@inproceedings{yavuz-etal-2018-calcs,
title = "{C}a{L}cs: Continuously Approximating Longest Common Subsequence for Sequence Level Optimization",
author = "Yavuz, Semih and
Chiu, Chung-Cheng and
Nguyen, Patrick and
Wu, Yonghui",
editor = "Riloff, Ellen and
Chiang, David and
Hockenmaier, Julia and
Tsujii, Jun{'}ichi",
booktitle = "Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing",
month = oct # "-" # nov,
year = "2018",
address = "Brussels, Belgium",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/D18-1406",
doi = "10.18653/v1/D18-1406",
pages = "3708--3718",
abstract = "Maximum-likelihood estimation (MLE) is one of the most widely used approaches for training structured prediction models for text-generation based natural language processing applications. However, besides exposure bias, models trained with MLE suffer from wrong objective problem where they are trained to maximize the word-level correct next step prediction, but are evaluated with respect to sequence-level discrete metrics such as ROUGE and BLEU. Several variants of policy-gradient methods address some of these problems by optimizing for final discrete evaluation metrics and showing improvements over MLE training for downstream tasks like text summarization and machine translation. However, policy-gradient methods suffers from high sample variance, making the training process very difficult and unstable. In this paper, we present an alternative direction towards mitigating this problem by introducing a new objective (CaLcs) based on a differentiable surrogate of longest common subsequence (LCS) measure that captures sequence-level structure similarity. Experimental results on abstractive summarization and machine translation validate the effectiveness of the proposed approach.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yavuz-etal-2018-calcs">
<titleInfo>
<title>CaLcs: Continuously Approximating Longest Common Subsequence for Sequence Level Optimization</title>
</titleInfo>
<name type="personal">
<namePart type="given">Semih</namePart>
<namePart type="family">Yavuz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chung-Cheng</namePart>
<namePart type="family">Chiu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Patrick</namePart>
<namePart type="family">Nguyen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yonghui</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2018-oct-nov</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Ellen</namePart>
<namePart type="family">Riloff</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Chiang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Julia</namePart>
<namePart type="family">Hockenmaier</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jun’ichi</namePart>
<namePart type="family">Tsujii</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Brussels, Belgium</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Maximum-likelihood estimation (MLE) is one of the most widely used approaches for training structured prediction models for text-generation based natural language processing applications. However, besides exposure bias, models trained with MLE suffer from wrong objective problem where they are trained to maximize the word-level correct next step prediction, but are evaluated with respect to sequence-level discrete metrics such as ROUGE and BLEU. Several variants of policy-gradient methods address some of these problems by optimizing for final discrete evaluation metrics and showing improvements over MLE training for downstream tasks like text summarization and machine translation. However, policy-gradient methods suffers from high sample variance, making the training process very difficult and unstable. In this paper, we present an alternative direction towards mitigating this problem by introducing a new objective (CaLcs) based on a differentiable surrogate of longest common subsequence (LCS) measure that captures sequence-level structure similarity. Experimental results on abstractive summarization and machine translation validate the effectiveness of the proposed approach.</abstract>
<identifier type="citekey">yavuz-etal-2018-calcs</identifier>
<identifier type="doi">10.18653/v1/D18-1406</identifier>
<location>
<url>https://aclanthology.org/D18-1406</url>
</location>
<part>
<date>2018-oct-nov</date>
<extent unit="page">
<start>3708</start>
<end>3718</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T CaLcs: Continuously Approximating Longest Common Subsequence for Sequence Level Optimization
%A Yavuz, Semih
%A Chiu, Chung-Cheng
%A Nguyen, Patrick
%A Wu, Yonghui
%Y Riloff, Ellen
%Y Chiang, David
%Y Hockenmaier, Julia
%Y Tsujii, Jun’ichi
%S Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing
%D 2018
%8 oct nov
%I Association for Computational Linguistics
%C Brussels, Belgium
%F yavuz-etal-2018-calcs
%X Maximum-likelihood estimation (MLE) is one of the most widely used approaches for training structured prediction models for text-generation based natural language processing applications. However, besides exposure bias, models trained with MLE suffer from wrong objective problem where they are trained to maximize the word-level correct next step prediction, but are evaluated with respect to sequence-level discrete metrics such as ROUGE and BLEU. Several variants of policy-gradient methods address some of these problems by optimizing for final discrete evaluation metrics and showing improvements over MLE training for downstream tasks like text summarization and machine translation. However, policy-gradient methods suffers from high sample variance, making the training process very difficult and unstable. In this paper, we present an alternative direction towards mitigating this problem by introducing a new objective (CaLcs) based on a differentiable surrogate of longest common subsequence (LCS) measure that captures sequence-level structure similarity. Experimental results on abstractive summarization and machine translation validate the effectiveness of the proposed approach.
%R 10.18653/v1/D18-1406
%U https://aclanthology.org/D18-1406
%U https://doi.org/10.18653/v1/D18-1406
%P 3708-3718
Markdown (Informal)
[CaLcs: Continuously Approximating Longest Common Subsequence for Sequence Level Optimization](https://aclanthology.org/D18-1406) (Yavuz et al., EMNLP 2018)
ACL