Zero-shot Domain Adaptation Based on Attribute Information

Masato Ishii, Takashi Takenouchi, Masashi Sugiyama
Proceedings of The Eleventh Asian Conference on Machine Learning, PMLR 101:473-488, 2019.

Abstract

In this paper, we propose a novel domain adaptation method that can be applied without target data. We consider the situation where domain shift is caused by a prior change of a specific factor and assume that we know how the prior changes between source and target domains. We call this factor an attribute, and reformulate the domain adaptation problem to utilize the attribute prior instead of target data. In our method, the source data are reweighted with the sample-wise weight estimated by the attribute prior and the data themselves so that they are useful in the target domain. We theoretically reveal that our method provides more precise estimation of sample-wise transferability than a straightforward attribute-based reweighting approach. Experimental results with both toy datasets and benchmark datasets show that our method can perform well, though it does not use any target data.

Cite this Paper

BibTeX
@InProceedings{pmlr-v101-ishii19a, title = {Zero-shot Domain Adaptation Based on Attribute Information}, author = {Ishii, Masato and Takenouchi, Takashi and Sugiyama, Masashi}, booktitle = {Proceedings of The Eleventh Asian Conference on Machine Learning}, pages = {473--488}, year = {2019}, editor = {Lee, Wee Sun and Suzuki, Taiji}, volume = {101}, series = {Proceedings of Machine Learning Research}, month = {17--19 Nov}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v101/ishii19a/ishii19a.pdf}, url = {https://proceedings.mlr.press/v101/ishii19a.html}, abstract = {In this paper, we propose a novel domain adaptation method that can be applied without target data. We consider the situation where domain shift is caused by a prior change of a specific factor and assume that we know how the prior changes between source and target domains. We call this factor an attribute, and reformulate the domain adaptation problem to utilize the attribute prior instead of target data. In our method, the source data are reweighted with the sample-wise weight estimated by the attribute prior and the data themselves so that they are useful in the target domain. We theoretically reveal that our method provides more precise estimation of sample-wise transferability than a straightforward attribute-based reweighting approach. Experimental results with both toy datasets and benchmark datasets show that our method can perform well, though it does not use any target data.} }
Endnote
%0 Conference Paper %T Zero-shot Domain Adaptation Based on Attribute Information %A Masato Ishii %A Takashi Takenouchi %A Masashi Sugiyama %B Proceedings of The Eleventh Asian Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2019 %E Wee Sun Lee %E Taiji Suzuki %F pmlr-v101-ishii19a %I PMLR %P 473--488 %U https://proceedings.mlr.press/v101/ishii19a.html %V 101 %X In this paper, we propose a novel domain adaptation method that can be applied without target data. We consider the situation where domain shift is caused by a prior change of a specific factor and assume that we know how the prior changes between source and target domains. We call this factor an attribute, and reformulate the domain adaptation problem to utilize the attribute prior instead of target data. In our method, the source data are reweighted with the sample-wise weight estimated by the attribute prior and the data themselves so that they are useful in the target domain. We theoretically reveal that our method provides more precise estimation of sample-wise transferability than a straightforward attribute-based reweighting approach. Experimental results with both toy datasets and benchmark datasets show that our method can perform well, though it does not use any target data.
APA
Ishii, M., Takenouchi, T. & Sugiyama, M.. (2019). Zero-shot Domain Adaptation Based on Attribute Information. Proceedings of The Eleventh Asian Conference on Machine Learning, in Proceedings of Machine Learning Research 101:473-488 Available from https://proceedings.mlr.press/v101/ishii19a.html.

Related Material