@inproceedings{srikanth-rudinger-2022-partial,
title = "Partial-input baselines show that {NLI} models can ignore context, but they don`t.",
author = "Srikanth, Neha and
Rudinger, Rachel",
editor = "Carpuat, Marine and
de Marneffe, Marie-Catherine and
Meza Ruiz, Ivan Vladimir",
booktitle = "Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies",
month = jul,
year = "2022",
address = "Seattle, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.naacl-main.350/",
doi = "10.18653/v1/2022.naacl-main.350",
pages = "4753--4763",
abstract = "When strong partial-input baselines reveal artifacts in crowdsourced NLI datasets, the performance of full-input models trained on such datasets is often dismissed as reliance on spurious correlations. We investigate whether state-of-the-art NLI models are capable of overriding default inferences made by a partial-input baseline. We introduce an evaluation set of 600 examples consisting of perturbed premises to examine a RoBERTa model`s sensitivity to edited contexts. Our results indicate that NLI models are still capable of learning to condition on context{---}a necessary component of inferential reasoning{---}despite being trained on artifact-ridden datasets."
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="srikanth-rudinger-2022-partial">
<titleInfo>
<title>Partial-input baselines show that NLI models can ignore context, but they don‘t.</title>
</titleInfo>
<name type="personal">
<namePart type="given">Neha</namePart>
<namePart type="family">Srikanth</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rachel</namePart>
<namePart type="family">Rudinger</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2022-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies</title>
</titleInfo>
<name type="personal">
<namePart type="given">Marine</namePart>
<namePart type="family">Carpuat</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marie-Catherine</namePart>
<namePart type="family">de Marneffe</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ivan</namePart>
<namePart type="given">Vladimir</namePart>
<namePart type="family">Meza Ruiz</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Seattle, United States</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>When strong partial-input baselines reveal artifacts in crowdsourced NLI datasets, the performance of full-input models trained on such datasets is often dismissed as reliance on spurious correlations. We investigate whether state-of-the-art NLI models are capable of overriding default inferences made by a partial-input baseline. We introduce an evaluation set of 600 examples consisting of perturbed premises to examine a RoBERTa model‘s sensitivity to edited contexts. Our results indicate that NLI models are still capable of learning to condition on context—a necessary component of inferential reasoning—despite being trained on artifact-ridden datasets.</abstract>
<identifier type="citekey">srikanth-rudinger-2022-partial</identifier>
<identifier type="doi">10.18653/v1/2022.naacl-main.350</identifier>
<location>
<url>https://aclanthology.org/2022.naacl-main.350/</url>
</location>
<part>
<date>2022-07</date>
<extent unit="page">
<start>4753</start>
<end>4763</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Partial-input baselines show that NLI models can ignore context, but they don‘t.
%A Srikanth, Neha
%A Rudinger, Rachel
%Y Carpuat, Marine
%Y de Marneffe, Marie-Catherine
%Y Meza Ruiz, Ivan Vladimir
%S Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies
%D 2022
%8 July
%I Association for Computational Linguistics
%C Seattle, United States
%F srikanth-rudinger-2022-partial
%X When strong partial-input baselines reveal artifacts in crowdsourced NLI datasets, the performance of full-input models trained on such datasets is often dismissed as reliance on spurious correlations. We investigate whether state-of-the-art NLI models are capable of overriding default inferences made by a partial-input baseline. We introduce an evaluation set of 600 examples consisting of perturbed premises to examine a RoBERTa model‘s sensitivity to edited contexts. Our results indicate that NLI models are still capable of learning to condition on context—a necessary component of inferential reasoning—despite being trained on artifact-ridden datasets.
%R 10.18653/v1/2022.naacl-main.350
%U https://aclanthology.org/2022.naacl-main.350/
%U https://doi.org/10.18653/v1/2022.naacl-main.350
%P 4753-4763
Markdown (Informal)
[Partial-input baselines show that NLI models can ignore context, but they don’t.](https://aclanthology.org/2022.naacl-main.350/) (Srikanth & Rudinger, NAACL 2022)
ACL