Title

A Hassle-free Unsupervised Domain Adaptation Method using Instance Similarity Features

Publication Type

Conference Proceeding Article

Publication Date

7-2015

Abstract

We present a simple yet effective unsu-pervised domain adaptation method that can be generally applied for different NLP tasks. Our method uses unlabeled tar-get domain instances to induce a set of instance similarity features. These features are then combined with the origi-nal features to represent labeled source do-main instances. Using three NLP tasks, we show that our method consistently out-performs a few baselines, including SCL, an existing general unsupervised domain adaptation method widely used in NLP. More importantly, our method is very easy to implement and incurs much less com-putational cost than SCL.

Discipline

Computer Sciences | Databases and Information Systems

Research Areas

Data Management and Analytics

Publication

Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing: Beijing, China, July 26-31, 2015

Volume

2

First Page

168

Last Page

173

ISBN

9781941643730

Publisher

Association for Computational Linguistics

City or Country

Stroudsburg, PA

Additional URL

http://www.aclweb.org/anthology/P15-2028

This document is currently not available here.

Share

COinS