Abstract

tens @ scs.lecd s.ac.uk Common wisdom has it that tile bias of stochastic grammars in favor of shorter deriwttions of a sentence is hamfful and should be redressed. We show that the common wisdom is wrong for stochastic grammars that use elementary trees instead o1 ' conlext-l'ree rules, such as Stochastic Tree-Substitution Grammars used by Data-Oriented Parsing models. For such grammars a non-probabilistic metric based on tile shortest derivation outperforms a probabilistic metric on the ATIS and OVIS corpora, while it obtains competitive results on the Wall Street Journal (WSJ) corpus. This paper also contains the first publislmd experiments with DOP on the WSJ. 1.

Citations

5971	A mathematical theory of communication - Shannon - 1948
2102	Building a large annotated corpus of English: the Penn treebank - Marcus, Santorini, et al. - 1993
955	Head-driven statistical models for natural language parsing - Collins - 1999
821	A maximum-entropy-inspired parser - Charniak - 2000
503	Three generative, lexicalised models for statistical parsing - Collins - 1997
373	The estimation of stochastic context-free grammars using the Inside-Outside algorithm. Computer Speech and Language - Lari, Young - 1990
366	Statistical parsing with a context-free grammar and word statistics - Charniak - 1997
231	A procedure for quantitatively comparing the syntactic coverage of English grammars - Black, Abney, et al. - 1991
230	Tree-bank grammars - Charniak - 1996
193	Generalised probabilistic LR parsing of natural language (corpora) with unification-based grammars - Briscoe, Carroll - 1993
165	Learning to parse natural language with maximum entropy models. Machine Learning 34, 151–175. Parsing 41 - Ratnaparkhi - 1999
145	Beyond grammar: an experience-based theory of language - Bod - 1998
135	Coping with ambiguity and unknown words through probabilistic models - Weischedel, Meteer, et al. - 1993
100	Computational complexity of probabilistic disambiguation by means of tree grammar - Sima’an - 1996
82	Parsing Inside-Out - Goodman - 1998
71	New figures of merit for best-first probabilistic chart parsing - Caraballo, Charniak - 1998
59	A probabilistic corpus-driven model for lexical-functional analysis - Bod, Kaplan - 1998
57	Pearl: A Probabilistic Chart Parser - Magerman, Marcus - 1991
48	Using an Annotated Corpus as a Stochastic Grammar - Bod - 1993
48	Bilexical grammars and a cubic-time probabilistic parser - Eisner - 1997
41	Statistical parsing of messages - Chitrao, Grishman - 1990
37	A DOP model for semantic interpretation - Bonnema, Bod, et al. - 1997
33	An optimized algorithm for Data Oriented Parsing - Sima'an - 1995
25	A CorpusBased Approach to Semantic Interpretation - Berg, Bod, et al. - 1994
22	Automatic extraction of stochastic lexicalized tree grammars from treebanks - Neumann - 1998
18	Learning Stochastic Lexicalized Tree Grammars from HPSG", DFKI - Neumann, Flickinger - 1999
14	A Hybrid Architecture for Robust MT using LFG-DOP - Way - 1999
8	An Empirical Evaluation of LFG-DOP - Bod
5	A New Probability Model for Data-Oriented Parsing - Bonnema, Buying, et al. - 1999
5	Extending DOP1 with the insertion operation - Hoogweg - 2000
2	The DOP Estimation Method is Biased and Inconsistent", squib - Johnson - 1998
1	l)ata-Oriented Parsing (I)OP - Bod - 1992
1	Using an Annotated Corpus as a Stochastic Grammar", l'roceedings European ChcqJter q/" the ACL'93 - Bod - 1993
1	Two Questions about l)ata Oriented Parsing - Bod - 1996
1	Beyond Grammar: An EaT,erience-l~ased Theot3, of Language - Bod - 1998
1	An Empirical Evaluation of LI~G-I)OP - Bod
1	A Probal)ilistic Corpusl)rivcn Model lkw l,cxical Ftmctional Analysis", l't'oceedings COIJNf;-ACI,'9 - Bed - 1998
1	A New I'robalfilily Model for I)ata-()ricntcd Parsing", lb-oceedings o.f the Amsterdam Colloqttittm - Bonncma, Buying, et al. - 1999
1	Generalized I'robabilistic IA~. I'arsing of Natural l,anguage (Corpora) with Uni ficalion-I~ased Gramnlars - Briscoe, Carroll - 1993
1	Monte Carlo Sampling Ior Nl'-hard Maximization I'roblems in the l"ramework of Wc ghtcd Parsing", iu Natmcd l.altguag<! l)roce.ssing - 2000, l.eclure Not,:!.,; i,q Arttl/h:ial lnlelligence 1835 - Chappclicr, l, et al. - 2000
1	Statistical Parsing wilh a Contcxl-l:rcc Grammar and Word Statistics", l'roceedinw AAAI-97. Menlo Park. Ca - Charniak - 1997
1	Three generalive lexicaliscd models tbr statistical parsing", l'roceedittgs EACI/ACL'97 - Collins - 1997
1	llead-l)riven Statistical Models./br Natural I,anguagc I'arsing - Collins - 1999
1	Bilexical Granlnlars and a Cubic-Time Probabilistic Parser - Eisner - 1997
1	Palwing lnsMe-Out, Ph.l). thesis, l larwu'd - Goodman - 1998
1	Extending DOPI with the Insertion Operation - loogweg - 2000
1	The I)OP Estimatkm Method is Biased and Inconsistent", squib - Johnson - 1998
1	The Estimation of Stochastic Context-Free Granunars Using the lnside-()utsidc Algorithm", ()mtl~ltlel",~'1~(3cc11 alld l,angttctge, d - ari, K, et al. - 1990
1	Pearl: A Probabilislic Chart Parser - Marcus - 1991
1	lhtikling a l'u-e - Marcus, Santorini, et al. - 1993