T5G2P: Text-to-Text Transfer Transformer Based Grapheme-to-Phoneme Conversion
The result's identifiers
Result code in IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F49777513%3A23520%2F24%3A43972436" target="_blank" >RIV/49777513:23520/24:43972436 - isvavai.cz</a>
Result on the web
<a href="https://ieeexplore.ieee.org/document/10592637" target="_blank" >https://ieeexplore.ieee.org/document/10592637</a>
DOI - Digital Object Identifier
<a href="http://dx.doi.org/10.1109/TASLP.2024.3426332" target="_blank" >10.1109/TASLP.2024.3426332</a>
Alternative languages
Result language
angličtina
Original language name
T5G2P: Text-to-Text Transfer Transformer Based Grapheme-to-Phoneme Conversion
Original language description
The present paper explores the use of several deep neural network architectures to carry out a grapheme-to-phoneme (G2P) conversion, aiming to find a universal and language-independent approach to the task. The models explored are trained on whole sentences in order to automatically capture cross-word context (such as voicedness assimilation) if it exists in the given language. Four different languages, English, Czech, Russian, and German, were chosen due to their different nature and requirements for the G2P task. Ultimately, the Text-to-Text Transfer Transformer (T5) based model achieved very high conversion accuracy on all the tested languages. Also, it exceeded the accuracy reached by a similar system, when trained on a public LibriSpeech database.
Czech name
—
Czech description
—
Classification
Type
J<sub>imp</sub> - Article in a specialist periodical, which is included in the Web of Science database
CEP classification
—
OECD FORD branch
20205 - Automation and control systems
Result continuities
Project
<a href="/en/project/GA22-27800S" target="_blank" >GA22-27800S: Transformers of multiple modalities for more natural spoken dialog</a><br>
Continuities
P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)
Others
Publication year
2024
Confidentiality
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů
Data specific for result type
Name of the periodical
IEEE/ACM Transactions on Audio, Speech, and Language Processing
ISSN
2329-9290
e-ISSN
2329-9304
Volume of the periodical
32
Issue of the periodical within the volume
July 2024
Country of publishing house
US - UNITED STATES
Number of pages
11
Pages from-to
3466-3476
UT code for WoS article
001283673700010
EID of the result in the Scopus database
2-s2.0-85198311174