@inproceedings{yu-etal-2022-interventional,
title = "Interventional Training for Out-Of-Distribution Natural Language Understanding",
author = "Yu, Sicheng and
Jiang, Jing and
Zhang, Hao and
Niu, Yulei and
Sun, Qianru and
Bing, Lidong",
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.799",
doi = "10.18653/v1/2022.emnlp-main.799",
pages = "11627--11638",
abstract = "Out-of-distribution (OOD) settings are used to measure a model{'}s performance when the distribution of the test data is different from that of the training data. NLU models are known to suffer in OOD. We study this issue from the perspective of causality, which sees confounding bias as the reason for models to learn spurious correlations. While a common solution is to perform intervention, existing methods handle only known and single confounder, but in many NLU tasks the confounders can be both unknown and multifactorial. In this paper, we propose a novel interventional training method called Bottom-up Automatic Intervention (BAI) that performs multi-granular intervention with identified multifactorial confounders. Our experiments on three NLU tasks, namely, natural language inference, fact verification and paraphrase identification, show the effectiveness of BAI for tackling OOD settings.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yu-etal-2022-interventional">
<titleInfo>
<title>Interventional Training for Out-Of-Distribution Natural Language Understanding</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sicheng</namePart>
<namePart type="family">Yu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jing</namePart>
<namePart type="family">Jiang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hao</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yulei</namePart>
<namePart type="family">Niu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qianru</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lidong</namePart>
<namePart type="family">Bing</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>Out-of-distribution (OOD) settings are used to measure a model’s performance when the distribution of the test data is different from that of the training data. NLU models are known to suffer in OOD. We study this issue from the perspective of causality, which sees confounding bias as the reason for models to learn spurious correlations. While a common solution is to perform intervention, existing methods handle only known and single confounder, but in many NLU tasks the confounders can be both unknown and multifactorial. In this paper, we propose a novel interventional training method called Bottom-up Automatic Intervention (BAI) that performs multi-granular intervention with identified multifactorial confounders. Our experiments on three NLU tasks, namely, natural language inference, fact verification and paraphrase identification, show the effectiveness of BAI for tackling OOD settings.</abstract>
<identifier type="citekey">yu-etal-2022-interventional</identifier>
<identifier type="doi">10.18653/v1/2022.emnlp-main.799</identifier>
<location>
<url>https://aclanthology.org/2022.emnlp-main.799</url>
</location>
<part>
<date>2022-12</date>
<extent unit="page">
<start>11627</start>
<end>11638</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Interventional Training for Out-Of-Distribution Natural Language Understanding
%A Yu, Sicheng
%A Jiang, Jing
%A Zhang, Hao
%A Niu, Yulei
%A Sun, Qianru
%A Bing, Lidong
%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 yu-etal-2022-interventional
%X Out-of-distribution (OOD) settings are used to measure a model’s performance when the distribution of the test data is different from that of the training data. NLU models are known to suffer in OOD. We study this issue from the perspective of causality, which sees confounding bias as the reason for models to learn spurious correlations. While a common solution is to perform intervention, existing methods handle only known and single confounder, but in many NLU tasks the confounders can be both unknown and multifactorial. In this paper, we propose a novel interventional training method called Bottom-up Automatic Intervention (BAI) that performs multi-granular intervention with identified multifactorial confounders. Our experiments on three NLU tasks, namely, natural language inference, fact verification and paraphrase identification, show the effectiveness of BAI for tackling OOD settings.
%R 10.18653/v1/2022.emnlp-main.799
%U https://aclanthology.org/2022.emnlp-main.799
%U https://doi.org/10.18653/v1/2022.emnlp-main.799
%P 11627-11638
Markdown (Informal)
[Interventional Training for Out-Of-Distribution Natural Language Understanding](https://aclanthology.org/2022.emnlp-main.799) (Yu et al., EMNLP 2022)
ACL