<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="rothe-etal-2020-leveraging">
    <titleInfo>
        <title>Leveraging Pre-trained Checkpoints for Sequence Generation Tasks</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Sascha</namePart>
        <namePart type="family">Rothe</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Shashi</namePart>
        <namePart type="family">Narayan</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Aliaksei</namePart>
        <namePart type="family">Severyn</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2020</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <genre authority="bibutilsgt">journal article</genre>
    <relatedItem type="host">
        <titleInfo>
            <title>Transactions of the Association for Computational Linguistics</title>
        </titleInfo>
        <originInfo>
            <issuance>continuing</issuance>
            <publisher>MIT Press</publisher>
            <place>
                <placeTerm type="text">Cambridge, MA</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">periodical</genre>
        <genre authority="bibutilsgt">academic journal</genre>
    </relatedItem>
    <abstract>Unsupervised pre-training of large neural models has recently revolutionized Natural Language Processing. By warm-starting from the publicly released checkpoints, NLP practitioners have pushed the state-of-the-art on multiple benchmarks while saving significant amounts of compute time. So far the focus has been mainly on the Natural Language Understanding tasks. In this paper, we demonstrate the efficacy of pre-trained checkpoints for Sequence Generation. We developed a Transformer-based sequence-to-sequence model that is compatible with publicly available pre-trained BERT, GPT-2, and RoBERTa checkpoints and conducted an extensive empirical study on the utility of initializing our model, both encoder and decoder, with these checkpoints. Our models result in new state-of-the-art results on Machine Translation, Text Summarization, Sentence Splitting, and Sentence Fusion.</abstract>
    <identifier type="citekey">rothe-etal-2020-leveraging</identifier>
    <identifier type="doi">10.1162/tacl_a_00313</identifier>
    <location>
        <url>https://aclanthology.org/2020.tacl-1.18</url>
    </location>
    <part>
        <date>2020</date>
        <detail type="volume"><number>8</number></detail>
        <extent unit="page">
            <start>264</start>
            <end>280</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Journal Article
%T Leveraging Pre-trained Checkpoints for Sequence Generation Tasks
%A Rothe, Sascha
%A Narayan, Shashi
%A Severyn, Aliaksei
%J Transactions of the Association for Computational Linguistics
%D 2020
%V 8
%I MIT Press
%C Cambridge, MA
%F rothe-etal-2020-leveraging
%X Unsupervised pre-training of large neural models has recently revolutionized Natural Language Processing. By warm-starting from the publicly released checkpoints, NLP practitioners have pushed the state-of-the-art on multiple benchmarks while saving significant amounts of compute time. So far the focus has been mainly on the Natural Language Understanding tasks. In this paper, we demonstrate the efficacy of pre-trained checkpoints for Sequence Generation. We developed a Transformer-based sequence-to-sequence model that is compatible with publicly available pre-trained BERT, GPT-2, and RoBERTa checkpoints and conducted an extensive empirical study on the utility of initializing our model, both encoder and decoder, with these checkpoints. Our models result in new state-of-the-art results on Machine Translation, Text Summarization, Sentence Splitting, and Sentence Fusion.
%R 10.1162/tacl_a_00313
%U https://aclanthology.org/2020.tacl-1.18
%U https://doi.org/10.1162/tacl_a_00313
%P 264-280

Markdown (Informal)

[Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://aclanthology.org/2020.tacl-1.18) (Rothe et al., TACL 2020)

• Leveraging Pre-trained Checkpoints for Sequence Generation Tasks (Rothe et al., TACL 2020)

ACL

• Sascha Rothe, Shashi Narayan, and Aliaksei Severyn. 2020. Leveraging Pre-trained Checkpoints for Sequence Generation Tasks. Transactions of the Association for Computational Linguistics, 8:264–280.