@inproceedings{chemmengath-etal-2022-cat,
title = "Let the {CAT} out of the bag: Contrastive Attributed explanations for Text",
author = "Chemmengath, Saneem and
Azad, Amar Prakash and
Luss, Ronny and
Dhurandhar, Amit",
editor = "Goldberg, Yoav and
Kozareva, Zornitsa and
Zhang, Yue",
booktitle = "Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing",
month = dec,
year = "2022",
address = "Abu Dhabi, United Arab Emirates",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.emnlp-main.484/",
doi = "10.18653/v1/2022.emnlp-main.484",
pages = "7190--7206",
abstract = "Contrastive explanations for understanding the behavior of black box models has gained a lot of attention recently as they provide potential for recourse. In this paper, we propose a method Contrastive Attributed explanations for Text (CAT) which provides contrastive explanations for natural language text data with a novel twist as we build and exploit attribute classifiers leading to more semantically meaningful explanations. To ensure that our contrastive generated text has the fewest possible edits with respect to the original text, while also being fluent and close to a human generated contrastive, we resort to a minimal perturbation approach regularized using a BERT language model and attribute classifiers trained on available attributes. We show through qualitative examples and a user study that our method not only conveys more insight because of these attributes, but also leads to better quality (contrastive) text. Quantitatively, we show that our method outperforms other state-of-the-art methods across four data sets on four benchmark metrics."
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chemmengath-etal-2022-cat">
<titleInfo>
<title>Let the CAT out of the bag: Contrastive Attributed explanations for Text</title>
</titleInfo>
<name type="personal">
<namePart type="given">Saneem</namePart>
<namePart type="family">Chemmengath</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Amar</namePart>
<namePart type="given">Prakash</namePart>
<namePart type="family">Azad</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ronny</namePart>
<namePart type="family">Luss</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Amit</namePart>
<namePart type="family">Dhurandhar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2022-12</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yoav</namePart>
<namePart type="family">Goldberg</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zornitsa</namePart>
<namePart type="family">Kozareva</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Abu Dhabi, United Arab Emirates</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Contrastive explanations for understanding the behavior of black box models has gained a lot of attention recently as they provide potential for recourse. In this paper, we propose a method Contrastive Attributed explanations for Text (CAT) which provides contrastive explanations for natural language text data with a novel twist as we build and exploit attribute classifiers leading to more semantically meaningful explanations. To ensure that our contrastive generated text has the fewest possible edits with respect to the original text, while also being fluent and close to a human generated contrastive, we resort to a minimal perturbation approach regularized using a BERT language model and attribute classifiers trained on available attributes. We show through qualitative examples and a user study that our method not only conveys more insight because of these attributes, but also leads to better quality (contrastive) text. Quantitatively, we show that our method outperforms other state-of-the-art methods across four data sets on four benchmark metrics.</abstract>
<identifier type="citekey">chemmengath-etal-2022-cat</identifier>
<identifier type="doi">10.18653/v1/2022.emnlp-main.484</identifier>
<location>
<url>https://aclanthology.org/2022.emnlp-main.484/</url>
</location>
<part>
<date>2022-12</date>
<extent unit="page">
<start>7190</start>
<end>7206</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Let the CAT out of the bag: Contrastive Attributed explanations for Text
%A Chemmengath, Saneem
%A Azad, Amar Prakash
%A Luss, Ronny
%A Dhurandhar, Amit
%Y Goldberg, Yoav
%Y Kozareva, Zornitsa
%Y Zhang, Yue
%S Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing
%D 2022
%8 December
%I Association for Computational Linguistics
%C Abu Dhabi, United Arab Emirates
%F chemmengath-etal-2022-cat
%X Contrastive explanations for understanding the behavior of black box models has gained a lot of attention recently as they provide potential for recourse. In this paper, we propose a method Contrastive Attributed explanations for Text (CAT) which provides contrastive explanations for natural language text data with a novel twist as we build and exploit attribute classifiers leading to more semantically meaningful explanations. To ensure that our contrastive generated text has the fewest possible edits with respect to the original text, while also being fluent and close to a human generated contrastive, we resort to a minimal perturbation approach regularized using a BERT language model and attribute classifiers trained on available attributes. We show through qualitative examples and a user study that our method not only conveys more insight because of these attributes, but also leads to better quality (contrastive) text. Quantitatively, we show that our method outperforms other state-of-the-art methods across four data sets on four benchmark metrics.
%R 10.18653/v1/2022.emnlp-main.484
%U https://aclanthology.org/2022.emnlp-main.484/
%U https://doi.org/10.18653/v1/2022.emnlp-main.484
%P 7190-7206
Markdown (Informal)
[Let the CAT out of the bag: Contrastive Attributed explanations for Text](https://aclanthology.org/2022.emnlp-main.484/) (Chemmengath et al., EMNLP 2022)
ACL