Skip to main navigation Skip to search Skip to main content

Maximum likelihood estimation of feature-based distributions

  • University of Delaware

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

4 Scopus citations

Abstract

Motivated by recent work in phonotactic learning (Hayes and Wilson 2008, Albright 2009), this paper shows how to define feature-based probability distributions whose parameters can be provably efficiently estimated. The main idea is that these distributions are defined as a product of simpler distributions (cf. Ghahramani and Jordan 1997). One advantage of this framework is it draws attention to what is minimally necessary to describe and learn phonological feature interactions in phonotactic patterns. The “bottom-up” approach adopted here is contrasted with the “top-down” approach in Hayes and Wilson (2008), and it is argued that the bottom-up approach is more analytically transparent.

Original languageEnglish
Title of host publicationSIGMORPHON 2010 - 11th Meeting of the ACL Special Interest Group on Computational Morphology and Phonology at the 48th Annual Meeting of the Association for Computational Linguistics, ACL 2010
EditorsJeffrey Heinz, Lynne Cahill, Richard Wicentowski
PublisherAssociation for Computational Linguistics (ACL)
Pages28-37
Number of pages10
ISBN (Electronic)1932432760, 9781932432763
StatePublished - 2010
Event11th Meeting of the ACL Special Interest Group on Computational Morphology and Phonology, SIGMORPHON 2010 at the 48th Annual Meeting of the Association for Computational Linguistics, ACL 2010 - Uppsala, Sweden
Duration: Jul 15 2010 → …

Publication series

NameProceedings of the Annual Meeting of the Association for Computational Linguistics
ISSN (Print)0736-587X

Conference

Conference11th Meeting of the ACL Special Interest Group on Computational Morphology and Phonology, SIGMORPHON 2010 at the 48th Annual Meeting of the Association for Computational Linguistics, ACL 2010
Country/TerritorySweden
CityUppsala
Period07/15/10 → …

Fingerprint

Dive into the research topics of 'Maximum likelihood estimation of feature-based distributions'. Together they form a unique fingerprint.

Cite this