@inproceedings{karisani-2023-neural,
title = "Neural Networks Against (and For) Self-Training: Classification with Small Labeled and Large Unlabeled Sets",
author = "Karisani, Payam",
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.769",
doi = "10.18653/v1/2023.findings-acl.769",
pages = "12148--12162",
abstract = "We propose a semi-supervised text classifier based on self-training using one positive and one negative property of neural networks. One of the weaknesses of self-training is the semantic drift problem, where noisy pseudo-labels accumulate over iterations and consequently the error rate soars. In order to tackle this challenge, we reshape the role of pseudo-labels and create a hierarchical order of information. In addition, a crucial step in self-training is to use the classifier confidence prediction to select the best candidate pseudo-labels. This step cannot be efficiently done by neural networks, because it is known that their output is poorly calibrated. To overcome this challenge, we propose a hybrid metric to replace the plain confidence measurement. Our metric takes into account the prediction uncertainty via a subsampling technique. We evaluate our model in a set of five standard benchmarks, and show that it significantly outperforms a set of ten diverse baseline models. Furthermore, we show that the improvement achieved by our model is additive to language model pretraining, which is a widely used technique for using unlabeled documents.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="karisani-2023-neural">
<titleInfo>
<title>Neural Networks Against (and For) Self-Training: Classification with Small Labeled and Large Unlabeled Sets</title>
</titleInfo>
<name type="personal">
<namePart type="given">Payam</namePart>
<namePart type="family">Karisani</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>We propose a semi-supervised text classifier based on self-training using one positive and one negative property of neural networks. One of the weaknesses of self-training is the semantic drift problem, where noisy pseudo-labels accumulate over iterations and consequently the error rate soars. In order to tackle this challenge, we reshape the role of pseudo-labels and create a hierarchical order of information. In addition, a crucial step in self-training is to use the classifier confidence prediction to select the best candidate pseudo-labels. This step cannot be efficiently done by neural networks, because it is known that their output is poorly calibrated. To overcome this challenge, we propose a hybrid metric to replace the plain confidence measurement. Our metric takes into account the prediction uncertainty via a subsampling technique. We evaluate our model in a set of five standard benchmarks, and show that it significantly outperforms a set of ten diverse baseline models. Furthermore, we show that the improvement achieved by our model is additive to language model pretraining, which is a widely used technique for using unlabeled documents.</abstract>
<identifier type="citekey">karisani-2023-neural</identifier>
<identifier type="doi">10.18653/v1/2023.findings-acl.769</identifier>
<location>
<url>https://aclanthology.org/2023.findings-acl.769</url>
</location>
<part>
<date>2023-07</date>
<extent unit="page">
<start>12148</start>
<end>12162</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Neural Networks Against (and For) Self-Training: Classification with Small Labeled and Large Unlabeled Sets
%A Karisani, Payam
%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 karisani-2023-neural
%X We propose a semi-supervised text classifier based on self-training using one positive and one negative property of neural networks. One of the weaknesses of self-training is the semantic drift problem, where noisy pseudo-labels accumulate over iterations and consequently the error rate soars. In order to tackle this challenge, we reshape the role of pseudo-labels and create a hierarchical order of information. In addition, a crucial step in self-training is to use the classifier confidence prediction to select the best candidate pseudo-labels. This step cannot be efficiently done by neural networks, because it is known that their output is poorly calibrated. To overcome this challenge, we propose a hybrid metric to replace the plain confidence measurement. Our metric takes into account the prediction uncertainty via a subsampling technique. We evaluate our model in a set of five standard benchmarks, and show that it significantly outperforms a set of ten diverse baseline models. Furthermore, we show that the improvement achieved by our model is additive to language model pretraining, which is a widely used technique for using unlabeled documents.
%R 10.18653/v1/2023.findings-acl.769
%U https://aclanthology.org/2023.findings-acl.769
%U https://doi.org/10.18653/v1/2023.findings-acl.769
%P 12148-12162
Markdown (Informal)
[Neural Networks Against (and For) Self-Training: Classification with Small Labeled and Large Unlabeled Sets](https://aclanthology.org/2023.findings-acl.769) (Karisani, Findings 2023)
ACL