Abstract

We propose the use of Lexicalized Tree Adjoining Grammar (LTAG) as a source of features that are useful for reranking the output of a statistical parser. In this paper, we extend the notion of a tree kernel over arbitrary sub-trees of the parse to the derivation trees and derived trees provided by the LTAG formalism, and in addition, we extend the original definition of the tree kernel, making it more lexicalized and more compact. We use LTAG based features for the parse reranking task and obtain labeled recall and precision of on WSJ section 23 of Penn � Treebank for sentences of length 100 words. Our results show that the use of LTAG based tree kernel gives rise to a � relative difference in �-score im-provement over the use of a linear kernel without LTAG based features. 1

Citations

6696	The Nature of Statistical Learning Theory - Vapnik - 1995
780	Head-Driven Statistical Models for Natural Language Parsing - Collins - 1999
549	Tree-adjoining grammars - Joshi, Schabes - 1997
204	Convolution kernels for natural languages - Collins, Duffy - 2001
164	N: New Ranking Algorithms for Parsing and Tagging: Kernels over Discrete Structures, and the Voted Perceptron - Collins, Duffy
148	Towards History-based Grammars: Using Richer Models for Probabilistic Parsing - Black, Jelinek, et al. - 1993
65	Learning Kernel Classifiers: Theory and Algorithms - Herbrich - 2002
54	2000. Automated extraction of tags from the penn treebank - Chen, Vijay-Shanker
33	Statistical Parsing with an AutomaticallyExtracted Tree Adjoining Grammar - Chiang - 2000
25	An SVM Based Voting Algorithm with Application to Parse ReRanking - Shen, Joshi - 2003