@inproceedings{de-coster-etal-2021-frozen,
title = "Frozen Pretrained Transformers for Neural Sign Language Translation",
author = "De Coster, Mathieu and
D{'}Oosterlinck, Karel and
Pizurica, Marija and
Rabaey, Paloma and
Verlinden, Severine and
Van Herreweghe, Mieke and
Dambre, Joni",
editor = "Shterionov, Dimitar",
booktitle = "Proceedings of the 1st International Workshop on Automatic Translation for Signed and Spoken Languages (AT4SSL)",
month = aug,
year = "2021",
address = "Virtual",
publisher = "Association for Machine Translation in the Americas",
url = "https://aclanthology.org/2021.mtsummit-at4ssl.10",
pages = "88--97",
abstract = "One of the major challenges in sign language translation from a sign language to a spoken language is the lack of parallel corpora. Recent works have achieved promising results on the RWTH-PHOENIX-Weather 2014T dataset, which consists of over eight thousand parallel sentences between German sign language and German. However, from the perspective of neural machine translation, this is still a tiny dataset. To improve the performance of models trained on small datasets, transfer learning can be used. While this has been previously applied in sign language translation for feature extraction, to the best of our knowledge, pretrained language models have not yet been investigated. We use pretrained BERT-base and mBART-50 models to initialize our sign language video to spoken language text translation model. To mitigate overfitting, we apply the frozen pretrained transformer technique: we freeze the majority of parameters during training. Using a pretrained BERT model, we outperform a baseline trained from scratch by 1 to 2 BLEU-4. Our results show that pretrained language models can be used to improve sign language translation performance and that the self-attention patterns in BERT transfer in zero-shot to the encoder and decoder of sign language translation models.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="de-coster-etal-2021-frozen">
<titleInfo>
<title>Frozen Pretrained Transformers for Neural Sign Language Translation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Mathieu</namePart>
<namePart type="family">De Coster</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Karel</namePart>
<namePart type="family">D’Oosterlinck</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marija</namePart>
<namePart type="family">Pizurica</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Paloma</namePart>
<namePart type="family">Rabaey</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Severine</namePart>
<namePart type="family">Verlinden</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mieke</namePart>
<namePart type="family">Van Herreweghe</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joni</namePart>
<namePart type="family">Dambre</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2021-08</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 1st International Workshop on Automatic Translation for Signed and Spoken Languages (AT4SSL)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Dimitar</namePart>
<namePart type="family">Shterionov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Machine Translation in the Americas</publisher>
<place>
<placeTerm type="text">Virtual</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>One of the major challenges in sign language translation from a sign language to a spoken language is the lack of parallel corpora. Recent works have achieved promising results on the RWTH-PHOENIX-Weather 2014T dataset, which consists of over eight thousand parallel sentences between German sign language and German. However, from the perspective of neural machine translation, this is still a tiny dataset. To improve the performance of models trained on small datasets, transfer learning can be used. While this has been previously applied in sign language translation for feature extraction, to the best of our knowledge, pretrained language models have not yet been investigated. We use pretrained BERT-base and mBART-50 models to initialize our sign language video to spoken language text translation model. To mitigate overfitting, we apply the frozen pretrained transformer technique: we freeze the majority of parameters during training. Using a pretrained BERT model, we outperform a baseline trained from scratch by 1 to 2 BLEU-4. Our results show that pretrained language models can be used to improve sign language translation performance and that the self-attention patterns in BERT transfer in zero-shot to the encoder and decoder of sign language translation models.</abstract>
<identifier type="citekey">de-coster-etal-2021-frozen</identifier>
<location>
<url>https://aclanthology.org/2021.mtsummit-at4ssl.10</url>
</location>
<part>
<date>2021-08</date>
<extent unit="page">
<start>88</start>
<end>97</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Frozen Pretrained Transformers for Neural Sign Language Translation
%A De Coster, Mathieu
%A D’Oosterlinck, Karel
%A Pizurica, Marija
%A Rabaey, Paloma
%A Verlinden, Severine
%A Van Herreweghe, Mieke
%A Dambre, Joni
%Y Shterionov, Dimitar
%S Proceedings of the 1st International Workshop on Automatic Translation for Signed and Spoken Languages (AT4SSL)
%D 2021
%8 August
%I Association for Machine Translation in the Americas
%C Virtual
%F de-coster-etal-2021-frozen
%X One of the major challenges in sign language translation from a sign language to a spoken language is the lack of parallel corpora. Recent works have achieved promising results on the RWTH-PHOENIX-Weather 2014T dataset, which consists of over eight thousand parallel sentences between German sign language and German. However, from the perspective of neural machine translation, this is still a tiny dataset. To improve the performance of models trained on small datasets, transfer learning can be used. While this has been previously applied in sign language translation for feature extraction, to the best of our knowledge, pretrained language models have not yet been investigated. We use pretrained BERT-base and mBART-50 models to initialize our sign language video to spoken language text translation model. To mitigate overfitting, we apply the frozen pretrained transformer technique: we freeze the majority of parameters during training. Using a pretrained BERT model, we outperform a baseline trained from scratch by 1 to 2 BLEU-4. Our results show that pretrained language models can be used to improve sign language translation performance and that the self-attention patterns in BERT transfer in zero-shot to the encoder and decoder of sign language translation models.
%U https://aclanthology.org/2021.mtsummit-at4ssl.10
%P 88-97
Markdown (Informal)
[Frozen Pretrained Transformers for Neural Sign Language Translation](https://aclanthology.org/2021.mtsummit-at4ssl.10) (De Coster et al., MTSummit 2021)
ACL
- Mathieu De Coster, Karel D’Oosterlinck, Marija Pizurica, Paloma Rabaey, Severine Verlinden, Mieke Van Herreweghe, and Joni Dambre. 2021. Frozen Pretrained Transformers for Neural Sign Language Translation. In Proceedings of the 1st International Workshop on Automatic Translation for Signed and Spoken Languages (AT4SSL), pages 88–97, Virtual. Association for Machine Translation in the Americas.