Abstract

We present a linguistically-motivated algorithm for reconstructing nonlocal dependency in broad-coverage context-free parse trees derived from treebanks. We use an algorithm based on loglinear classifiers to augment and reshape context-free trees so as to reintroduce underlying nonlocal dependencies lost in the context-free approximation. We find that our algorithm compares favorably with prior work on English using an existing evaluation metric, and also introduce and argue for a new dependency-based evaluation metric. By this new evaluation metric our algorithm achieves 60 % error reduction on gold-standard input trees and 5 % error reduction on state-ofthe-art machine-parsed input trees, when compared with the best previous work. We also present the first results on nonlocal dependency reconstruction for a language other than English, comparing performance on English and German. Our new evaluation metric quantitatively corroborates the intuition that in a language with freer word order, the surface dependencies in context-free parse trees are a poorer approximation to underlying dependency structure. 1

Citations

780	Head-Driven Statistical Models for Natural Language Parsing - Collins - 1999
671	A Maximum-Entropy-Inspired Parser - Charniak - 2000
466	Automatic Labeling of Semantic Roles - Gildea, Jurafsky - 2002
422	Accurate unlexicalized parsing - Klein, Manning - 2003
397	Generalized phrase structure grammar - Gazdar - 1985
255	Three models for the description of language - Chomsky - 1956
95	Parsing the Wall Street Journal Using a Lexical-Functional Grammar and Discriminative Estimation Techniques - Riezler - 2002
86	The Interface between Phrasal and Functional Constraints - Maxwell, Kaplan - 1993
60	Generative models for statistical parsing with Combinatory Categorial Grammar - Hockenmaier, Steedman - 2002
43	High Performance question/answering - Pasca, Harabagiu - 2001
39	Speed and accuracy in shallow and deep stochastic parsing - Kaplan, Riezler, et al. - 2004
15	A simple pattern-matching algorithm for recovering empty nodes and their antecedents - Johnson - 2001
12	Antecedent recovery: Experiments with a trace tagger - Dienes, Dienes, et al. - 2003
10	Parsing german with sister-head dependencies - Dubey, Keller - 2003
9	A theory of non-constituent coordination based on finite-state rules - Maxwell, Manning - 1996
4	Computing the most probable parse for a discontinuous phrase structure grammar - Plaehn - 2004
3	Statistical Parsing with Non-local Dependencies - Dienes - 2003
2	Deep processing by combining shallow methods - Dienes, Dubey - 2003
2	Annotating unrestricted German text. In Fachtagung der Sektion Computerlinguistik der Deutschen Gesellschaft fr Sprachwissenschaft - Skut, Brants, et al. - 1997
1	Learning linearization rules from treebanks - Kruijff - 2002
1	Probabilistic Models of Syntactic Discontinuity - Levy - 2004