Abstract
This paper deals with robust estimations of data statistics used for the adaptation. The statistics are accumulated before the adaptation process from available adaptation data. In general, only small amount of adaptation data is assumed. These data are often corrupted by noise, channel, they do not contain only clean speech. Also, when training Hidden Markov Models (HMM) several assumptions are made that could not have been fulfilled in the praxis, etc. Therefore, we described several techniques that aim to make the adaptation as robust as possible in order to increase the accuracy of the adapted system. One of the methods consists in initialization of the adaptation statistics in order to prevent ill-conditioned transformation matrices. Another problem arises when an acoustic feature is assigned to an improper HMM state even if the reference transcription is available. Such situations can occur because of the forced alignment process used to align frames to states. Thus, it is quite handy to accumulate data statistic utilizing only reliable frames (in the sense of data likelihood). We are focusing on Maximum Likelihood Linear Transformations and the experiments were performed utilizing the feature Maximum Likelihood Linear Regression (fMLLR). Experiments are aimed to describe the behavior of the system extended by proposed methods.
This research was supported by the Ministry of Education of the Czech Republic, project No. MŠMT LC536, the Grant Agency of the Czech Republic, project No. GAČR 102/08/0707 and the grant of The University of West Bohemia, project No. SGS-2010-054.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Psutka, J., Šmídl, L., Pražák, A.: Searching for a Robust MFCC-Based Parameterization for ASR Application. In: SIGMAP, Lisabon, pp. 196–199 (2007) ISBN: 978-989-8111-13-5
Pražák, A., Zajíc, Z., Machlica, L., Psutka, J.V.: Fast Speaker Adaptation in Automatic Online Subtitling. In: SIGMAP, Italy, pp. 126–130 (2009)
Gauvain, L., Lee, C.H.: Maximum A-Posteriori Estimation for Multivariate Gaussian Mixture Observations of Markov Chains. IEEE Transactions SAP 2, 291–298 (1994)
Gales, M.J.F.: Maximum Likelihood Linear Transformation for HMM-based Speech Recognition. Tech. Report, CUED/FINFENG/TR291, Cambridge Univ. (1997)
Povey, D., Saon, G.: Feature and Model Space Speaker Adaptation with Full Covariance Gaussians. In: Interspeech, paper 2050-Tue2BuP.14 (2006)
Gales, M. J. F.: The Generation and Use of Regression Class Trees for MLLR Adaptation. Cambridge University Engineering Department (1996)
Yu, K.: Adaptive Training for Large Vocabulary Continuous Speech Recognition. Ph.D. thesis, Hughes Hall College and Cambridge University Engineering Department (2006)
Li, Y., et al.: Incremental On-line Feature Space MLLR Adaptation for Telephony Speech Recognition. In: International Conference on Spoken Language Processing, Denver (2002)
Byrne, W., Gunawardana, A.: Discounted Likelihood Linear Regression for Rapid Adaptation. In: Eurospeech, Budapest, pp. 203–206 (1999)
Pollak, P., et al.: SpeechDat(E) – Eastern European Telephone Speech Databases, XLDB – Very Large Telephone Speech Databases. In: European Language Recources Association (ELRA), Paris (2000)
Young, S., et al.: The HTK Book (for HTK Version 3.4). Cambridge University Engineering Department (2001-2006)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2010 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Zajíc, Z., Machlica, L., Müller, L. (2010). Robust Statistic Estimates for Adaptation in the Task of Speech Recognition . In: Sojka, P., Horák, A., Kopeček, I., Pala, K. (eds) Text, Speech and Dialogue. TSD 2010. Lecture Notes in Computer Science(), vol 6231. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-15760-8_59
Download citation
DOI: https://doi.org/10.1007/978-3-642-15760-8_59
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-15759-2
Online ISBN: 978-3-642-15760-8
eBook Packages: Computer ScienceComputer Science (R0)