@inproceedings{xu-etal-2023-ctc,
title = "{CTC}-based Non-autoregressive Speech Translation",
author = "Xu, Chen and
Liu, Xiaoqian and
Liu, Xiaowen and
Sun, Qingxuan and
Zhang, Yuhao and
Yang, Murun and
Dong, Qianqian and
Ko, Tom and
Wang, Mingxuan and
Xiao, Tong and
Ma, Anxiang and
Zhu, Jingbo",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.744",
doi = "10.18653/v1/2023.acl-long.744",
pages = "13321--13339",
abstract = "Combining end-to-end speech translation (ST) and non-autoregressive (NAR) generation is promising in language and speech processing for their advantages of less error propagation and low latency. In this paper, we investigate the potential of connectionist temporal classification (CTC) for non-autoregressive speech translation (NAST).In particular, we develop a model consisting of two encoders that are guided by CTC to predict the source and target texts, respectively. Introducing CTC into NAST on both language sides has obvious challenges: 1) the conditional independent generation somewhat breaks the interdependency among tokens, and 2) the monotonic alignment assumption in standard CTC does not hold in translation tasks. In response, we develop a prediction-aware encoding approach and a cross-layer attention approach to address these issues. We also use curriculum learning to improve convergence of training. Experiments on the MuST-C ST benchmarks show that our NAST model achieves an average BLEU score of 29.5 with a speed-up of 5.67$\times$, which is comparable to the autoregressive counterpart and even outperforms the previous best result of 0.9 BLEU points.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="xu-etal-2023-ctc">
<titleInfo>
<title>CTC-based Non-autoregressive Speech Translation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Chen</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaoqian</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaowen</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qingxuan</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuhao</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Murun</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qianqian</namePart>
<namePart type="family">Dong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tom</namePart>
<namePart type="family">Ko</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mingxuan</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tong</namePart>
<namePart type="family">Xiao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Anxiang</namePart>
<namePart type="family">Ma</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jingbo</namePart>
<namePart type="family">Zhu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Anna</namePart>
<namePart type="family">Rogers</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jordan</namePart>
<namePart type="family">Boyd-Graber</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Naoaki</namePart>
<namePart type="family">Okazaki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Toronto, Canada</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Combining end-to-end speech translation (ST) and non-autoregressive (NAR) generation is promising in language and speech processing for their advantages of less error propagation and low latency. In this paper, we investigate the potential of connectionist temporal classification (CTC) for non-autoregressive speech translation (NAST).In particular, we develop a model consisting of two encoders that are guided by CTC to predict the source and target texts, respectively. Introducing CTC into NAST on both language sides has obvious challenges: 1) the conditional independent generation somewhat breaks the interdependency among tokens, and 2) the monotonic alignment assumption in standard CTC does not hold in translation tasks. In response, we develop a prediction-aware encoding approach and a cross-layer attention approach to address these issues. We also use curriculum learning to improve convergence of training. Experiments on the MuST-C ST benchmarks show that our NAST model achieves an average BLEU score of 29.5 with a speed-up of 5.67\times, which is comparable to the autoregressive counterpart and even outperforms the previous best result of 0.9 BLEU points.</abstract>
<identifier type="citekey">xu-etal-2023-ctc</identifier>
<identifier type="doi">10.18653/v1/2023.acl-long.744</identifier>
<location>
<url>https://aclanthology.org/2023.acl-long.744</url>
</location>
<part>
<date>2023-07</date>
<extent unit="page">
<start>13321</start>
<end>13339</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T CTC-based Non-autoregressive Speech Translation
%A Xu, Chen
%A Liu, Xiaoqian
%A Liu, Xiaowen
%A Sun, Qingxuan
%A Zhang, Yuhao
%A Yang, Murun
%A Dong, Qianqian
%A Ko, Tom
%A Wang, Mingxuan
%A Xiao, Tong
%A Ma, Anxiang
%A Zhu, Jingbo
%Y Rogers, Anna
%Y Boyd-Graber, Jordan
%Y Okazaki, Naoaki
%S Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2023
%8 July
%I Association for Computational Linguistics
%C Toronto, Canada
%F xu-etal-2023-ctc
%X Combining end-to-end speech translation (ST) and non-autoregressive (NAR) generation is promising in language and speech processing for their advantages of less error propagation and low latency. In this paper, we investigate the potential of connectionist temporal classification (CTC) for non-autoregressive speech translation (NAST).In particular, we develop a model consisting of two encoders that are guided by CTC to predict the source and target texts, respectively. Introducing CTC into NAST on both language sides has obvious challenges: 1) the conditional independent generation somewhat breaks the interdependency among tokens, and 2) the monotonic alignment assumption in standard CTC does not hold in translation tasks. In response, we develop a prediction-aware encoding approach and a cross-layer attention approach to address these issues. We also use curriculum learning to improve convergence of training. Experiments on the MuST-C ST benchmarks show that our NAST model achieves an average BLEU score of 29.5 with a speed-up of 5.67\times, which is comparable to the autoregressive counterpart and even outperforms the previous best result of 0.9 BLEU points.
%R 10.18653/v1/2023.acl-long.744
%U https://aclanthology.org/2023.acl-long.744
%U https://doi.org/10.18653/v1/2023.acl-long.744
%P 13321-13339
Markdown (Informal)
[CTC-based Non-autoregressive Speech Translation](https://aclanthology.org/2023.acl-long.744) (Xu et al., ACL 2023)
ACL
- Chen Xu, Xiaoqian Liu, Xiaowen Liu, Qingxuan Sun, Yuhao Zhang, Murun Yang, Qianqian Dong, Tom Ko, Mingxuan Wang, Tong Xiao, Anxiang Ma, and Jingbo Zhu. 2023. CTC-based Non-autoregressive Speech Translation. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 13321–13339, Toronto, Canada. Association for Computational Linguistics.