@inproceedings{kumar-etal-2021-interpreting,
title = "Interpreting Text Classifiers by Learning Context-sensitive Influence of Words",
author = "Kumar, Sawan and
Dixit, Kalpit and
Shah, Kashif",
editor = "Pruksachatkun, Yada and
Ramakrishna, Anil and
Chang, Kai-Wei and
Krishna, Satyapriya and
Dhamala, Jwala and
Guha, Tanaya and
Ren, Xiang",
booktitle = "Proceedings of the First Workshop on Trustworthy Natural Language Processing",
month = jun,
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2021.trustnlp-1.7/",
doi = "10.18653/v1/2021.trustnlp-1.7",
pages = "55--67",
abstract = "Many existing approaches for interpreting text classification models focus on providing importance scores for parts of the input text, such as words, but without a way to test or improve the interpretation method itself. This has the effect of compounding the problem of understanding or building trust in the model, with the interpretation method itself adding to the opacity of the model. Further, importance scores on individual examples are usually not enough to provide a sufficient picture of model behavior. To address these concerns, we propose MOXIE (MOdeling conteXt-sensitive InfluencE of words) with an aim to enable a richer interface for a user to interact with the model being interpreted and to produce testable predictions. In particular, we aim to make predictions for importance scores, counterfactuals and learned biases with MOXIE. In addition, with a global learning objective, MOXIE provides a clear path for testing and improving itself. We evaluate the reliability and efficiency of MOXIE on the task of sentiment analysis."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="kumar-etal-2021-interpreting">
<titleInfo>
<title>Interpreting Text Classifiers by Learning Context-sensitive Influence of Words</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sawan</namePart>
<namePart type="family">Kumar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kalpit</namePart>
<namePart type="family">Dixit</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kashif</namePart>
<namePart type="family">Shah</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2021-06</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the First Workshop on Trustworthy Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yada</namePart>
<namePart type="family">Pruksachatkun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Anil</namePart>
<namePart type="family">Ramakrishna</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kai-Wei</namePart>
<namePart type="family">Chang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Satyapriya</namePart>
<namePart type="family">Krishna</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jwala</namePart>
<namePart type="family">Dhamala</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tanaya</namePart>
<namePart type="family">Guha</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiang</namePart>
<namePart type="family">Ren</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>Many existing approaches for interpreting text classification models focus on providing importance scores for parts of the input text, such as words, but without a way to test or improve the interpretation method itself. This has the effect of compounding the problem of understanding or building trust in the model, with the interpretation method itself adding to the opacity of the model. Further, importance scores on individual examples are usually not enough to provide a sufficient picture of model behavior. To address these concerns, we propose MOXIE (MOdeling conteXt-sensitive InfluencE of words) with an aim to enable a richer interface for a user to interact with the model being interpreted and to produce testable predictions. In particular, we aim to make predictions for importance scores, counterfactuals and learned biases with MOXIE. In addition, with a global learning objective, MOXIE provides a clear path for testing and improving itself. We evaluate the reliability and efficiency of MOXIE on the task of sentiment analysis.</abstract>
<identifier type="citekey">kumar-etal-2021-interpreting</identifier>
<identifier type="doi">10.18653/v1/2021.trustnlp-1.7</identifier>
<location>
<url>https://aclanthology.org/2021.trustnlp-1.7/</url>
</location>
<part>
<date>2021-06</date>
<extent unit="page">
<start>55</start>
<end>67</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Interpreting Text Classifiers by Learning Context-sensitive Influence of Words
%A Kumar, Sawan
%A Dixit, Kalpit
%A Shah, Kashif
%Y Pruksachatkun, Yada
%Y Ramakrishna, Anil
%Y Chang, Kai-Wei
%Y Krishna, Satyapriya
%Y Dhamala, Jwala
%Y Guha, Tanaya
%Y Ren, Xiang
%S Proceedings of the First Workshop on Trustworthy Natural Language Processing
%D 2021
%8 June
%I Association for Computational Linguistics
%C Online
%F kumar-etal-2021-interpreting
%X Many existing approaches for interpreting text classification models focus on providing importance scores for parts of the input text, such as words, but without a way to test or improve the interpretation method itself. This has the effect of compounding the problem of understanding or building trust in the model, with the interpretation method itself adding to the opacity of the model. Further, importance scores on individual examples are usually not enough to provide a sufficient picture of model behavior. To address these concerns, we propose MOXIE (MOdeling conteXt-sensitive InfluencE of words) with an aim to enable a richer interface for a user to interact with the model being interpreted and to produce testable predictions. In particular, we aim to make predictions for importance scores, counterfactuals and learned biases with MOXIE. In addition, with a global learning objective, MOXIE provides a clear path for testing and improving itself. We evaluate the reliability and efficiency of MOXIE on the task of sentiment analysis.
%R 10.18653/v1/2021.trustnlp-1.7
%U https://aclanthology.org/2021.trustnlp-1.7/
%U https://doi.org/10.18653/v1/2021.trustnlp-1.7
%P 55-67
Markdown (Informal)
[Interpreting Text Classifiers by Learning Context-sensitive Influence of Words](https://aclanthology.org/2021.trustnlp-1.7/) (Kumar et al., TrustNLP 2021)
ACL