Paper: Simple Semi-Supervised Training of Part-Of-Speech Taggers

ACL ID P10-2038
Title Simple Semi-Supervised Training of Part-Of-Speech Taggers
Venue Annual Meeting of the Association of Computational Linguistics
Session Short Paper
Year 2010
Authors

Most attempts to train part-of-speech tag- gers on a mixture of labeled and unlabeled data have failed. In this work stacked learning is used to reduce tagging to a classification task. This simplifies semi- supervised training considerably. Our prefered semi-supervised method com- bines tri-training (Li and Zhou, 2005) and disagreement-based co-training. On the Wall Street Journal, we obtain an error re- duction of 4.2% with SVMTool (Gimenez and Marquez, 2004).