@inproceedings{goyal-etal-2017-differentiable,
title = "Differentiable Scheduled Sampling for Credit Assignment",
author = "Goyal, Kartik and
Dyer, Chris and
Berg-Kirkpatrick, Taylor",
editor = "Barzilay, Regina and
Kan, Min-Yen",
booktitle = "Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)",
month = jul,
year = "2017",
address = "Vancouver, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/P17-2058/",
doi = "10.18653/v1/P17-2058",
pages = "366--371",
abstract = "We demonstrate that a continuous relaxation of the argmax operation can be used to create a differentiable approximation to greedy decoding in sequence-to-sequence (seq2seq) models. By incorporating this approximation into the scheduled sampling training procedure{--}a well-known technique for correcting exposure bias{--}we introduce a new training objective that is continuous and differentiable everywhere and can provide informative gradients near points where previous decoding decisions change their value. By using a related approximation, we also demonstrate a similar approach to sampled-based training. We show that our approach outperforms both standard cross-entropy training and scheduled sampling procedures in two sequence prediction tasks: named entity recognition and machine translation."
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="goyal-etal-2017-differentiable">
<titleInfo>
<title>Differentiable Scheduled Sampling for Credit Assignment</title>
</titleInfo>
<name type="personal">
<namePart type="given">Kartik</namePart>
<namePart type="family">Goyal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chris</namePart>
<namePart type="family">Dyer</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Taylor</namePart>
<namePart type="family">Berg-Kirkpatrick</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2017-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Regina</namePart>
<namePart type="family">Barzilay</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Min-Yen</namePart>
<namePart type="family">Kan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Vancouver, Canada</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>We demonstrate that a continuous relaxation of the argmax operation can be used to create a differentiable approximation to greedy decoding in sequence-to-sequence (seq2seq) models. By incorporating this approximation into the scheduled sampling training procedure–a well-known technique for correcting exposure bias–we introduce a new training objective that is continuous and differentiable everywhere and can provide informative gradients near points where previous decoding decisions change their value. By using a related approximation, we also demonstrate a similar approach to sampled-based training. We show that our approach outperforms both standard cross-entropy training and scheduled sampling procedures in two sequence prediction tasks: named entity recognition and machine translation.</abstract>
<identifier type="citekey">goyal-etal-2017-differentiable</identifier>
<identifier type="doi">10.18653/v1/P17-2058</identifier>
<location>
<url>https://aclanthology.org/P17-2058/</url>
</location>
<part>
<date>2017-07</date>
<extent unit="page">
<start>366</start>
<end>371</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Differentiable Scheduled Sampling for Credit Assignment
%A Goyal, Kartik
%A Dyer, Chris
%A Berg-Kirkpatrick, Taylor
%Y Barzilay, Regina
%Y Kan, Min-Yen
%S Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)
%D 2017
%8 July
%I Association for Computational Linguistics
%C Vancouver, Canada
%F goyal-etal-2017-differentiable
%X We demonstrate that a continuous relaxation of the argmax operation can be used to create a differentiable approximation to greedy decoding in sequence-to-sequence (seq2seq) models. By incorporating this approximation into the scheduled sampling training procedure–a well-known technique for correcting exposure bias–we introduce a new training objective that is continuous and differentiable everywhere and can provide informative gradients near points where previous decoding decisions change their value. By using a related approximation, we also demonstrate a similar approach to sampled-based training. We show that our approach outperforms both standard cross-entropy training and scheduled sampling procedures in two sequence prediction tasks: named entity recognition and machine translation.
%R 10.18653/v1/P17-2058
%U https://aclanthology.org/P17-2058/
%U https://doi.org/10.18653/v1/P17-2058
%P 366-371
Markdown (Informal)
[Differentiable Scheduled Sampling for Credit Assignment](https://aclanthology.org/P17-2058/) (Goyal et al., ACL 2017)
ACL
- Kartik Goyal, Chris Dyer, and Taylor Berg-Kirkpatrick. 2017. Differentiable Scheduled Sampling for Credit Assignment. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 366–371, Vancouver, Canada. Association for Computational Linguistics.