@inproceedings{duan-etal-2022-barle,
title = "{BARLE}: Background-Aware Representation Learning for Background Shift Out-of-Distribution Detection",
author = "Duan, Hanyu and
Yang, Yi and
Abbasi, Ahmed and
Tam, Kar Yan",
editor = "Goldberg, Yoav and
Kozareva, Zornitsa and
Zhang, Yue",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2022",
month = dec,
year = "2022",
address = "Abu Dhabi, United Arab Emirates",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.findings-emnlp.53",
doi = "10.18653/v1/2022.findings-emnlp.53",
pages = "750--764",
abstract = "Machine learning models often suffer from a performance drop when they are applied to out-of-distribution (OOD) samples, i.e., those drawn far away from the training data distribution. Existing OOD detection work mostly focuses on identifying semantic-shift OOD samples, e.g., instances from unseen new classes. However, background-shift OOD detection, which identifies samples with domain or style-change, represents a more practical yet challenging task. In this paper, we propose Background-Aware Representation Learning (BARLE) for background-shift OOD detection in NLP. Specifically, we generate semantics-preserving background-shifted pseudo OOD samples from pretrained masked language models. We then contrast the in-distribution (ID) samples with their pseudo OOD counterparts. Unlike prior semantic-shift OOD detection work that often leverages an external text corpus, BARLE only uses ID data, which is more flexible and cost-efficient. In experiments across several text classification tasks, we demonstrate that BARLE is capable of improving background-shift OOD detection performance while maintaining ID classification accuracy. We further investigate the properties of the generated pseudo OOD samples, uncovering the working mechanism of BARLE.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="duan-etal-2022-barle">
<titleInfo>
<title>BARLE: Background-Aware Representation Learning for Background Shift Out-of-Distribution Detection</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hanyu</namePart>
<namePart type="family">Duan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yi</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ahmed</namePart>
<namePart type="family">Abbasi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kar</namePart>
<namePart type="given">Yan</namePart>
<namePart type="family">Tam</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>Findings of the Association for Computational Linguistics: EMNLP 2022</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>Machine learning models often suffer from a performance drop when they are applied to out-of-distribution (OOD) samples, i.e., those drawn far away from the training data distribution. Existing OOD detection work mostly focuses on identifying semantic-shift OOD samples, e.g., instances from unseen new classes. However, background-shift OOD detection, which identifies samples with domain or style-change, represents a more practical yet challenging task. In this paper, we propose Background-Aware Representation Learning (BARLE) for background-shift OOD detection in NLP. Specifically, we generate semantics-preserving background-shifted pseudo OOD samples from pretrained masked language models. We then contrast the in-distribution (ID) samples with their pseudo OOD counterparts. Unlike prior semantic-shift OOD detection work that often leverages an external text corpus, BARLE only uses ID data, which is more flexible and cost-efficient. In experiments across several text classification tasks, we demonstrate that BARLE is capable of improving background-shift OOD detection performance while maintaining ID classification accuracy. We further investigate the properties of the generated pseudo OOD samples, uncovering the working mechanism of BARLE.</abstract>
<identifier type="citekey">duan-etal-2022-barle</identifier>
<identifier type="doi">10.18653/v1/2022.findings-emnlp.53</identifier>
<location>
<url>https://aclanthology.org/2022.findings-emnlp.53</url>
</location>
<part>
<date>2022-12</date>
<extent unit="page">
<start>750</start>
<end>764</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T BARLE: Background-Aware Representation Learning for Background Shift Out-of-Distribution Detection
%A Duan, Hanyu
%A Yang, Yi
%A Abbasi, Ahmed
%A Tam, Kar Yan
%Y Goldberg, Yoav
%Y Kozareva, Zornitsa
%Y Zhang, Yue
%S Findings of the Association for Computational Linguistics: EMNLP 2022
%D 2022
%8 December
%I Association for Computational Linguistics
%C Abu Dhabi, United Arab Emirates
%F duan-etal-2022-barle
%X Machine learning models often suffer from a performance drop when they are applied to out-of-distribution (OOD) samples, i.e., those drawn far away from the training data distribution. Existing OOD detection work mostly focuses on identifying semantic-shift OOD samples, e.g., instances from unseen new classes. However, background-shift OOD detection, which identifies samples with domain or style-change, represents a more practical yet challenging task. In this paper, we propose Background-Aware Representation Learning (BARLE) for background-shift OOD detection in NLP. Specifically, we generate semantics-preserving background-shifted pseudo OOD samples from pretrained masked language models. We then contrast the in-distribution (ID) samples with their pseudo OOD counterparts. Unlike prior semantic-shift OOD detection work that often leverages an external text corpus, BARLE only uses ID data, which is more flexible and cost-efficient. In experiments across several text classification tasks, we demonstrate that BARLE is capable of improving background-shift OOD detection performance while maintaining ID classification accuracy. We further investigate the properties of the generated pseudo OOD samples, uncovering the working mechanism of BARLE.
%R 10.18653/v1/2022.findings-emnlp.53
%U https://aclanthology.org/2022.findings-emnlp.53
%U https://doi.org/10.18653/v1/2022.findings-emnlp.53
%P 750-764
Markdown (Informal)
[BARLE: Background-Aware Representation Learning for Background Shift Out-of-Distribution Detection](https://aclanthology.org/2022.findings-emnlp.53) (Duan et al., Findings 2022)
ACL