Source-Free Adaptation to Measurement Shift via Bottom-Up Feature Restoration

07/12/2021
by   Cian Eastwood, et al.
2

Source-free domain adaptation (SFDA) aims to adapt a model trained on labelled data in a source domain to unlabelled data in a target domain without access to the source-domain data during adaptation. Existing methods for SFDA leverage entropy-minimization techniques which: (i) apply only to classification; (ii) destroy model calibration; and (iii) rely on the source model achieving a good level of feature-space class-separation in the target domain. We address these issues for a particularly pervasive type of domain shift called measurement shift, characterized by a change in measurement system (e.g. a change in sensor or lighting). In the source domain, we store a lightweight and flexible approximation of the feature distribution under the source data. In the target domain, we adapt the feature-extractor such that the approximate feature distribution under the target data realigns with that saved on the source. We call this method Feature Restoration (FR) as it seeks to extract features with the same semantics from the target domain as were previously extracted from the source. We additionally propose Bottom-Up Feature Restoration (BUFR), a bottom-up training scheme for FR which boosts performance by preserving learnt structure in the later layers of a network. Through experiments we demonstrate that BUFR often outperforms existing SFDA methods in terms of accuracy, calibration, and data efficiency, while being less reliant on the performance of the source model in the target domain.

READ FULL TEXT

page 16

page 18

page 19

page 20

research
12/15/2022

Co-Learning with Pre-Trained Networks Improves Source-Free Domain Adaptation

Source-free domain adaptation aims to adapt a source model trained on fu...
research
11/21/2019

Improving Unsupervised Domain Adaptation with Variational Information Bottleneck

Domain adaptation aims to leverage the supervision signal of source doma...
research
11/28/2019

Light-weight Calibrator: a Separable Component for Unsupervised Domain Adaptation

Existing domain adaptation methods aim at learning features that can be ...
research
10/03/2018

A Generalized Neyman-Pearson Criterion for Optimal Domain Adaptation

In the problem domain adaptation for binary classification, the learner ...
research
03/10/2023

Knowledge Transfer via Multi-Head Feature Adaptation for Whole Slide Image Classification

Transferring prior knowledge from a source domain to the same or similar...
research
04/20/2023

Activity Classification Using Unsupervised Domain Transfer from Body Worn Sensors

Activity classification has become a vital feature of wearable health tr...
research
08/03/2023

Consistency Regularization for Generalizable Source-free Domain Adaptation

Source-free domain adaptation (SFDA) aims to adapt a well-trained source...

Please sign up or login with your details

Forgot password? Click here to reset