Towards automatic transcription of large spoken archives in agglutinating languages - Hungarian ASR for the MALACH project

Authors:
Péter Mihajlik;Tibor Fegyó;Bottyán Németh;Zoltán Tüske;Viktor Trón
Affiliations:
Department of Telecommunications and Media Informatics, Budapest University of Technology and Economics, Budapest, Hungary;Department of Telecommunications and Media Informatics, Budapest University of Technology and Economics, Budapest, Hungary and AITIA International, Inc., Budapest, Hungary;Department of Telecommunications and Media Informatics, Budapest University of Technology and Economics, Budapest, Hungary and AITIA International, Inc., Budapest, Hungary;Department of Telecommunications and Media Informatics, Budapest University of Technology and Economics, Budapest, Hungary;University of Edinburgh, Edinburgh, United Kingdom
Venue:
TSD'07 Proceedings of the 10th international conference on Text, speech and dialogue
Year:
2007

Citing 3
Cited 1

Korean large vocabulary continuous speech recognition with morpheme-based recognition units

Speech Communication
Unlimited vocabulary speech recognition for agglutinative languages

HLT-NAACL '06 Proceedings of the main conference on Human Language Technology Conference of the North American Chapter of the Association of Computational Linguistics
Hunmorph: open source word analysis

Software '05 Proceedings of the Workshop on Software

Improved recognition of spontaneous Hungarian speech: morphological and acoustic modeling techniques for a less resourced task

IEEE Transactions on Audio, Speech, and Language Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

The paper describes automatic speech recognition experiments and results on the spontaneous Hungarian MALACH speech corpus. A novel morph-based lexical modeling approach is compared to the traditional word-based one and to another, previously best performing morph-based one in terms of word and letter error rates. The applied language and acoustic modeling techniques are also detailed. Using unsupervised speaker adaptations along with morph based lexical models 14.4%-8.1% absolute word error rate reductions have been achieved on a 2 speakers, 2 hours test set as compared to the speaker independent baseline results.