Efficient language model adaptation for automatic speech recognition of spoken translations
- Author
- Joris Pelemans, Tom Vanallemeersch, Kris Demuynck (UGent) , Hugo Van hamme and Patrick Wambacq
- Organization
- Abstract
- Direct integration of translation model (TM) probabilities into a language model (LM) with the purpose of improving automatic speech recognition (ASR) of spoken translations typically requires a number of complex operations for each sentence. Many if not all of the LM probabilities need to be updated, the model needs to be renormalized and the ASR system needs to load a new, updated LM for each sentence. In computer-aided translation environments the time loss induced by these complex operations seriously reduces the potential of ASR as an efficient input method. In this paper we present a novel LM adaptation technique that drastically reduces the complexity of each of these operations. The technique consists of LM probability updates using exponential weights based on TM probabilities for each sentence and does not enforce probability renormalization. Instead of storing each resulting language model in its entirety, we only store the update weights which also reduces disk storage and loading time during ASR. Experiments on Dutch read speech translated from English show that both disk storage and recognition time drop dramatically compared to a baseline system that employs a more conventional way of updating the LM.
- Keywords
- efficient adaptation, machine translation, computer-aided translation, language models, speech recognition, INTEGRATION
Downloads
-
(...).pdf
- full text
- |
- UGent only
- |
- |
- 141.46 KB
Citation
Please use this url to cite or link to this publication: http://hdl.handle.net/1854/LU-7227884
- MLA
- Pelemans, Joris, et al. “Efficient Language Model Adaptation for Automatic Speech Recognition of Spoken Translations.” 16th Annual Conference of the International Speech Communication Association (INTERSPEECH 2015), Vols 1-5, International Speech Communication Association (ISCA), 2015, pp. 2262–66.
- APA
- Pelemans, J., Vanallemeersch, T., Demuynck, K., Van hamme, H., & Wambacq, P. (2015). Efficient language model adaptation for automatic speech recognition of spoken translations. 16th Annual Conference of the International Speech Communication Association (INTERSPEECH 2015), Vols 1-5, 2262–2266. Baixas, France: International Speech Communication Association (ISCA).
- Chicago author-date
- Pelemans, Joris, Tom Vanallemeersch, Kris Demuynck, Hugo Van hamme, and Patrick Wambacq. 2015. “Efficient Language Model Adaptation for Automatic Speech Recognition of Spoken Translations.” In 16th Annual Conference of the International Speech Communication Association (INTERSPEECH 2015), Vols 1-5, 2262–66. Baixas, France: International Speech Communication Association (ISCA).
- Chicago author-date (all authors)
- Pelemans, Joris, Tom Vanallemeersch, Kris Demuynck, Hugo Van hamme, and Patrick Wambacq. 2015. “Efficient Language Model Adaptation for Automatic Speech Recognition of Spoken Translations.” In 16th Annual Conference of the International Speech Communication Association (INTERSPEECH 2015), Vols 1-5, 2262–2266. Baixas, France: International Speech Communication Association (ISCA).
- Vancouver
- 1.Pelemans J, Vanallemeersch T, Demuynck K, Van hamme H, Wambacq P. Efficient language model adaptation for automatic speech recognition of spoken translations. In: 16th Annual conference of the International Speech Communication Association (INTERSPEECH 2015), vols 1-5. Baixas, France: International Speech Communication Association (ISCA); 2015. p. 2262–6.
- IEEE
- [1]J. Pelemans, T. Vanallemeersch, K. Demuynck, H. Van hamme, and P. Wambacq, “Efficient language model adaptation for automatic speech recognition of spoken translations,” in 16th Annual conference of the International Speech Communication Association (INTERSPEECH 2015), vols 1-5, Dresden, Germany, 2015, pp. 2262–2266.
@inproceedings{7227884,
abstract = {{Direct integration of translation model (TM) probabilities into a language model (LM) with the purpose of improving automatic speech recognition (ASR) of spoken translations typically requires a number of complex operations for each sentence. Many if not all of the LM probabilities need to be updated, the model needs to be renormalized and the ASR system needs to load a new, updated LM for each sentence. In computer-aided translation environments the time loss induced by these complex operations seriously reduces the potential of ASR as an efficient input method.
In this paper we present a novel LM adaptation technique that drastically reduces the complexity of each of these operations. The technique consists of LM probability updates using exponential weights based on TM probabilities for each sentence and does not enforce probability renormalization. Instead of storing each resulting language model in its entirety, we only store the update weights which also reduces disk storage and loading time during ASR. Experiments on Dutch read speech translated from English show that both disk storage and recognition time drop dramatically compared to a baseline system that employs a more conventional way of updating the LM.}},
author = {{Pelemans, Joris and Vanallemeersch, Tom and Demuynck, Kris and Van hamme, Hugo and Wambacq, Patrick}},
booktitle = {{16th Annual conference of the International Speech Communication Association (INTERSPEECH 2015), vols 1-5}},
isbn = {{9781510817906}},
issn = {{2308-457X}},
keywords = {{efficient adaptation,machine translation,computer-aided translation,language models,speech recognition,INTEGRATION}},
language = {{eng}},
location = {{Dresden, Germany}},
pages = {{2262--2266}},
publisher = {{International Speech Communication Association (ISCA)}},
title = {{Efficient language model adaptation for automatic speech recognition of spoken translations}},
year = {{2015}},
}