A time-delay neural network architecture for isolated word recognition /

A transaction-invariant back-propagation network is described that performs better than a sophisticated continuous acoustic parameter hiden Markov model on a noisy, 100-speaker confusable vocabulary isolated word recognition task. The network's replicated architecture permits it to extract prec...

Full description

Bibliographic Details
Main Authors: 262544 Lang, Kevin J., Waibel, Alex H, Hinton, Geoffrey E.
Format:
Subjects:
_version_ 1826366012378841088
author 262544 Lang, Kevin J.
Waibel, Alex H
Hinton, Geoffrey E.
author_facet 262544 Lang, Kevin J.
Waibel, Alex H
Hinton, Geoffrey E.
author_sort 262544 Lang, Kevin J.
collection OCEAN
description A transaction-invariant back-propagation network is described that performs better than a sophisticated continuous acoustic parameter hiden Markov model on a noisy, 100-speaker confusable vocabulary isolated word recognition task. The network's replicated architecture permits it to extract precise information from unaligned training patterns selected by a naive segmentation rule.
first_indexed 2024-03-04T15:13:26Z
format
id KOHA-OAI-TEST:54940
institution Universiti Teknologi Malaysia - OCEAN
last_indexed 2024-03-04T15:13:26Z
record_format dspace
spelling KOHA-OAI-TEST:549402020-12-19T16:58:53ZA time-delay neural network architecture for isolated word recognition / 262544 Lang, Kevin J. Waibel, Alex H Hinton, Geoffrey E. A transaction-invariant back-propagation network is described that performs better than a sophisticated continuous acoustic parameter hiden Markov model on a noisy, 100-speaker confusable vocabulary isolated word recognition task. The network's replicated architecture permits it to extract precise information from unaligned training patterns selected by a naive segmentation rule.A transaction-invariant back-propagation network is described that performs better than a sophisticated continuous acoustic parameter hiden Markov model on a noisy, 100-speaker confusable vocabulary isolated word recognition task. The network's replicated architecture permits it to extract precise information from unaligned training patterns selected by a naive segmentation rule.12PSZJBLNeural circuitryWord recognition
spellingShingle Neural circuitry
Word recognition
262544 Lang, Kevin J.
Waibel, Alex H
Hinton, Geoffrey E.
A time-delay neural network architecture for isolated word recognition /
title A time-delay neural network architecture for isolated word recognition /
title_full A time-delay neural network architecture for isolated word recognition /
title_fullStr A time-delay neural network architecture for isolated word recognition /
title_full_unstemmed A time-delay neural network architecture for isolated word recognition /
title_short A time-delay neural network architecture for isolated word recognition /
title_sort time delay neural network architecture for isolated word recognition
topic Neural circuitry
Word recognition
work_keys_str_mv AT 262544langkevinj atimedelayneuralnetworkarchitectureforisolatedwordrecognition
AT waibelalexh atimedelayneuralnetworkarchitectureforisolatedwordrecognition
AT hintongeoffreye atimedelayneuralnetworkarchitectureforisolatedwordrecognition
AT 262544langkevinj timedelayneuralnetworkarchitectureforisolatedwordrecognition
AT waibelalexh timedelayneuralnetworkarchitectureforisolatedwordrecognition
AT hintongeoffreye timedelayneuralnetworkarchitectureforisolatedwordrecognition