论文信息 - A Bayesian Model of Natural Language Phonology: Generating Alternations from Underlying Forms

A Bayesian Model of Natural Language Phonology: Generating Alternations from Underlying Forms

A stochastic approach to learning phonology. The model presented captures 7--15% more phonologically plausible underlying forms than a simple majority solution, because it prefers "pure" alternations. It could be useful in cases where an approximate solution is needed, or as a seed for more complex models. A similar process could be involved in some stages of child language acquisition; in particular, early learning of phonotactics.

David Ellis

[1] Zoubin Ghahramani,et al. MCMC for Doubly-intractable Distributions , 2006, UAI.

[2] Gertjan van Noord,et al. Approximation and Exactness in Finite State Optimality Theory , 2000, ACL 2000.

[3] Nasser M. Nasrabadi,et al. Pattern Recognition and Machine Learning , 2006, Technometrics.

[4] Mark Johnson,et al. Learning OT constraint rankings using a maximum entropy model , 2003 .

[5] Daniel Gildea,et al. Learning Bias and Phonological-Rule Induction , 1996, CL.

[6] Mikko Kurimo,et al. Overview of Morpho Challenge in CLEF 2007 , 2007, CLEF.

[7] Christopher M. Bishop,et al. Pattern Recognition and Machine Learning (Information Science and Statistics) , 2006 .

[8] Lauri Karttunen,et al. The Proper Treatment of Optimality in Computational Phonology , 1998, ArXiv.

[9] K. Demuth,et al. Revisiting the acquisition of Sesotho noun class prefixes , 2009 .

[10] Mark Johnson,et al. Priors in Bayesian Learning of Phonological Rules , 2004, SIGMORPHON@ACL.

[11] Thomas L. Griffiths,et al. Adaptor Grammars: A Framework for Specifying Compositional Nonparametric Bayesian Models , 2006, NIPS.

[12] Lauri Karttunen,et al. Finite-State Non-Concatenative Morphotactics , 2000, ACL.

[13] Donald Geman,et al. Stochastic Relaxation, Gibbs Distributions, and the Bayesian Restoration of Images , 1984, IEEE Transactions on Pattern Analysis and Machine Intelligence.