Evolutionary TBL template generation

Evolutionary TBL template generation

Transformation Based Learning (TBL) is a Machine Learning technique frequently used in some Natural Language Processing (NLP) tasks. TBL uses rule templates to identify error-correcting patterns. A critical requirement in TBL is the availability of a problem domain expert to build these rule templates. In this work, we propose an evolutionary approach based on Genetic Algorithms to automatically implement the template generation process. Additionally, we report our findings on five experiments with useful NLP tasks. We observe that our approach provides template sets with a mean loss of performance of 0.5% when compared to human built templates

Saved in:
Bibliographic Details
Main Authors: Milidiú,Ruy Luiz, Duarte,Julio Cesar, Santos,Cícero Nogueira dos
Format: Digital revista
Language:English
Published: Sociedade Brasileira de Computação 2007
Online Access:http://old.scielo.br/scielo.php?script=sci_arttext&pid=S0104-65002007000400004
Tags: Add Tag
No Tags, Be the first to tag this record!
id oai:scielo:S0104-65002007000400004
record_format ojs
spelling oai:scielo:S0104-650020070004000042010-05-24Evolutionary TBL template generationMilidiú,Ruy LuizDuarte,Julio CesarSantos,Cícero Nogueira dos Machine Learning Genetic Algorithms Transformation Error-Driven Based Learning Transformation Based Learning (TBL) is a Machine Learning technique frequently used in some Natural Language Processing (NLP) tasks. TBL uses rule templates to identify error-correcting patterns. A critical requirement in TBL is the availability of a problem domain expert to build these rule templates. In this work, we propose an evolutionary approach based on Genetic Algorithms to automatically implement the template generation process. Additionally, we report our findings on five experiments with useful NLP tasks. We observe that our approach provides template sets with a mean loss of performance of 0.5% when compared to human built templatesinfo:eu-repo/semantics/openAccessSociedade Brasileira de ComputaçãoJournal of the Brazilian Computer Society v.13 n.4 20072007-12-01info:eu-repo/semantics/articletext/htmlhttp://old.scielo.br/scielo.php?script=sci_arttext&pid=S0104-65002007000400004en10.1007/BF03194255
institution SCIELO
collection OJS
country Brasil
countrycode BR
component Revista
access En linea
databasecode rev-scielo-br
tag revista
region America del Sur
libraryname SciELO
language English
format Digital
author Milidiú,Ruy Luiz
Duarte,Julio Cesar
Santos,Cícero Nogueira dos
spellingShingle Milidiú,Ruy Luiz
Duarte,Julio Cesar
Santos,Cícero Nogueira dos
Evolutionary TBL template generation
author_facet Milidiú,Ruy Luiz
Duarte,Julio Cesar
Santos,Cícero Nogueira dos
author_sort Milidiú,Ruy Luiz
title Evolutionary TBL template generation
title_short Evolutionary TBL template generation
title_full Evolutionary TBL template generation
title_fullStr Evolutionary TBL template generation
title_full_unstemmed Evolutionary TBL template generation
title_sort evolutionary tbl template generation
description Transformation Based Learning (TBL) is a Machine Learning technique frequently used in some Natural Language Processing (NLP) tasks. TBL uses rule templates to identify error-correcting patterns. A critical requirement in TBL is the availability of a problem domain expert to build these rule templates. In this work, we propose an evolutionary approach based on Genetic Algorithms to automatically implement the template generation process. Additionally, we report our findings on five experiments with useful NLP tasks. We observe that our approach provides template sets with a mean loss of performance of 0.5% when compared to human built templates
publisher Sociedade Brasileira de Computação
publishDate 2007
url http://old.scielo.br/scielo.php?script=sci_arttext&pid=S0104-65002007000400004
work_keys_str_mv AT milidiuruyluiz evolutionarytbltemplategeneration
AT duartejuliocesar evolutionarytbltemplategeneration
AT santosciceronogueirados evolutionarytbltemplategeneration
_version_ 1756411106312585216

Similar Items