@inproceedings{gu-etal-2023-critical,
title = "A Critical Analysis of Document Out-of-Distribution Detection",
author = "Gu, Jiuxiang and
Ming, Yifei and
Zhou, Yi and
Kuen, Jason and
Morariu, Vlad and
Zhao, Handong and
Zhang, Ruiyi and
Barmpalios, Nikolaos and
Liu, Anqi and
Li, Yixuan and
Sun, Tong and
Nenkova, Ani",
editor = "Bouamor, Houda and
Pino, Juan and
Bali, Kalika",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
month = dec,
year = "2023",
address = "Singapore",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-emnlp.332",
doi = "10.18653/v1/2023.findings-emnlp.332",
pages = "4973--4999",
abstract = "Large-scale pre-training is widely used in recent document understanding tasks. During deployment, one may expect that models should trigger a conservative fallback policy when encountering out-of-distribution (OOD) samples, which highlights the importance of OOD detection. However, most existing OOD detection methods focus on single-modal inputs such as images or texts. While documents are multi-modal in nature, it is underexplored if and how multi-modal information in documents can be exploited for OOD detection. In this work, we first provide a systematic and in-depth analysis on OOD detection for document understanding models. We study the effects of model modality, pre-training, and fine-tuning across various types of OOD inputs. In particular, we find that spatial information is critical for document OOD detection. To better exploit spatial information, we propose a spatial-aware adapter, which serves as a parameter-efficient add-on module to adapt transformer-based language models to the document domain. Extensive experiments show that adding the spatial-aware adapter significantly improves the OOD detection performance compared to directly using the language model and achieves superior performance compared to competitive baselines.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="gu-etal-2023-critical">
<titleInfo>
<title>A Critical Analysis of Document Out-of-Distribution Detection</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jiuxiang</namePart>
<namePart type="family">Gu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yifei</namePart>
<namePart type="family">Ming</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yi</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jason</namePart>
<namePart type="family">Kuen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Vlad</namePart>
<namePart type="family">Morariu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Handong</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruiyi</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nikolaos</namePart>
<namePart type="family">Barmpalios</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Anqi</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yixuan</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tong</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ani</namePart>
<namePart type="family">Nenkova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-12</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2023</title>
</titleInfo>
<name type="personal">
<namePart type="given">Houda</namePart>
<namePart type="family">Bouamor</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Juan</namePart>
<namePart type="family">Pino</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kalika</namePart>
<namePart type="family">Bali</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Singapore</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Large-scale pre-training is widely used in recent document understanding tasks. During deployment, one may expect that models should trigger a conservative fallback policy when encountering out-of-distribution (OOD) samples, which highlights the importance of OOD detection. However, most existing OOD detection methods focus on single-modal inputs such as images or texts. While documents are multi-modal in nature, it is underexplored if and how multi-modal information in documents can be exploited for OOD detection. In this work, we first provide a systematic and in-depth analysis on OOD detection for document understanding models. We study the effects of model modality, pre-training, and fine-tuning across various types of OOD inputs. In particular, we find that spatial information is critical for document OOD detection. To better exploit spatial information, we propose a spatial-aware adapter, which serves as a parameter-efficient add-on module to adapt transformer-based language models to the document domain. Extensive experiments show that adding the spatial-aware adapter significantly improves the OOD detection performance compared to directly using the language model and achieves superior performance compared to competitive baselines.</abstract>
<identifier type="citekey">gu-etal-2023-critical</identifier>
<identifier type="doi">10.18653/v1/2023.findings-emnlp.332</identifier>
<location>
<url>https://aclanthology.org/2023.findings-emnlp.332</url>
</location>
<part>
<date>2023-12</date>
<extent unit="page">
<start>4973</start>
<end>4999</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T A Critical Analysis of Document Out-of-Distribution Detection
%A Gu, Jiuxiang
%A Ming, Yifei
%A Zhou, Yi
%A Kuen, Jason
%A Morariu, Vlad
%A Zhao, Handong
%A Zhang, Ruiyi
%A Barmpalios, Nikolaos
%A Liu, Anqi
%A Li, Yixuan
%A Sun, Tong
%A Nenkova, Ani
%Y Bouamor, Houda
%Y Pino, Juan
%Y Bali, Kalika
%S Findings of the Association for Computational Linguistics: EMNLP 2023
%D 2023
%8 December
%I Association for Computational Linguistics
%C Singapore
%F gu-etal-2023-critical
%X Large-scale pre-training is widely used in recent document understanding tasks. During deployment, one may expect that models should trigger a conservative fallback policy when encountering out-of-distribution (OOD) samples, which highlights the importance of OOD detection. However, most existing OOD detection methods focus on single-modal inputs such as images or texts. While documents are multi-modal in nature, it is underexplored if and how multi-modal information in documents can be exploited for OOD detection. In this work, we first provide a systematic and in-depth analysis on OOD detection for document understanding models. We study the effects of model modality, pre-training, and fine-tuning across various types of OOD inputs. In particular, we find that spatial information is critical for document OOD detection. To better exploit spatial information, we propose a spatial-aware adapter, which serves as a parameter-efficient add-on module to adapt transformer-based language models to the document domain. Extensive experiments show that adding the spatial-aware adapter significantly improves the OOD detection performance compared to directly using the language model and achieves superior performance compared to competitive baselines.
%R 10.18653/v1/2023.findings-emnlp.332
%U https://aclanthology.org/2023.findings-emnlp.332
%U https://doi.org/10.18653/v1/2023.findings-emnlp.332
%P 4973-4999
Markdown (Informal)
[A Critical Analysis of Document Out-of-Distribution Detection](https://aclanthology.org/2023.findings-emnlp.332) (Gu et al., Findings 2023)
ACL
- Jiuxiang Gu, Yifei Ming, Yi Zhou, Jason Kuen, Vlad Morariu, Handong Zhao, Ruiyi Zhang, Nikolaos Barmpalios, Anqi Liu, Yixuan Li, Tong Sun, and Ani Nenkova. 2023. A Critical Analysis of Document Out-of-Distribution Detection. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 4973–4999, Singapore. Association for Computational Linguistics.