Heuristic Hyper-minimization of Finite State Lexicons

Show full item record



Permalink

http://hdl.handle.net/10138/136266

Citation

Drobac , S , Linden , K , Pirinen , T & Silfverberg , M 2014 , Heuristic Hyper-minimization of Finite State Lexicons . in Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC'14) . vol. 9 , #784 , European Language Resources Association (ELRA) , Reykjavik, Iceland , Language Resource and Evaluation Conference , Reykjavik , Iceland , 26/05/2014 .

Title: Heuristic Hyper-minimization of Finite State Lexicons
Author: Drobac, Senka; Linden, Krister; Pirinen, Tommi; Silfverberg, Miikka
Contributor: University of Helsinki, Department of Modern Languages 2010-2017
University of Helsinki, Department of Modern Languages 2010-2017
University of Helsinki, Phonetics and Speech Synthesis
University of Helsinki, Department of Modern Languages 2010-2017
Publisher: European Language Resources Association (ELRA)
Date: 2014-05-26
Language: eng
Number of pages: 6
Belongs to series: Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC'14)
ISBN: 978-2-9517408-8-4
978-2-9517408-8-4
URI: http://hdl.handle.net/10138/136266
Abstract: Flag diacritics, which are special multi-character symbols executed at runtime, enable optimising finite-state networks by combining identical sub-graphs of its transition graph. Traditionally, the feature has required linguists to devise the optimisations to the graph by hand alongside the morphological description. In this paper, we present a novel method for discovering flag positions in morphological lexicons automatically, based on the morpheme structure implicit in the language description. With this approach, we have gained significant decrease in the size of finite-state networks while maintaining reasonable application speed. The algorithm can be applied to any language description, where the biggest achievements are expected in large and complex morphologies. The most noticeable reduction in size we got with a morphological transducer for Greenlandic, whose original size is on average about 15 times larger than other morphologies. With the presented hyper-minimization method, the transducer is reduced to 10,1% of the original size, with lookup speed decreased only by 9,5%.
Subject: 113 Computer and information sciences
finite-state transducers
hyper-minimization
6121 Languages
lexicon
Rights:


Files in this item

Total number of downloads: Loading...

Files Size Format View
Hyperminimization1.pdf 528.8Kb PDF View/Open

This item appears in the following Collection(s)

Show full item record