Abstract

This work introduces the novel task of Source-free Multi-target Domain Adaptation and proposes adaptation framework comprising of Consistency with Nuclear-Norm Maximization and MixUp knowledge distillation CoNMix as a solution to this problem.

The main motive of this work is to solve for Single and Multi target Domain Adaptation (SMTDA) for the source-free paradigm, which enforces a constraint where the labeled source data is not available during target adaptation due to various privacy-related restrictions on data sharing. The source-free approach leverages target pseudo labels, which can be noisy to improve the target adaptation. We introduce consistency between label preserving augmentations and utilize pseudo label refinement methods to reduce noisy pseudo labels. Further, for better generalization of multiple target domains model using multiple source-free STDA models, we propose novel MixUp Knowledge Distillation (MKD).

We also show that Vision Transformer (VT) backbone gives better feature representation with improved domain transferability and class discriminability. Our proposed framework achieves the state-of-the-art (SOTA) results in various paradigms of source-free STDA and MTDA settings on popular domain adaptation datasets like Office-Home, Office-Caltech, and DomainNet.

Ablation Study

License: 

This project is licenced under an [MIT License].