@inproceedings{bulla-etal-2023-towards,
title = "Towards Distribution-shift Robust Text Classification of Emotional Content",
author = "Bulla, Luana and
Gangemi, Aldo and
Mongiovi{'}, Misael",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Findings of the Association for Computational Linguistics: ACL 2023",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-acl.524",
doi = "10.18653/v1/2023.findings-acl.524",
pages = "8256--8268",
abstract = "Supervised models based on Transformers have been shown to achieve impressive performances in many natural language processing tasks. However, besides requiring a large amount of costly manually annotated data, supervised models tend to adapt to the characteristics of the training dataset, which are usually created ad-hoc and whose data distribution often differs from the one in real applications, showing significant performance degradation in real-world scenarios. We perform an extensive assessment of the out-of-distribution performances of supervised models for classification in the emotion and hate-speech detection tasks and show that NLI-based zero-shot models often outperform them, making task-specific annotation useless when the characteristics of final-user data are not known in advance. To benefit from both supervised and zero-shot approaches, we propose to fine-tune an NLI-based model on the task-specific dataset. The resulting model often outperforms all available supervised models both in distribution and out of distribution, with only a few thousand training samples.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="bulla-etal-2023-towards">
<titleInfo>
<title>Towards Distribution-shift Robust Text Classification of Emotional Content</title>
</titleInfo>
<name type="personal">
<namePart type="given">Luana</namePart>
<namePart type="family">Bulla</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aldo</namePart>
<namePart type="family">Gangemi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Misael</namePart>
<namePart type="family">Mongiovi’</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2023</title>
</titleInfo>
<name type="personal">
<namePart type="given">Anna</namePart>
<namePart type="family">Rogers</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jordan</namePart>
<namePart type="family">Boyd-Graber</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Naoaki</namePart>
<namePart type="family">Okazaki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Toronto, Canada</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Supervised models based on Transformers have been shown to achieve impressive performances in many natural language processing tasks. However, besides requiring a large amount of costly manually annotated data, supervised models tend to adapt to the characteristics of the training dataset, which are usually created ad-hoc and whose data distribution often differs from the one in real applications, showing significant performance degradation in real-world scenarios. We perform an extensive assessment of the out-of-distribution performances of supervised models for classification in the emotion and hate-speech detection tasks and show that NLI-based zero-shot models often outperform them, making task-specific annotation useless when the characteristics of final-user data are not known in advance. To benefit from both supervised and zero-shot approaches, we propose to fine-tune an NLI-based model on the task-specific dataset. The resulting model often outperforms all available supervised models both in distribution and out of distribution, with only a few thousand training samples.</abstract>
<identifier type="citekey">bulla-etal-2023-towards</identifier>
<identifier type="doi">10.18653/v1/2023.findings-acl.524</identifier>
<location>
<url>https://aclanthology.org/2023.findings-acl.524</url>
</location>
<part>
<date>2023-07</date>
<extent unit="page">
<start>8256</start>
<end>8268</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Towards Distribution-shift Robust Text Classification of Emotional Content
%A Bulla, Luana
%A Gangemi, Aldo
%A Mongiovi’, Misael
%Y Rogers, Anna
%Y Boyd-Graber, Jordan
%Y Okazaki, Naoaki
%S Findings of the Association for Computational Linguistics: ACL 2023
%D 2023
%8 July
%I Association for Computational Linguistics
%C Toronto, Canada
%F bulla-etal-2023-towards
%X Supervised models based on Transformers have been shown to achieve impressive performances in many natural language processing tasks. However, besides requiring a large amount of costly manually annotated data, supervised models tend to adapt to the characteristics of the training dataset, which are usually created ad-hoc and whose data distribution often differs from the one in real applications, showing significant performance degradation in real-world scenarios. We perform an extensive assessment of the out-of-distribution performances of supervised models for classification in the emotion and hate-speech detection tasks and show that NLI-based zero-shot models often outperform them, making task-specific annotation useless when the characteristics of final-user data are not known in advance. To benefit from both supervised and zero-shot approaches, we propose to fine-tune an NLI-based model on the task-specific dataset. The resulting model often outperforms all available supervised models both in distribution and out of distribution, with only a few thousand training samples.
%R 10.18653/v1/2023.findings-acl.524
%U https://aclanthology.org/2023.findings-acl.524
%U https://doi.org/10.18653/v1/2023.findings-acl.524
%P 8256-8268
Markdown (Informal)
[Towards Distribution-shift Robust Text Classification of Emotional Content](https://aclanthology.org/2023.findings-acl.524) (Bulla et al., Findings 2023)
ACL