Data Augmentation with Large Language Models for Vietnamese Abstractive Text Summarization

Abstract

Text summarization plays a crucial role in managing the overwhelming volume of information available today. This task aims to condense large amounts of information into summaries. However, the lack of large-scale annotated data in certain languages, such as Vietnamese, poses a substantial challenge for developing effective summarization models. With the recent advancements in large language models, such as GPT-3.5, there is an opportunity to leverage these models to augment data for improving the performance of deep learning models in Vietnamese text summarization. In this paper, we propose an automatic approach that utilizes a large language model to generate additional training examples and to enhance the summarization process for Vietnamese texts.