@inproceedings{edwards-etal-2022-guiding,
title = "Guiding Generative Language Models for Data Augmentation in Few-Shot Text Classification",
author = "Edwards, Aleksandra and
Ushio, Asahi and
Camacho-collados, Jose and
Ribaupierre, Helene and
Preece, Alun",
editor = "Dragut, Eduard and
Li, Yunyao and
Popa, Lucian and
Vucetic, Slobodan and
Srivastava, Shashank",
booktitle = "Proceedings of the Fourth Workshop on Data Science with Human-in-the-Loop (Language Advances)",
month = dec,
year = "2022",
address = "Abu Dhabi, United Arab Emirates (Hybrid)",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.dash-1.8",
pages = "51--63",
abstract = "Data augmentation techniques are widely used for enhancing the performance of machine learning models by tackling class imbalance issues and data sparsity. State-of-the-art generative language models have been shown to provide significant gains across different NLP tasks. However, their applicability to data augmentation for text classification tasks in few-shot settings have not been fully explored, especially for specialised domains. In this paper, we leverage GPT-2 (Radford et al, 2019) for generating artificial training instances in order to improve classification performance. Our aim is to analyse the impact the selection process of seed training examples has over the quality of GPT-generated samples and consequently the classifier performance. We propose a human-in-the-loop approach for selecting seed samples. Further, we compare the approach to other seed selection strategies that exploit the characteristics of specialised domains such as human-created class hierarchical structure and the presence of noun phrases. Our results show that fine-tuning GPT-2 in a handful of label instances leads to consistent classification improvements and outperform competitive baselines. The seed selection strategies developed in this work lead to significant improvements over random seed selection for specialised domains. We show that guiding text generation through domain expert selection can lead to further improvements, which opens up interesting research avenues for combining generative models and active learning.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="edwards-etal-2022-guiding">
<titleInfo>
<title>Guiding Generative Language Models for Data Augmentation in Few-Shot Text Classification</title>
</titleInfo>
<name type="personal">
<namePart type="given">Aleksandra</namePart>
<namePart type="family">Edwards</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Asahi</namePart>
<namePart type="family">Ushio</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jose</namePart>
<namePart type="family">Camacho-collados</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Helene</namePart>
<namePart type="family">Ribaupierre</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alun</namePart>
<namePart type="family">Preece</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>Proceedings of the Fourth Workshop on Data Science with Human-in-the-Loop (Language Advances)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Eduard</namePart>
<namePart type="family">Dragut</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yunyao</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lucian</namePart>
<namePart type="family">Popa</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Slobodan</namePart>
<namePart type="family">Vucetic</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shashank</namePart>
<namePart type="family">Srivastava</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 (Hybrid)</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Data augmentation techniques are widely used for enhancing the performance of machine learning models by tackling class imbalance issues and data sparsity. State-of-the-art generative language models have been shown to provide significant gains across different NLP tasks. However, their applicability to data augmentation for text classification tasks in few-shot settings have not been fully explored, especially for specialised domains. In this paper, we leverage GPT-2 (Radford et al, 2019) for generating artificial training instances in order to improve classification performance. Our aim is to analyse the impact the selection process of seed training examples has over the quality of GPT-generated samples and consequently the classifier performance. We propose a human-in-the-loop approach for selecting seed samples. Further, we compare the approach to other seed selection strategies that exploit the characteristics of specialised domains such as human-created class hierarchical structure and the presence of noun phrases. Our results show that fine-tuning GPT-2 in a handful of label instances leads to consistent classification improvements and outperform competitive baselines. The seed selection strategies developed in this work lead to significant improvements over random seed selection for specialised domains. We show that guiding text generation through domain expert selection can lead to further improvements, which opens up interesting research avenues for combining generative models and active learning.</abstract>
<identifier type="citekey">edwards-etal-2022-guiding</identifier>
<location>
<url>https://aclanthology.org/2022.dash-1.8</url>
</location>
<part>
<date>2022-12</date>
<extent unit="page">
<start>51</start>
<end>63</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Guiding Generative Language Models for Data Augmentation in Few-Shot Text Classification
%A Edwards, Aleksandra
%A Ushio, Asahi
%A Camacho-collados, Jose
%A Ribaupierre, Helene
%A Preece, Alun
%Y Dragut, Eduard
%Y Li, Yunyao
%Y Popa, Lucian
%Y Vucetic, Slobodan
%Y Srivastava, Shashank
%S Proceedings of the Fourth Workshop on Data Science with Human-in-the-Loop (Language Advances)
%D 2022
%8 December
%I Association for Computational Linguistics
%C Abu Dhabi, United Arab Emirates (Hybrid)
%F edwards-etal-2022-guiding
%X Data augmentation techniques are widely used for enhancing the performance of machine learning models by tackling class imbalance issues and data sparsity. State-of-the-art generative language models have been shown to provide significant gains across different NLP tasks. However, their applicability to data augmentation for text classification tasks in few-shot settings have not been fully explored, especially for specialised domains. In this paper, we leverage GPT-2 (Radford et al, 2019) for generating artificial training instances in order to improve classification performance. Our aim is to analyse the impact the selection process of seed training examples has over the quality of GPT-generated samples and consequently the classifier performance. We propose a human-in-the-loop approach for selecting seed samples. Further, we compare the approach to other seed selection strategies that exploit the characteristics of specialised domains such as human-created class hierarchical structure and the presence of noun phrases. Our results show that fine-tuning GPT-2 in a handful of label instances leads to consistent classification improvements and outperform competitive baselines. The seed selection strategies developed in this work lead to significant improvements over random seed selection for specialised domains. We show that guiding text generation through domain expert selection can lead to further improvements, which opens up interesting research avenues for combining generative models and active learning.
%U https://aclanthology.org/2022.dash-1.8
%P 51-63
Markdown (Informal)
[Guiding Generative Language Models for Data Augmentation in Few-Shot Text Classification](https://aclanthology.org/2022.dash-1.8) (Edwards et al., DaSH 2022)
ACL