augmentation, and Alignment: A Pseudo-Source Domain Based Method for Source-Free Domain Adaptation.

A Framework for Self-Supervised Federated Domain Adaptation

Transformer-Based Multi-Source Domain Adaptation Without Source Data.

Attracting and Dispersing: A Simple Approach for Source-free Domain Adaptation

Transformer-Based Source-Free Domain Adaptation

Exploiting the Intrinsic Neighborhood Structure for Source-free Domain Adaptation

SF(DA)$^2$: Source-free Domain Adaptation Through the Lens of Data Augmentation

Unleashing Knowledge Potential of Source Hypothesis for Source-Free Domain Adaptation

A Comprehensive Survey on Source-Free Domain Adaptation

OneRing: A Simple Method for Source-free Open-partial Domain Adaptation

SIDE: Self-supervised Intermediate Domain Exploration for Source-free Domain Adaptation

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

Hierarchical Unsupervised Relation Distillation for Source Free Domain Adaptation

Source-free Domain Adaptation Requires Penalized Diversity

BMD: A General Class-balanced Multicentric Dynamic Prototype Strategy for Source-free Domain Adaptation

CNG-SFDA:Clean-and-Noisy Region Guided Online-Offline Source-Free Domain Adaptation

Learn from the Learnt: Source-Free Active Domain Adaptation via Contrastive Sampling and Visual Persistence

SFDA-T: A novel source-free domain adaptation method with strong generalization ability for fault diagnosis