ScholarWorks@UMassAmherst (University of Massachusetts Amherst) · 2012 · 26 citations · 10 references
Open access
Language StudiesComputer ScienceFrenchComputational LinguisticsMachine LearningPhonologyLanguage ModelsOptimality TheoryPhonological AlternationsPhonological RepresentationRepresentation LearningVowel RepresentationAlternation RulesLexically Conditioned AlternationsObjective Function FormulationOptimality Theory In PhonologyPhonological GeneralizationPhonological LearningSemi-supervised Representation LearningVowel Pattern Alternations
We show that a class of cases that has been previously studied in terms of learning of abstract phonological underlying representations (URs) can be handled by a learner that chooses URs from a contextually conditioned distribution over observed surface representations. We implement such a learner in a Maximum Entropy version of Optimality Theory, in which UR learning is an instance of semisupervised learning. Our objective function incorporates a term aimed to ensure generalization, independently required for phonotactic learning in Optimality Theory, and does not have a bias for single URs for morphemes. This learner is successful on a test language provided by Tesar (2006) as a challenge for UR learning. We also provide successful results on learning of a toy case modeled on French vowel alternations, which have also URs. This case includes lexically conditioned variation, an aspect of the data that cannot be handled by abstract URs, showing that in this respect our approach is more general. 1
10
On Information and Sufficiency
S. Kullback, R. A. Leibler · The Annals of Mathematical Statistics · 1951 · 20K citations · Full text
Solutions of Ill-Posed Problems.
John B. Bell, V. Y. Arsenin · Mathematics of Computation · 1978 · 8.2K citations
A Limited Memory Algorithm for Bound Constrained Optimization
Richard H. Byrd, Peihuang Lu, Jorge Nocedal et al. · SIAM Journal on Scientific Computing · 1995 · 6.2K citations
Natural and unnatural constraints in Hungarian vowel harmony
Bruce Hayes, Kie Zuraw, Péter Siptár et al. · Language · 2009 · 394 citations · Full text
Learning OT constraint rankings using a maximum entropy model
Sharon Goldwater, Mark Johnson · 2003 · 275 citations