Balancing Discriminability and Transferability for Source-Free Domain Adaptation

06/16/2022
by   Jogendra Nath Kundu, et al.
3

Conventional domain adaptation (DA) techniques aim to improve domain transferability by learning domain-invariant representations; while concurrently preserving the task-discriminability knowledge gathered from the labeled source data. However, the requirement of simultaneous access to labeled source and unlabeled target renders them unsuitable for the challenging source-free DA setting. The trivial solution of realizing an effective original to generic domain mapping improves transferability but degrades task discriminability. Upon analyzing the hurdles from both theoretical and empirical standpoints, we derive novel insights to show that a mixup between original and corresponding translated generic samples enhances the discriminability-transferability trade-off while duly respecting the privacy-oriented source-free setting. A simple but effective realization of the proposed insights on top of the existing source-free DA approaches yields state-of-the-art performance with faster convergence. Beyond single-source, we also outperform multi-source prior-arts across both classification and semantic segmentation benchmarks.

READ FULL TEXT

page 2

page 5

page 19

research
07/27/2022

Concurrent Subsidiary Supervision for Unsupervised Source-Free Domain Adaptation

The prime challenge in unsupervised domain adaptation (DA) is to mitigat...
research
08/25/2021

Generalize then Adapt: Source-Free Domain Adaptive Semantic Segmentation

Unsupervised domain adaptation (DA) has gained substantial interest in s...
research
04/09/2020

Towards Inheritable Models for Open-Set Domain Adaptation

There has been a tremendous progress in Domain Adaptation (DA) for visua...
research
04/09/2020

Universal Source-Free Domain Adaptation

There is a strong incentive to develop versatile learning techniques tha...
research
08/27/2023

Domain-Specificity Inducing Transformers for Source-Free Domain Adaptation

Conventional Domain Adaptation (DA) methods aim to learn domain-invarian...
research
12/26/2019

Domain Adaptation Regularization for Spectral Pruning

Deep Neural Networks (DNNs) have recently been achieving state-of-the-ar...
research
02/09/2022

Amplitude Spectrum Transformation for Open Compound Domain Adaptive Semantic Segmentation

Open compound domain adaptation (OCDA) has emerged as a practical adapta...

Please sign up or login with your details

Forgot password? Click here to reset