Paper: Cross Lingual Adaptation: An Experiment on Sentiment Classifications

ACL ID P10-2048
Title Cross Lingual Adaptation: An Experiment on Sentiment Classifications
Venue Annual Meeting of the Association of Computational Linguistics
Session Short Paper
Year 2010
Authors

In this paper, we study the problem of using an annotated corpus in English for the same natural language processing task in another language. While various ma- chinetranslationsystemsareavailable, au- tomated translation is still far from per- fect. To minimize the noise introduced by translations, we propose to use only key ‘reliable” parts from the translations and apply structural correspondence learn- ing (SCL) to find a low dimensional rep- resentation shared by the two languages. We perform experiments on an English- Chinese sentiment classification task and compare our results with a previous co- training approach. To alleviate the prob- lem of data sparseness, we create ex- tra pseudo-examples for SCL by making queries to a search engine. Experiments on real-world on-line revi...