@inproceedings{moradi-etal-2021-measuring,
title = "Measuring and Improving Faithfulness of Attention in Neural Machine Translation",
author = "Moradi, Pooya and
Kambhatla, Nishant and
Sarkar, Anoop",
editor = "Merlo, Paola and
Tiedemann, Jorg and
Tsarfaty, Reut",
booktitle = "Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume",
month = apr,
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2021.eacl-main.243",
doi = "10.18653/v1/2021.eacl-main.243",
pages = "2791--2802",
abstract = "While the attention heatmaps produced by neural machine translation (NMT) models seem insightful, there is little evidence that they reflect a model{'}s true internal reasoning. We provide a measure of faithfulness for NMT based on a variety of stress tests where attention weights which are crucial for prediction are perturbed and the model should alter its predictions if the learned weights are a faithful explanation of the predictions. We show that our proposed faithfulness measure for NMT models can be improved using a novel differentiable objective that rewards faithful behaviour by the model through probability divergence. Our experimental results on multiple language pairs show that our objective function is effective in increasing faithfulness and can lead to a useful analysis of NMT model behaviour and more trustworthy attention heatmaps. Our proposed objective improves faithfulness without reducing the translation quality and has a useful regularization effect on the NMT model and can even improve translation quality in some cases.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="moradi-etal-2021-measuring">
<titleInfo>
<title>Measuring and Improving Faithfulness of Attention in Neural Machine Translation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Pooya</namePart>
<namePart type="family">Moradi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nishant</namePart>
<namePart type="family">Kambhatla</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Anoop</namePart>
<namePart type="family">Sarkar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2021-04</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume</title>
</titleInfo>
<name type="personal">
<namePart type="given">Paola</namePart>
<namePart type="family">Merlo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jorg</namePart>
<namePart type="family">Tiedemann</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Reut</namePart>
<namePart type="family">Tsarfaty</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Online</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>While the attention heatmaps produced by neural machine translation (NMT) models seem insightful, there is little evidence that they reflect a model’s true internal reasoning. We provide a measure of faithfulness for NMT based on a variety of stress tests where attention weights which are crucial for prediction are perturbed and the model should alter its predictions if the learned weights are a faithful explanation of the predictions. We show that our proposed faithfulness measure for NMT models can be improved using a novel differentiable objective that rewards faithful behaviour by the model through probability divergence. Our experimental results on multiple language pairs show that our objective function is effective in increasing faithfulness and can lead to a useful analysis of NMT model behaviour and more trustworthy attention heatmaps. Our proposed objective improves faithfulness without reducing the translation quality and has a useful regularization effect on the NMT model and can even improve translation quality in some cases.</abstract>
<identifier type="citekey">moradi-etal-2021-measuring</identifier>
<identifier type="doi">10.18653/v1/2021.eacl-main.243</identifier>
<location>
<url>https://aclanthology.org/2021.eacl-main.243</url>
</location>
<part>
<date>2021-04</date>
<extent unit="page">
<start>2791</start>
<end>2802</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Measuring and Improving Faithfulness of Attention in Neural Machine Translation
%A Moradi, Pooya
%A Kambhatla, Nishant
%A Sarkar, Anoop
%Y Merlo, Paola
%Y Tiedemann, Jorg
%Y Tsarfaty, Reut
%S Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume
%D 2021
%8 April
%I Association for Computational Linguistics
%C Online
%F moradi-etal-2021-measuring
%X While the attention heatmaps produced by neural machine translation (NMT) models seem insightful, there is little evidence that they reflect a model’s true internal reasoning. We provide a measure of faithfulness for NMT based on a variety of stress tests where attention weights which are crucial for prediction are perturbed and the model should alter its predictions if the learned weights are a faithful explanation of the predictions. We show that our proposed faithfulness measure for NMT models can be improved using a novel differentiable objective that rewards faithful behaviour by the model through probability divergence. Our experimental results on multiple language pairs show that our objective function is effective in increasing faithfulness and can lead to a useful analysis of NMT model behaviour and more trustworthy attention heatmaps. Our proposed objective improves faithfulness without reducing the translation quality and has a useful regularization effect on the NMT model and can even improve translation quality in some cases.
%R 10.18653/v1/2021.eacl-main.243
%U https://aclanthology.org/2021.eacl-main.243
%U https://doi.org/10.18653/v1/2021.eacl-main.243
%P 2791-2802
Markdown (Informal)
[Measuring and Improving Faithfulness of Attention in Neural Machine Translation](https://aclanthology.org/2021.eacl-main.243) (Moradi et al., EACL 2021)
ACL