Abstract
This paper addresses the problem of modeling Korean pronunciation variation as a sequential labeling task where tokens in the source language (phonemic symbols) are labeled with tokens in the target language (orthographic Korean transcription). This is done by utilizing conditional random fields (CRFs), which are undirected graphical models that maximize the posterior probabilities of the label target sequence given the input source sequence. In this study, the proposed CRFbased pronunciation variation model is applied to our Korean LVCSR after we perform the standard hidden Markov model (HMM)-based recognition of the phonemic syllable of the actual pronunciation (surface forms). The goal is then to output a sequence of Korean orthography given a sequence of phonemic syllable surface forms. Experimental results show that the proposed CRF model could help enhance our Korean large-vocabulary continuous speech recognition system.
This is a preview of subscription content, log in via an institution to check access.
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
H. Strik and C. Cucchiarini, “Modeling pronunciation variation for ASR: A survey of the literature,” Speech Communication, vol. 29, pp. 225–246, 1999.
B. Kim, G.G. Lee, and J. Lee, “Morpheme-based grapheme to phoneme conversion using phonetic patterns and morphophonemic connectivity information,” ACM Transactions on Asian Language Information Processing, vol. 1, no. 1, pp. 6582, 2002.
J. Jeon, S. Wee, and M. Chung, “Generating pronunciation dictionary by analyzing phonological variations frequently found in spoken Korean,” in Proc. of International Conference on Speech Processing, 1997, pp. 519–524.
J. Jeon, S. Cha, M. Chung, and J. Park, “Automatic generation of Korean pronunciation variants by multistage applications of phonological rules,” in Proc. of ICSLP, Sydney, Australia, 1998, pp. 1943–1946.
J. Lafferty, A. McCallum, and F. Pereira, “Conditional random fields: Probabilistic models for segmenting and labeling sequence data,” in Proc. of ICML, Williamstown, MA, USA, 2001, pp. 282–289.
T. Kudo, K. Yamamoto, and Y. Matsumoto, “Applying conditional random fields to Japanese morphological analysis,” in Proc of EMNLP, 2004, pp. 230–237.
F. Sha and F. Pereira, “Shallow parsing with conditional random fields,” in Proc. of HLTNAACL, Edmonton, Canada, 2003, pp. 213–220.
J.R. Finkel and C.D. Manning, “Joint parsing and named entity recognition,” in Proc. of NAACL, Boulder, Colorado, USA, 2009, pp. 326–334.
J. Hammersley and P. Clifford, “Markov fields and finite graphs and lattices,” 1971.
T. Kudo, “CRF++: Yet another CRF toolkit,” http://crfpp.sourceforge.net/, 2005.
M. Kim, Y.R. Oh, and H.K. Kim, “Non-native pronunciation variation modeling using an indirect data driven method,” in Proc. of ASRU, Kyoto, Japan, 2007, pp. 231–236.
T. Jitsuhiro, T. Matsui, and S. Nakamura, “Automatic generation of non-uniform HMM topologies based on the MDL criterion,” IEICE Trans. Inf. & Syst., vol. E87-D, no. 8, pp. 2121–2129, 2004.
H. Li, M. Zhang, and J. Su, “A joint source-channel model for machine transliteration,” in Proc. of ACL, Barcelona, Spain, 2004, pp. 160–167.
M. Bisani and H. Ney, “A joint-sequence models for grapheme-to-phoneme conversion,” Speech Communication, vol. 50, pp. 434–451, 2008.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2011 Springer Science+Business Media, LLC
About this paper
Cite this paper
Sakti, S., Finch, A., Hori, C., Kashioka, H., Nakamura, S. (2011). Conditional Random Fields for Modeling Korean Pronunciation Variation. In: Delgado, RC., Kobayashi, T. (eds) Proceedings of the Paralinguistic Information and its Integration in Spoken Dialogue Systems Workshop. Springer, New York, NY. https://doi.org/10.1007/978-1-4614-1335-6_7
Download citation
DOI: https://doi.org/10.1007/978-1-4614-1335-6_7
Published:
Publisher Name: Springer, New York, NY
Print ISBN: 978-1-4614-1334-9
Online ISBN: 978-1-4614-1335-6
eBook Packages: EngineeringEngineering (R0)