<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="huang-etal-2017-finish">
    <titleInfo>
        <title>When to Finish? Optimal Beam Search for Neural Text Generation (modulo beam size)</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Liang</namePart>
        <namePart type="family">Huang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Kai</namePart>
        <namePart type="family">Zhao</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Mingbo</namePart>
        <namePart type="family">Ma</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2017-09</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Martha</namePart>
            <namePart type="family">Palmer</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Rebecca</namePart>
            <namePart type="family">Hwa</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Sebastian</namePart>
            <namePart type="family">Riedel</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Copenhagen, Denmark</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
    </relatedItem>
    <abstract>In neural text generation such as neural machine translation, summarization, and image captioning, beam search is widely used to improve the output text quality. However, in the neural generation setting, hypotheses can finish in different steps, which makes it difficult to decide when to end beam search to ensure optimality. We propose a provably optimal beam search algorithm that will always return the optimal-score complete hypothesis (modulo beam size), and finish as soon as the optimality is established. To counter neural generation’s tendency for shorter hypotheses, we also introduce a bounded length reward mechanism which allows a modified version of our beam search algorithm to remain optimal. Experiments on neural machine translation demonstrate that our principled beam search algorithm leads to improvement in BLEU score over previously proposed alternatives.</abstract>
    <identifier type="citekey">huang-etal-2017-finish</identifier>
    <identifier type="doi">10.18653/v1/D17-1227</identifier>
    <location>
        <url>https://aclanthology.org/D17-1227</url>
    </location>
    <part>
        <date>2017-09</date>
        <extent unit="page">
            <start>2134</start>
            <end>2139</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T When to Finish? Optimal Beam Search for Neural Text Generation (modulo beam size)
%A Huang, Liang
%A Zhao, Kai
%A Ma, Mingbo
%Y Palmer, Martha
%Y Hwa, Rebecca
%Y Riedel, Sebastian
%S Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing
%D 2017
%8 September
%I Association for Computational Linguistics
%C Copenhagen, Denmark
%F huang-etal-2017-finish
%X In neural text generation such as neural machine translation, summarization, and image captioning, beam search is widely used to improve the output text quality. However, in the neural generation setting, hypotheses can finish in different steps, which makes it difficult to decide when to end beam search to ensure optimality. We propose a provably optimal beam search algorithm that will always return the optimal-score complete hypothesis (modulo beam size), and finish as soon as the optimality is established. To counter neural generation’s tendency for shorter hypotheses, we also introduce a bounded length reward mechanism which allows a modified version of our beam search algorithm to remain optimal. Experiments on neural machine translation demonstrate that our principled beam search algorithm leads to improvement in BLEU score over previously proposed alternatives.
%R 10.18653/v1/D17-1227
%U https://aclanthology.org/D17-1227
%U https://doi.org/10.18653/v1/D17-1227
%P 2134-2139

Markdown (Informal)

[When to Finish? Optimal Beam Search for Neural Text Generation (modulo beam size)](https://aclanthology.org/D17-1227) (Huang et al., EMNLP 2017)

• When to Finish? Optimal Beam Search for Neural Text Generation (modulo beam size) (Huang et al., EMNLP 2017)

ACL

• Liang Huang, Kai Zhao, and Mingbo Ma. 2017. When to Finish? Optimal Beam Search for Neural Text Generation (modulo beam size). In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 2134–2139, Copenhagen, Denmark. Association for Computational Linguistics.