Abstract

We formulate the problem of nonprojective dependency parsing as a polynomial-sized integer linear program. Our formulation is able to handle non-local output features in an efficient manner; not only is it compatible with prior knowledge encoded as hard constraints, it can also learn soft constraints from data. In particular, our model is able to learn correlations among neighboring arcs (siblings and grandparents), word valency, and tendencies toward nearlyprojective parses. The model parameters are learned in a max-margin framework by employing a linear programming relaxation. We evaluate the performance of our parser on data in several natural languages, achieving improvements over existing state-of-the-art methods. 1

Citations

2333	Convex Analysis - Rockafellar - 1970
371	Combinatorial Optimization: Polyhedra and Efficiency - Schrijver - 2003
363	Markov logic networks - Richardson, Domingos - 2006
200	Three New Probabilistic Models for Dependency Parsing: An Exploration - Eisner - 1996
181	Online passiveaggressive algorithms - Crammer, Dekel, et al. - 2006
177	2005b. Non-projective dependency parsing using spanning tree algorithms - McDonald, Pereira, et al.
161	CoNLL-X shared task on multilingual dependency parsing - Buchholz, Marsi - 2006
136	Optimum Branchings - Edmonds - 1967
132	Dependency tree kernels for relation extraction - Culotta, Soresen
111	2006. Online learning of approximate dependency parsing algorithms - McDonald, Pereira
76	On the shortest arborescence of a directed graph - Chu, Liu - 1965
74	Efficent Parsing for Bilexical Context-Free Grammars and Head Automaton Grammars - Eisner, Satta - 1999
72	Pseudo-boolean optimization - Boros, Hammer
62	D.: Semantic role labeling via integer linear programming inference - Punyakanok, Roth, et al. - 2004
57	Integer Linear Programming Inference for Conditional Random Fields - Roth, Yih - 2005
51	Machine translation using probabilistic synchronous dependency insertion grammars - Ding, Palmer - 2005
48	Joint Determination of Anaphoricity and Coreference Resolution Using Integer Programming - Denis, Baldridge - 2007
47	Dependency parsing by belief propagation - Smith, Eisner - 2008
41	Global inference for sentence compression: An integer linear programming approach - Clarke, Lapata - 2008
37	Experiments with a higher-order projective dependency parser - Carreras - 2007
36	Integrating graphbased and transition-based dependency parsers - Nivre, McDonald - 2008
34	Pseudoprojectivity: A polynomially parsable non-projective dependency grammar - Kahane, Nasr, et al. - 1998
33	Optimal trees - Magnanti, Wolsey - 1994
30	Word alignment via quadratic assignment - Lacoste-Julien, Taskar, et al. - 2006
30	Incremental integer linear programming for non-projective dependency parsing - Riedel, Clarke - 2006
29	The CoNLL-2008 shared task on joint parsing of syntactic and semantic dependencies - Surdeanu, Johansson, et al. - 2008
27	Stacking dependency parsers - Martins, Das, et al. - 2008
22	On the complexity of non-projective data-driven dependency parsing - McDonald, Satta - 2007
15	Finding Optimum Branchings. Networks - Tarjan - 1977
11	Polyhedral outer approximations with application to natural language parsing - Martins, Smith, et al. - 2009
9	Constraints as prior knowledge - Chang, Ratinov, et al. - 2008