Statistical A* Dependency Parsing

Peter Dienes, Alexander Koller, and Marco Kuhlmann

Proceedings of the Workshop on Prospects and Advances in the Syntax/Semantics Interface, Nancy, 2003.

We propose a probability model for XDG dependency trees and an A* search control regime for the XDG parsing algorithm that guarantees the best parse to be found first. To the best of our knowledge, this is the first application of statistical guidance in the context of constraint programming for computational linguistics. Our probabilistic model is most closely related to Resnik's (1992) model for stochastic TAG.

Download: Download (39 K)

BibTex Entry

@InProceedings{DieKolKuh03, 
   author = {Peter Dienes and Alexander Koller and Marco Kuhlmann}, 
   title  = {Statistical A* Dependency Parsing}, 
   year   = 2003, 
   booktitle = {Proceedings of the Workshop on Prospects and 
             Advances in the Syntax/Semantics Interface}, 
   address = {Nancy} 
} 

Back: Publications