@inproceedings{lee-etal-2023-consistency,
title = "Consistency is Key: On Data-Efficient Modality Transfer in Speech Translation",
author = "Lee, Hojin and
Lee, Changmin and
Hwang, Seung-won",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.905/",
doi = "10.18653/v1/2023.findings-emnlp.905",
pages = "13572--13581",
abstract = "End-to-end approaches have shown promising results for speech translation (ST), but they suffer from its data scarcity compared to machine translation (MT). To address this, progressive training has become a common practice, of using external MT data during the fine-tuning phase. Despite of its prevalence and computational overhead, its validity is not extensively corroborated yet. This paper conducts an empirical investigation and finds that progressive training is ineffective. We identify learning-forgetting trade-off as a critical obstacle, then hypothesize and verify that consistency learning (CL) breaks the dilemma of learning-forgetting. The proposed method, which combines knowledge distillation (KD) and CL, outperforms the previous methods on MuST-C dataset even without additional data, and our proposed consistency-informed KD achieves additional improvements against KD+CL. Code and models are availble at https://github.com/hjlee1371/consistency-s2tt."
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lee-etal-2023-consistency">
<titleInfo>
<title>Consistency is Key: On Data-Efficient Modality Transfer in Speech Translation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hojin</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Changmin</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Seung-won</namePart>
<namePart type="family">Hwang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-12</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2023</title>
</titleInfo>
<name type="personal">
<namePart type="given">Houda</namePart>
<namePart type="family">Bouamor</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Juan</namePart>
<namePart type="family">Pino</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kalika</namePart>
<namePart type="family">Bali</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Singapore</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>End-to-end approaches have shown promising results for speech translation (ST), but they suffer from its data scarcity compared to machine translation (MT). To address this, progressive training has become a common practice, of using external MT data during the fine-tuning phase. Despite of its prevalence and computational overhead, its validity is not extensively corroborated yet. This paper conducts an empirical investigation and finds that progressive training is ineffective. We identify learning-forgetting trade-off as a critical obstacle, then hypothesize and verify that consistency learning (CL) breaks the dilemma of learning-forgetting. The proposed method, which combines knowledge distillation (KD) and CL, outperforms the previous methods on MuST-C dataset even without additional data, and our proposed consistency-informed KD achieves additional improvements against KD+CL. Code and models are availble at https://github.com/hjlee1371/consistency-s2tt.</abstract>
<identifier type="citekey">lee-etal-2023-consistency</identifier>
<identifier type="doi">10.18653/v1/2023.findings-emnlp.905</identifier>
<location>
<url>https://aclanthology.org/2023.findings-emnlp.905/</url>
</location>
<part>
<date>2023-12</date>
<extent unit="page">
<start>13572</start>
<end>13581</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Consistency is Key: On Data-Efficient Modality Transfer in Speech Translation
%A Lee, Hojin
%A Lee, Changmin
%A Hwang, Seung-won
%Y Bouamor, Houda
%Y Pino, Juan
%Y Bali, Kalika
%S Findings of the Association for Computational Linguistics: EMNLP 2023
%D 2023
%8 December
%I Association for Computational Linguistics
%C Singapore
%F lee-etal-2023-consistency
%X End-to-end approaches have shown promising results for speech translation (ST), but they suffer from its data scarcity compared to machine translation (MT). To address this, progressive training has become a common practice, of using external MT data during the fine-tuning phase. Despite of its prevalence and computational overhead, its validity is not extensively corroborated yet. This paper conducts an empirical investigation and finds that progressive training is ineffective. We identify learning-forgetting trade-off as a critical obstacle, then hypothesize and verify that consistency learning (CL) breaks the dilemma of learning-forgetting. The proposed method, which combines knowledge distillation (KD) and CL, outperforms the previous methods on MuST-C dataset even without additional data, and our proposed consistency-informed KD achieves additional improvements against KD+CL. Code and models are availble at https://github.com/hjlee1371/consistency-s2tt.
%R 10.18653/v1/2023.findings-emnlp.905
%U https://aclanthology.org/2023.findings-emnlp.905/
%U https://doi.org/10.18653/v1/2023.findings-emnlp.905
%P 13572-13581
Markdown (Informal)
[Consistency is Key: On Data-Efficient Modality Transfer in Speech Translation](https://aclanthology.org/2023.findings-emnlp.905/) (Lee et al., Findings 2023)
ACL