Surprisingly simple semi-supervised domain adaptation with pretraining and consistency
Files
Published version
Date
2021
DOI
Authors
Mishra, Samarth
Saenko, Kate
Saligrama, Venkatesh
Version
Published version
OA Version
Citation
S. Mishra, K. Saenko, V. Saligrama. 2021. "Surprisingly Simple Semi-Supervised Domain Adaptation with Pretraining and Consistency." https://arxiv.org/abs/2101.12727.
Abstract
Visual domain adaptation involves learning to classify
images from a target visual domain using labels available
in a different source domain. A range of prior work uses
adversarial domain alignment to try and learn a domain invariant
feature space, where a good source classifier can
perform well on target data. This however, can lead to errors
where class A features in the target domain get aligned
to class B features in source. We show that in the presence
of a few target labels, simple techniques like selfsupervision
(via rotation prediction) and consistency regularization
can be effective without any adversarial alignment
to learn a good target classifier. Our Pretraining and
Consistency (PAC) approach, can achieve state of the art
accuracy on this semi-supervised domain adaptation task,
surpassing multiple adversarial domain alignment methods,
across multiple datasets. Notably, it outperforms all
recent approaches by 3-5% on the large and challenging
DomainNet benchmark, showing the strength of these simple
techniques in fixing errors made by adversarial alignment