Paper: Clustering Words by Syntactic Similarity improves Dependency Parsing of Predicate-argument Structures

ACL ID W09-3829
Title Clustering Words by Syntactic Similarity improves Dependency Parsing of Predicate-argument Structures
Venue International Conference on Parsing Technologies
Session Main Conference
Year 2009
Authors

We present an aproach for deriving syntactic word clusters from parsed text, grouping words acording to their unlexicalized syntac- tic contexts. We then explore the use of these syntactic clusters in leveraging a large corpus of tres generated by a high-acuracy parser to improve the acuracy of another parser based on a diferent formalism for representing a dif- ferent level of sentence structure. In our ex- periments, we use phrase-structure tres to produce syntactic word clusters that are used by a predicate-argument dependency parser, significantly improving its acuracy.