Minimal Dependency Length in Realization Ranking

Comprehension and corpus studies have found that the tendency to minimize dependency length has a strong influence on constituent ordering choices. In this paper, we investigate dependency length minimization in the context of discriminative realization ranking, focusing on its potential to eliminate egregious ordering errors as well as better match the distributional characteristics of sentence orderings in news text. We find that with a state-of-the-art, comprehensive realization ranking model, dependency length minimization yields statistically significant improvements in BLEU scores and significantly reduces the number of heavy/light ordering errors. Through distributional analyses, we also show that with simpler ranking models, dependency length minimization can go overboard, too often sacrificing canonical word order to shorten dependencies, while richer models manage to better counterbalance the dependency length minimization preference against (sometimes) competing canonical word order preferences.

[1]  Karin Harbusch,et al.  Generating Natural Word Orders in a Semi?free Word Order Language: Treebank-Based Linearization Preferences for German , 2004, CICLing.

[2]  Michael Strube,et al.  Tree Linearization in English: Improving Language Model Based Approaches , 2009, NAACL.

[3]  Michael White,et al.  Efficient Realization of Coordinate Structures in Combinatory Categorial Grammar , 2006 .

[4]  R. Harald Baayen,et al.  Predicting the dative alternation , 2007 .

[5]  Michael Speriosu,et al.  The role of prosody in the English dative alternation , 2010 .

[6]  Salim Roukos,et al.  Bleu: a Method for Automatic Evaluation of Machine Translation , 2002, ACL.

[7]  Frank Keller,et al.  Data from eye-tracking corpora as evidence for theories of syntactic processing complexity , 2008, Cognition.

[8]  Richard L. Lewis,et al.  An Activation-Based Model of Sentence Processing as Skilled Memory Retrieval , 2005, Cogn. Sci..

[9]  Benoît Favre,et al.  StuMaBa : From Deep Representation to Surface , 2011, ENLG.

[10]  Michael Gamon,et al.  Linguistically Informed Statistical Models of Constituent Structure for Ordering in Sentence Realization , 2004, COLING.

[11]  Charles Callaway.  The Types and Distributions of Errors in a Wide Coverage Surface Realizer Evaluation , 2005, ENLG.

[12]  Michael White,et al.  Perceptron Reranking for CCG Realization , 2009, EMNLP.

[13]  Holly P. Branigan,et al.  Contributions of animacy to grammatical function assignment and word order during production , 2008 .

[14]  John A. Hawkins,et al.  The relative order of prepositional phrases in English: Going beyond Manner–Place–Time , 1999, Language Variation and Change.

[15]  Julia Hockenmaier,et al.  Data and models for statistical parsing with combinatory categorial grammar , 2003 .

[16]  Benoit Favre,et al.  from deep representation to surface , 2011 .

[17]  Josef van Genabith,et al.  Exploiting Multi-Word Units in History-Based Probabilistic Generation , 2007, EMNLP-CoNLL.

[18]  James R. Curran,et al.  Wide-Coverage Efficient Statistical Parsing with CCG and Log-Linear Models , 2007, Computational Linguistics.

[19]  David Temperley,et al.  Minimization of dependency length in written English , 2007, Cognition.

[20]  Martin Kay,et al.  Syntactic Process , 1979, ACL.

[21]  John A. Hawkins,et al.  Why are categories adjacent , 2001 .

[22]  E. Gibson.  The dependency locality theory: A distance-based theory of linguistic complexity. , 2000 .

[23]  Stephan Oepen,et al.  Maximum Entropy Models for Realization Ranking , 2005 .

[24]  Irene Langkilde-Geary,et al.  An Empirical Verification of Coverage and Correctness for a General-Purpose Sentence Generator , 2002, INLG.

[25]  Daniel Gildea,et al.  Optimizing Grammars for Minimum Dependency Length , 2007, ACL.

[26]  F. Chang,et al.  “Long before short” preference in the production of a head-final language , 2001, Cognition.

[27]  Philipp Koehn,et al.  Statistical Significance Tests for Machine Translation Evaluation , 2004, EMNLP.

[28]  Thomas Wasov,et al.  Postverbal behavior , 2002, CSLI lecture notes series.

[29]  Michael White,et al.  Exploiting Named Entity Classes in CCG Surface Realization , 2009, HLT-NAACL.

[30]  Michael White,et al.  Designing Agreement Features for Realization Ranking , 2010, COLING.

[31]  Thomas Wasow,et al.  Post-verbal constituent ordering in English , 2003 .

[32]  Aoife Cahill,et al.  Incorporating Information Status into Generation Ranking , 2009, ACL/IJCNLP.

[33]  Annie Zaenen,et al.  Animacy and Syntactic Structure : Fronted NPs in English , 2007 .

[34]  Jennifer E. Arnold,et al.  HEAVINESS VS . , 2000 .

[35]  Jennifer E. Arnold,et al.  Heaviness vs. newness: The effects of structural complexity and discourse status on constituent ordering , 2015 .

[36]  E. Gibson.  Linguistic complexity: locality of syntactic dependencies , 1998, Cognition.

[37]  Christian Rohrer,et al.  DESIGNING FEATURES FOR PARSE DISAMBIGUATION AND REALISATION RANKING , 2007 .

[38]  Anja Belz,et al.  The First Surface Realisation Shared Task: Overview and Evaluation Results , 2011, ENLG.

[39]  Mark Steedman,et al.  CCGbank: A Corpus of CCG Derivations and Dependency Structures Extracted from the Penn Treebank , 2007, CL.

[40]  Jun'ichi Tsujii,et al.  Probabilistic Models for Disambiguation of an HPSG-Based Chart Generator , 2005, IWPT.

[41]  Richard L. Lewis,et al.  Computational principles of working memory in sentence comprehension , 2006, Trends in Cognitive Sciences.

[42]  John A. Hawkins,et al.  A Performance Theory of Order and Constituency , 1995 .

[43]  Jason Baldridge,et al.  Coupling CCG and Hybrid Logic Dependency Semantics , 2002, ACL.

[44]  Michael Strube,et al.  Generating Constituent Order in German Clauses , 2007, ACL.

[45]  Josef van Genabith,et al.  Dependency-Based N-Gram Models for General Purpose Sentence Realisation , 2008, COLING.