@inproceedings{route-etal-2019-multimodal,
title = "Multimodal, Multilingual Grapheme-to-Phoneme Conversion for Low-Resource Languages",
author = "Route, James and
Hillis, Steven and
Czeresnia Etinger, Isak and
Zhang, Han and
Black, Alan W",
editor = "Cherry, Colin and
Durrett, Greg and
Foster, George and
Haffari, Reza and
Khadivi, Shahram and
Peng, Nanyun and
Ren, Xiang and
Swayamdipta, Swabha",
booktitle = "Proceedings of the 2nd Workshop on Deep Learning Approaches for Low-Resource NLP (DeepLo 2019)",
month = nov,
year = "2019",
address = "Hong Kong, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/D19-6121",
doi = "10.18653/v1/D19-6121",
pages = "192--201",
abstract = "Grapheme-to-phoneme conversion (g2p) is the task of predicting the pronunciation of words from their orthographic representation. His- torically, g2p systems were transition- or rule- based, making generalization beyond a mono- lingual (high resource) domain impractical. Recently, neural architectures have enabled multilingual systems to generalize widely; however, all systems to date have been trained only on spelling-pronunciation pairs. We hy- pothesize that the sequences of IPA characters used to represent pronunciation do not capture its full nuance, especially when cleaned to fa- cilitate machine learning. We leverage audio data as an auxiliary modality in a multi-task training process to learn a more optimal inter- mediate representation of source graphemes; this is the first multimodal model proposed for multilingual g2p. Our approach is highly ef- fective: on our in-domain test set, our mul- timodal model reduces phoneme error rate to 2.46{\%}, a more than 65{\%} decrease compared to our implementation of a unimodal spelling- pronunciation model{---}which itself achieves state-of-the-art results on the Wiktionary test set. The advantages of the multimodal model generalize to wholly unseen languages, reduc- ing phoneme error rate on our out-of-domain test set to 6.39{\%} from the unimodal 8.21{\%}, a more than 20{\%} relative decrease. Further- more, our training and test sets are composed primarily of low-resource languages, demon- strating that our multimodal approach remains useful when training data are constrained.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="route-etal-2019-multimodal">
<titleInfo>
<title>Multimodal, Multilingual Grapheme-to-Phoneme Conversion for Low-Resource Languages</title>
</titleInfo>
<name type="personal">
<namePart type="given">James</namePart>
<namePart type="family">Route</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Steven</namePart>
<namePart type="family">Hillis</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Isak</namePart>
<namePart type="family">Czeresnia Etinger</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Han</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alan</namePart>
<namePart type="given">W</namePart>
<namePart type="family">Black</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2019-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2nd Workshop on Deep Learning Approaches for Low-Resource NLP (DeepLo 2019)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Colin</namePart>
<namePart type="family">Cherry</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Greg</namePart>
<namePart type="family">Durrett</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">George</namePart>
<namePart type="family">Foster</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Reza</namePart>
<namePart type="family">Haffari</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shahram</namePart>
<namePart type="family">Khadivi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nanyun</namePart>
<namePart type="family">Peng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiang</namePart>
<namePart type="family">Ren</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Swabha</namePart>
<namePart type="family">Swayamdipta</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Hong Kong, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Grapheme-to-phoneme conversion (g2p) is the task of predicting the pronunciation of words from their orthographic representation. His- torically, g2p systems were transition- or rule- based, making generalization beyond a mono- lingual (high resource) domain impractical. Recently, neural architectures have enabled multilingual systems to generalize widely; however, all systems to date have been trained only on spelling-pronunciation pairs. We hy- pothesize that the sequences of IPA characters used to represent pronunciation do not capture its full nuance, especially when cleaned to fa- cilitate machine learning. We leverage audio data as an auxiliary modality in a multi-task training process to learn a more optimal inter- mediate representation of source graphemes; this is the first multimodal model proposed for multilingual g2p. Our approach is highly ef- fective: on our in-domain test set, our mul- timodal model reduces phoneme error rate to 2.46%, a more than 65% decrease compared to our implementation of a unimodal spelling- pronunciation model—which itself achieves state-of-the-art results on the Wiktionary test set. The advantages of the multimodal model generalize to wholly unseen languages, reduc- ing phoneme error rate on our out-of-domain test set to 6.39% from the unimodal 8.21%, a more than 20% relative decrease. Further- more, our training and test sets are composed primarily of low-resource languages, demon- strating that our multimodal approach remains useful when training data are constrained.</abstract>
<identifier type="citekey">route-etal-2019-multimodal</identifier>
<identifier type="doi">10.18653/v1/D19-6121</identifier>
<location>
<url>https://aclanthology.org/D19-6121</url>
</location>
<part>
<date>2019-11</date>
<extent unit="page">
<start>192</start>
<end>201</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Multimodal, Multilingual Grapheme-to-Phoneme Conversion for Low-Resource Languages
%A Route, James
%A Hillis, Steven
%A Czeresnia Etinger, Isak
%A Zhang, Han
%A Black, Alan W.
%Y Cherry, Colin
%Y Durrett, Greg
%Y Foster, George
%Y Haffari, Reza
%Y Khadivi, Shahram
%Y Peng, Nanyun
%Y Ren, Xiang
%Y Swayamdipta, Swabha
%S Proceedings of the 2nd Workshop on Deep Learning Approaches for Low-Resource NLP (DeepLo 2019)
%D 2019
%8 November
%I Association for Computational Linguistics
%C Hong Kong, China
%F route-etal-2019-multimodal
%X Grapheme-to-phoneme conversion (g2p) is the task of predicting the pronunciation of words from their orthographic representation. His- torically, g2p systems were transition- or rule- based, making generalization beyond a mono- lingual (high resource) domain impractical. Recently, neural architectures have enabled multilingual systems to generalize widely; however, all systems to date have been trained only on spelling-pronunciation pairs. We hy- pothesize that the sequences of IPA characters used to represent pronunciation do not capture its full nuance, especially when cleaned to fa- cilitate machine learning. We leverage audio data as an auxiliary modality in a multi-task training process to learn a more optimal inter- mediate representation of source graphemes; this is the first multimodal model proposed for multilingual g2p. Our approach is highly ef- fective: on our in-domain test set, our mul- timodal model reduces phoneme error rate to 2.46%, a more than 65% decrease compared to our implementation of a unimodal spelling- pronunciation model—which itself achieves state-of-the-art results on the Wiktionary test set. The advantages of the multimodal model generalize to wholly unseen languages, reduc- ing phoneme error rate on our out-of-domain test set to 6.39% from the unimodal 8.21%, a more than 20% relative decrease. Further- more, our training and test sets are composed primarily of low-resource languages, demon- strating that our multimodal approach remains useful when training data are constrained.
%R 10.18653/v1/D19-6121
%U https://aclanthology.org/D19-6121
%U https://doi.org/10.18653/v1/D19-6121
%P 192-201
Markdown (Informal)
[Multimodal, Multilingual Grapheme-to-Phoneme Conversion for Low-Resource Languages](https://aclanthology.org/D19-6121) (Route et al., 2019)
ACL