@inproceedings{li-chan-2019-big,
title = "Big Bidirectional Insertion Representations for Documents",
author = "Li, Lala and
Chan, William",
booktitle = "Proceedings of the 3rd Workshop on Neural Generation and Translation",
month = nov,
year = "2019",
address = "Hong Kong",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/D19-5620",
doi = "10.18653/v1/D19-5620",
pages = "194--198",
abstract = "The Insertion Transformer is well suited for long form text generation due to its parallel generation capabilities, requiring $O(\log_2 n)$ generation steps to generate $n$ tokens. However, modeling long sequences is difficult, as there is more ambiguity captured in the attention mechanism. This work proposes the Big Bidirectional Insertion Representations for Documents (Big BIRD), an insertion-based model for document-level translation tasks. We scale up the insertion-based models to long form documents. Our key contribution is introducing sentence alignment via sentence-positional embeddings between the source and target document. We show an improvement of +4.3 BLEU on the WMT{'}19 English-{\textgreater}German document-level translation task compared with the Insertion Transformer baseline.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="li-chan-2019-big">
<titleInfo>
<title>Big Bidirectional Insertion Representations for Documents</title>
</titleInfo>
<name type="personal">
<namePart type="given">Lala</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">William</namePart>
<namePart type="family">Chan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2019-nov</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 3rd Workshop on Neural Generation and Translation</title>
</titleInfo>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Hong Kong</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>The Insertion Transformer is well suited for long form text generation due to its parallel generation capabilities, requiring $O(\log_2 n)$ generation steps to generate $n$ tokens. However, modeling long sequences is difficult, as there is more ambiguity captured in the attention mechanism. This work proposes the Big Bidirectional Insertion Representations for Documents (Big BIRD), an insertion-based model for document-level translation tasks. We scale up the insertion-based models to long form documents. Our key contribution is introducing sentence alignment via sentence-positional embeddings between the source and target document. We show an improvement of +4.3 BLEU on the WMT’19 English-\textgreaterGerman document-level translation task compared with the Insertion Transformer baseline.</abstract>
<identifier type="citekey">li-chan-2019-big</identifier>
<identifier type="doi">10.18653/v1/D19-5620</identifier>
<location>
<url>https://aclanthology.org/D19-5620</url>
</location>
<part>
<date>2019-nov</date>
<extent unit="page">
<start>194</start>
<end>198</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Big Bidirectional Insertion Representations for Documents
%A Li, Lala
%A Chan, William
%S Proceedings of the 3rd Workshop on Neural Generation and Translation
%D 2019
%8 nov
%I Association for Computational Linguistics
%C Hong Kong
%F li-chan-2019-big
%X The Insertion Transformer is well suited for long form text generation due to its parallel generation capabilities, requiring $O(\log_2 n)$ generation steps to generate $n$ tokens. However, modeling long sequences is difficult, as there is more ambiguity captured in the attention mechanism. This work proposes the Big Bidirectional Insertion Representations for Documents (Big BIRD), an insertion-based model for document-level translation tasks. We scale up the insertion-based models to long form documents. Our key contribution is introducing sentence alignment via sentence-positional embeddings between the source and target document. We show an improvement of +4.3 BLEU on the WMT’19 English-\textgreaterGerman document-level translation task compared with the Insertion Transformer baseline.
%R 10.18653/v1/D19-5620
%U https://aclanthology.org/D19-5620
%U https://doi.org/10.18653/v1/D19-5620
%P 194-198
Markdown (Informal)
[Big Bidirectional Insertion Representations for Documents](https://aclanthology.org/D19-5620) (Li & Chan, EMNLP 2019)
ACL