论文信息 - Estimators for Stochastic “Unification-Based” Grammars

Estimators for Stochastic “Unification-Based” Grammars

Log-linear models provide a statistically sound framework for Stochastic "Unification-Based" Grammars (SUBGs) and stochastic versions of other kinds of grammars. We describe two computationally-tractable ways of estimating the parameters of such grammars from a training corpus of syntactic analyses, and apply these to estimate a stochastic version of Lexical-Functional Grammar.

[1] W. J. Langford. Statistical Methods , 1959, Nature.

[2] J. Besag. Spatial Interaction and the Statistical Analysis of Lattice Systems , 1974 .

[3] J. Besag. Statistical Analysis of Non-Lattice Data , 1975 .

[4] F. A. Seiler,et al. Numerical Recipes in C: The Art of Scientific Computing , 1989 .

[5] Jerry R. Hobbs,et al. Two Principles of Parse Preference , 1990, COLING.

[6] A. Hasman,et al. Probabilistic reasoning in intelligent systems: Networks of plausible inference , 1991 .

[7] Ronald M. Kaplan,et al. A Method for Disjunctive Constraint Satisfaction , 1991 .

[8] Adam L. Berger,et al. A Maximum Entropy Approach to Natural Language Processing , 1996, CL.

[9] Frederick Jelinek,et al. Statistical methods for speech recognition , 1997 .

[10] Steven P. Abney. Stochastic Attribute-Value Grammars , 1996, CL.

[11] John T. Maxwell,et al. Formal issues in lexical-functional grammar , 1998 .

[12] Brendan J. Frey,et al. Graphical Models for Machine Learning and Digital Communication , 1998 .

[13] Stuart Gonan,et al. Probability models for complex systems , 1998 .

[14] Michael I. Jordan. Graphical Models , 2003 .