Saved in:
Bibliographic Details
Main Authors: Xue, Chenhao, Jin, Yuanzhe, Carrasco-Revilla, Adrian, Chakraborty, Joyraj, Chen, Min
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2508.10000
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909736531656704
author Xue, Chenhao
Jin, Yuanzhe
Carrasco-Revilla, Adrian
Chakraborty, Joyraj
Chen, Min
author_facet Xue, Chenhao
Jin, Yuanzhe
Carrasco-Revilla, Adrian
Chakraborty, Joyraj
Chen, Min
contents When developing text classification models for real world applications, one major challenge is the difficulty to collect sufficient data for all text classes. In this work, we address this challenge by utilizing large language models (LLMs) to generate synthetic data and using such data to improve the performance of the models without waiting for more real data to be collected and labelled. As an LLM generates different synthetic data in response to different input examples, we formulate an automated workflow, which searches for input examples that lead to more ``effective'' synthetic data for improving the model concerned. We study three search strategies with an extensive set of experiments, and use experiment results to inform an ensemble algorithm that selects a search strategy according to the characteristics of a class. Our further experiments demonstrate that this ensemble approach is more effective than each individual strategy in our automated workflow for improving classification models using LLMs.
format Preprint
id arxiv_https___arxiv_org_abs_2508_10000
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle AutoGeTS: Knowledge-based Automated Generation of Text Synthetics for Improving Text Classification
Xue, Chenhao
Jin, Yuanzhe
Carrasco-Revilla, Adrian
Chakraborty, Joyraj
Chen, Min
Computation and Language
Machine Learning
When developing text classification models for real world applications, one major challenge is the difficulty to collect sufficient data for all text classes. In this work, we address this challenge by utilizing large language models (LLMs) to generate synthetic data and using such data to improve the performance of the models without waiting for more real data to be collected and labelled. As an LLM generates different synthetic data in response to different input examples, we formulate an automated workflow, which searches for input examples that lead to more ``effective'' synthetic data for improving the model concerned. We study three search strategies with an extensive set of experiments, and use experiment results to inform an ensemble algorithm that selects a search strategy according to the characteristics of a class. Our further experiments demonstrate that this ensemble approach is more effective than each individual strategy in our automated workflow for improving classification models using LLMs.
title AutoGeTS: Knowledge-based Automated Generation of Text Synthetics for Improving Text Classification
topic Computation and Language
Machine Learning
url https://arxiv.org/abs/2508.10000