Paper: Co-Parsing with Competitive Models

ACL ID W09-3816
Title Co-Parsing with Competitive Models
Venue International Conference on Parsing Technologies
Session Main Conference
Year 2009

We present an asymmetric approach to a run-time combination of two parsers where one component serves as a predic- tor to the other one. Predictions are inte- grated by means of weighted constraints and therefore are subject to preferential decisions. Previously, the same architec- ture has been successfully used with pre- dictors providing partial or inferior infor- mation about the parsing problem. It has now been applied to a situation where the predictor produces exactly the same type of information at a fully competitive qual- ity level. Results show that the combined system outperforms its individual compo- nents, even though their performance in isolation is already fairly high.