HMM based Viterbi paths for rejection correction in a convolutional neural network classifier
Identifieur interne : 005F82 ( Main/Merge ); précédent : 005F81; suivant : 005F83HMM based Viterbi paths for rejection correction in a convolutional neural network classifier
Auteurs : Hubert Cecotti ; Szilard Vajda ; Abdel Belaïd [France]Source :
English descriptors
- KwdEn :
Abstract
In this paper a fast data driven learning-corpus building algorithm (FDDLCB) is proposed. The generic technique allows to build dynamically a representative and compact learning corpus for a connectionist model. The constructed dataset contains just a reduced number of patterns but sufficiently descriptive to characterize the different classes which should be separated. The method is based on a double least mean squares (LMS) error minimization mechanism trying to find the optimal boundaries of the different pattern classes. In the classical learning process the LMS is serving to minimize the error during the learning and this process is improved with a second one, as the new samples selection is also based on the idea to minimize the recognition error. Reinforcing the class boundaries where the recognition fails let us achieve a rapid and good generalization without any loss of occuracy. A modified version of the algorithm will be also presented. The experiments were performed on MNIST (Modified NIST) separated digit dataset. The encouraging result (98.51%) using just 1.85% of the available patterns from the original training dataset is comparable even with the state of the art techniques.
Links toward previous steps (curation, corpus...)
- to stream Crin, to step Corpus: 004162
- to stream Crin, to step Curation: 004162
- to stream Crin, to step Checkpoint: 000207
Links to Exploration step
CRIN:cecotti05cLe document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en" wicri:score="485">HMM based Viterbi paths for rejection correction in a convolutional neural network classifier</title></titleStmt><publicationStmt><idno type="RBID">CRIN:cecotti05c
</idno><date when="2005" year="2005">2005</date><idno type="wicri:Area/Crin/Corpus">004162</idno><idno type="wicri:Area/Crin/Curation">004162</idno><idno type="wicri:explorRef" wicri:stream="Crin" wicri:step="Curation">004162</idno><idno type="wicri:Area/Crin/Checkpoint">000207</idno><idno type="wicri:explorRef" wicri:stream="Crin" wicri:step="Checkpoint">000207</idno><idno type="wicri:Area/Main/Merge">005F82</idno></publicationStmt><sourceDesc><biblStruct><analytic><title xml:lang="en">HMM based Viterbi paths for rejection correction in a convolutional neural network classifier</title><author><name sortKey="Cecotti, Hubert" sort="Cecotti, Hubert" uniqKey="Cecotti H" first="Hubert" last="Cecotti">Hubert Cecotti</name></author><author><name sortKey="Vajda, Szilard" sort="Vajda, Szilard" uniqKey="Vajda S" first="Szilard" last="Vajda">Szilard Vajda</name></author><author><name sortKey="Belaid, Abdel" sort="Belaid, Abdel" uniqKey="Belaid A" first="Abdel" last="Belaïd">Abdel Belaïd</name><affiliation><country>France</country><placeName><settlement type="city">Nancy</settlement><region type="region" nuts="2">Grand Est</region><region type="region" nuts="2">Lorraine (région)</region></placeName><orgName type="laboratoire" n="5">Laboratoire lorrain de recherche en informatique et ses applications</orgName><orgName type="university">Université de Lorraine</orgName><orgName type="institution">Centre national de la recherche scientifique</orgName><orgName type="institution">Institut national de recherche en informatique et en automatique</orgName></affiliation></author></analytic></biblStruct></sourceDesc></fileDesc><profileDesc><textClass><keywords scheme="KwdEn" xml:lang="en"><term>classifier combination</term><term>handwritten digit recognition</term><term>hidden markov model</term><term>neural network</term><term>stochastic model</term></keywords></textClass></profileDesc></teiHeader><front><div type="abstract" xml:lang="en" wicri:score="2984">In this paper a fast data driven learning-corpus building algorithm (FDDLCB) is proposed. The generic technique allows to build dynamically a representative and compact learning corpus for a connectionist model. The constructed dataset contains just a reduced number of patterns but sufficiently descriptive to characterize the different classes which should be separated. The method is based on a double least mean squares (LMS) error minimization mechanism trying to find the optimal boundaries of the different pattern classes. In the classical learning process the LMS is serving to minimize the error during the learning and this process is improved with a second one, as the new samples selection is also based on the idea to minimize the recognition error. Reinforcing the class boundaries where the recognition fails let us achieve a rapid and good generalization without any loss of occuracy. A modified version of the algorithm will be also presented. The experiments were performed on MNIST (Modified NIST) separated digit dataset. The encouraging result (98.51%) using just 1.85% of the available patterns from the original training dataset is comparable even with the state of the art techniques.</div></front></TEI></record>Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Wicri/Lorraine/explor/InforLorV4/Data/Main/Merge
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 005F82 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Main/Merge/biblio.hfd -nk 005F82 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien
|wiki= Wicri/Lorraine
|area= InforLorV4
|flux= Main
|étape= Merge
|type= RBID
|clé= CRIN:cecotti05c
|texte= HMM based Viterbi paths for rejection correction in a convolutional neural network classifier
}}
| This area was generated with Dilib version V0.6.33. |  |