Project Details

Rozpoznávání řeči pro jazyky s omezeným množstvím trénovacích zdrojů

Project Period: 1. 1. 2012 - 31. 12. 2014

Project Type: grant

Code: GPP202/12/P604

Agency: Czech Science Foundation

Program: Postdoktorandské granty

English title
Speech recognition for low-resource languages
Type
grant
Keywords

speech recognition, multilingual speech recognition, automatic dictionary generation, SGMM

Abstract

The project aims at speech recognition in situations, where little training data and limited or no knowledge of liguistic and phonetics of the target language are available. In the domain of acoustic models, we will investigate modern techniques of GMM/HMM parameter representations in sub-spaces. The project will also focus on unsupervised creation of pronunciation dictionaries where sequences of phonemes will be replaced by graphemes or clusters of acoustic unites trained on data. Important parts of the project are tests on standard data and participation in international evaluations.

Team members
Karafiát Martin, Ing., Ph.D. (UPGM FIT VUT) , research leader
Janda Miloš, Ing. (UPGM FIT VUT) , team leader
Publications

2014

  • GRÉZL František, KARAFIÁT Martin and VESELÝ Karel. Adaptation of Multilingual Stacked Bottle-neck Neural Network Structure for New Language. In: Proceedings of ICASSP 2014. Florencie: IEEE Signal Processing Society, 2014, pp. 7704-7708. ISBN 978-1-4799-2892-7. Detail
  • GRÉZL František and KARAFIÁT Martin. Adapting Multilingual Neural Network Hierarchy to a New Language. In: Proceedings of the 4th International Workshop on Spoken Language Technologies for Under- resourced Languages SLTU-2014. – St. Petersburg, Russia, 2014. St. Petersburg: International Speech Communication Association, 2014, pp. 39-45. ISBN 978-5-8088-0908-6. Detail
  • KARAFIÁT Martin, GRÉZL František, VESELÝ Karel, HANNEMANN Mirko, SZŐKE Igor and ČERNOCKÝ Jan. BUT 2014 Babel System: Analysis of adaptation in NN based systems. In: Proceedings of Interspeech 2014. Singapore: International Speech Communication Association, 2014, pp. 3002-3006. ISBN 978-1-63439-435-2. Detail
  • KARAFIÁT Martin, VESELÝ Karel, SZŐKE Igor, BURGET Lukáš, GRÉZL František, HANNEMANN Mirko and ČERNOCKÝ Jan. BUT ASR System for BABEL Surprise Evaluation 2014. In: Proceedings of 2014 Spoken Language Technology Workshop. South Lake Tahoe, Nevada: IEEE Signal Processing Society, 2014, pp. 501-506. ISBN 978-1-4799-7129-9. Detail
  • KARAFIÁT Martin, GRÉZL František, HANNEMANN Mirko and ČERNOCKÝ Jan. BUT Neural Network Features for Spontaneous Vietnamese in BABEL. In: Proceedings of ICASSP 2014. Florencie: IEEE Signal Processing Society, 2014, pp. 5659-5663. ISBN 978-1-4799-2892-7. Detail
  • GRÉZL František and KARAFIÁT Martin. Combination of Multilingual and Semi-Supervised Training for Under-Resourced Languages. In: Proceedings of Interspeech 2014. Singapore: International Speech Communication Association, 2014, pp. 820-824. ISBN 978-1-63439-435-2. Detail
  • GRÉZL František, EGOROVA Ekaterina and KARAFIÁT Martin. Further Investigation into Multilingual Training and Adaptation of Stacked Bottle-neck Neural Network Structure. In: Proceedings of 2014 Spoken Language Technology Workshop. South Lake Tahoe, Nevada: IEEE Signal Processing Society, 2014, pp. 48-53. ISBN 978-1-4799-7129-9. Detail
  • NG Tim, HSIAO Roger, ZHANG Le, KARAKOS Damianos, MALLIDI Sri Harish, KARAFIÁT Martin, VESELÝ Karel, SZŐKE Igor, ZHANG Bing, NGUYEN Long and SCHWARTZ Richard. Progress in the BBN Keyword Search System for the DARPA RATS Program. In: Proceedings of Interspeech 2014. Singapore: International Speech Communication Association, 2014, pp. 959-963. ISBN 978-1-63439-435-2. Detail

2013

  • KARAFIÁT Martin, GRÉZL František, HANNEMANN Mirko, VESELÝ Karel and ČERNOCKÝ Jan. BUT BABEL System for Spontaneous Cantonese. In: Proceedings of Interspeech 2013. Lyon: International Speech Communication Association, 2013, pp. 2589-2593. ISBN 978-1-62993-443-3. ISSN 2308-457X. Detail
  • MOTLÍČEK Petr, POVEY Daniel and KARAFIÁT Martin. Feature And Score Level Combination Of Subspace Gaussians In LVCSR Task. In: Proceedings of ICASSP 2013. Vancouver: IEEE Signal Processing Society, 2013, pp. 7604-7608. ISBN 978-1-4799-0355-9. Detail
  • EGOROVA Ekaterina, VESELÝ Karel, KARAFIÁT Martin, JANDA Miloš and ČERNOCKÝ Jan. Manual and Semi-Automatic Approaches to Building a Multilingual Phoneme Set. In: Proceedings of ICASSP 2013. Vancouver: IEEE Signal Processing Society, 2013, pp. 7324-7328. ISBN 978-1-4799-0355-9. Detail
  • GRÉZL František and KARAFIÁT Martin. Semi-Supervised Bootstrapping Approach For Neural Network Feature Extractor Training. In: Proceedings of ASRU 2013. Olomouc: IEEE Signal Processing Society, 2013, pp. 470-475. ISBN 978-1-4799-2755-5. Detail

2012

  • KOMBRINK Stefan, MIKOLOV Tomáš, KARAFIÁT Martin and BURGET Lukáš. Improving Language Models for ASR Using Translated In-domain Data. In: Proceedings of 2012 IEEE International Conference on Acoustics, Speech and Signal Processing. Kyoto: IEEE Signal Processing Society, 2012, pp. 4405-4408. ISBN 978-1-4673-0044-5. Detail
  • KARAFIÁT Martin, JANDA Miloš, ČERNOCKÝ Jan and BURGET Lukáš. Region Dependent Linear Transforms in Multilingual Speech Recognition. In: Proc. International Conference on Acoustics, Speech, and Signal Processing 2012. Kyoto: IEEE Signal Processing Society, 2012, pp. 4885-4888. ISBN 978-1-4673-0044-5. Detail
  • VESELÝ Karel, KARAFIÁT Martin, GRÉZL František, JANDA Miloš and EGOROVA Ekaterina. The Language-Independent Bottleneck Features. In: Proceedings of IEEE 2012 Workshop on Spoken Language Technology. Miami: IEEE Signal Processing Society, 2012, pp. 336-341. ISBN 978-1-4673-5124-9. Detail
Back to top