Please use this identifier to cite or link to this item:https://hdl.handle.net/20.500.12259/101536
Type of publication: Straipsnis Clarivate Analytics Web of Science ar/ir Scopus / Article in Clarivate Analytics Web of Science or / and Scopus (S1)
Field of Science: Informatika / Computer science (N009)
Author(s): Raškinis, Gailius;Paškauskaitė, Gintarė;Saudargienė, Aušra;Kazlauskienė, Asta;Vaičiūnas, Airenas
Title: Comparison of phonemic and graphemic word to sub-word unit mappings for Lithuanian phone-level speech transcription
Is part of: Informatica : an international journal. Vilnius : Vilnius university Institute of mathematics and informatics, 2019, vol. 30, no. 3
Extent: p. 573-593
Date: 2019
Keywords: Speech recognition;G2P conversion;Lithuanian language;Grapheme;Phoneme
Abstract: Conventional large vocabulary automatic speech recognition (ASR) systems require a mapping from words into sub-word units to generalize over the words that were absent in the training data and to enable the robust estimation of acoustic model parameters. This paper surveys the research done during the last 15 years on the topic of word to sub-word mappings for Lithuanian ASR systems. It also compares various phoneme and grapheme based mappings across a broad range of acoustic modelling techniques including monophone and triphone based Hidden Markov models (HMM), speaker adaptively trained HMMs, subspace gaussian mixture models (SGMM), feed-forward time delay neural network (TDNN), and state-of-the-art low frame rate bidirectional long short term memory (LFR BLSTM) recurrent deep neural network. Experimental comparisons are based on a 50-hour speech corpus. This paper shows that the best phone-based mapping significantly outperforms a grapheme-based mapping. It also shows that the lowest phone error rate of an ASR system is achieved by the phoneme-based lexicon that explicitly models syllable stress and represents diphthongs as single phonetic units
Internet: https://www.mii.lt/informatica/pdf/INFO1233.pdf
https://www.mii.lt/informatica/pdf/INFO1233.pdf
http://dx.doi.org/10.15388/Informatica.2019.219
Affiliation(s): Humanitarinių mokslų fakultetas
Informatikos fakultetas
Kauno technologijos universitetas
Lietuvos sveikatos mokslų universitetas
Lituanistikos katedra
Taikomosios informatikos katedra
Vytauto Didžiojo universitetas
Appears in Collections:Universiteto mokslo publikacijos / University Research Publications

Files in This Item:
marc.xml7.78 kBXMLView/Open

MARC21 XML metadata

Show full item record

Google ScholarTM

Check

Altmetric


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.