@inproceedings{li-zhang-2021-semi,
title = "Semi-supervised Meta-learning for Cross-domain Few-shot Intent Classification",
author = "Li, Yue and
Zhang, Jiong",
booktitle = "Proceedings of the 1st Workshop on Meta Learning and Its Applications to Natural Language Processing",
month = aug,
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2021.metanlp-1.8",
doi = "10.18653/v1/2021.metanlp-1.8",
pages = "67--75",
abstract = "Meta learning aims to optimize the model{'}s capability to generalize to new tasks and domains. Lacking a data-efficient way to create meta training tasks has prevented the application of meta-learning to the real-world few shot learning scenarios. Recent studies have proposed unsupervised approaches to create meta-training tasks from unlabeled data for free, e.g., the SMLMT method (Bansal et al., 2020a) constructs unsupervised multi-class classification tasks from the unlabeled text by randomly masking words in the sentence and let the meta learner choose which word to fill in the blank. This study proposes a semi-supervised meta-learning approach that incorporates both the representation power of large pre-trained language models and the generalization capability of prototypical networks enhanced by SMLMT. The semi-supervised meta training approach avoids overfitting prototypical networks on a small number of labeled training examples and quickly learns cross-domain task-specific representation only from a few supporting examples. By incorporating SMLMT with prototypical networks, the meta learner generalizes better to unseen domains and gains higher accuracy on out-of-scope examples without the heavy lifting of pre-training. We observe significant improvement in few-shot generalization after training only a few epochs on the intent classification tasks evaluated in a multi-domain setting.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="li-zhang-2021-semi">
<titleInfo>
<title>Semi-supervised Meta-learning for Cross-domain Few-shot Intent Classification</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiong</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2021-aug</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 1st Workshop on Meta Learning and Its Applications to Natural Language Processing</title>
</titleInfo>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Online</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Meta learning aims to optimize the model’s capability to generalize to new tasks and domains. Lacking a data-efficient way to create meta training tasks has prevented the application of meta-learning to the real-world few shot learning scenarios. Recent studies have proposed unsupervised approaches to create meta-training tasks from unlabeled data for free, e.g., the SMLMT method (Bansal et al., 2020a) constructs unsupervised multi-class classification tasks from the unlabeled text by randomly masking words in the sentence and let the meta learner choose which word to fill in the blank. This study proposes a semi-supervised meta-learning approach that incorporates both the representation power of large pre-trained language models and the generalization capability of prototypical networks enhanced by SMLMT. The semi-supervised meta training approach avoids overfitting prototypical networks on a small number of labeled training examples and quickly learns cross-domain task-specific representation only from a few supporting examples. By incorporating SMLMT with prototypical networks, the meta learner generalizes better to unseen domains and gains higher accuracy on out-of-scope examples without the heavy lifting of pre-training. We observe significant improvement in few-shot generalization after training only a few epochs on the intent classification tasks evaluated in a multi-domain setting.</abstract>
<identifier type="citekey">li-zhang-2021-semi</identifier>
<identifier type="doi">10.18653/v1/2021.metanlp-1.8</identifier>
<location>
<url>https://aclanthology.org/2021.metanlp-1.8</url>
</location>
<part>
<date>2021-aug</date>
<extent unit="page">
<start>67</start>
<end>75</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Semi-supervised Meta-learning for Cross-domain Few-shot Intent Classification
%A Li, Yue
%A Zhang, Jiong
%S Proceedings of the 1st Workshop on Meta Learning and Its Applications to Natural Language Processing
%D 2021
%8 aug
%I Association for Computational Linguistics
%C Online
%F li-zhang-2021-semi
%X Meta learning aims to optimize the model’s capability to generalize to new tasks and domains. Lacking a data-efficient way to create meta training tasks has prevented the application of meta-learning to the real-world few shot learning scenarios. Recent studies have proposed unsupervised approaches to create meta-training tasks from unlabeled data for free, e.g., the SMLMT method (Bansal et al., 2020a) constructs unsupervised multi-class classification tasks from the unlabeled text by randomly masking words in the sentence and let the meta learner choose which word to fill in the blank. This study proposes a semi-supervised meta-learning approach that incorporates both the representation power of large pre-trained language models and the generalization capability of prototypical networks enhanced by SMLMT. The semi-supervised meta training approach avoids overfitting prototypical networks on a small number of labeled training examples and quickly learns cross-domain task-specific representation only from a few supporting examples. By incorporating SMLMT with prototypical networks, the meta learner generalizes better to unseen domains and gains higher accuracy on out-of-scope examples without the heavy lifting of pre-training. We observe significant improvement in few-shot generalization after training only a few epochs on the intent classification tasks evaluated in a multi-domain setting.
%R 10.18653/v1/2021.metanlp-1.8
%U https://aclanthology.org/2021.metanlp-1.8
%U https://doi.org/10.18653/v1/2021.metanlp-1.8
%P 67-75
Markdown (Informal)
[Semi-supervised Meta-learning for Cross-domain Few-shot Intent Classification](https://aclanthology.org/2021.metanlp-1.8) (Li & Zhang, MetaNLP 2021)
ACL