RESEACH ON TEXT SUMMARIZATION AND SOME EFFECTIVE SUMMARIZATION METHODS

Abstract

This study examines extractive and abstractive text summarization, emphasizing deep learning techniques such as Transformer architectures and reinforcement learning. Extractive summarization selects key sentences, while abstractive summarization generates human-like summaries by rephrasing content. Key datasets like Vietnews and Wikilingua are highlighted for their role in training models for low-resource languages like Vietnamese. The research addresses challenges in maintaining coherence and semantic accuracy, proposing solutions to enhance summarization quality. Future directions include improving evaluation metrics, refining coherence in Vietnamese summaries, and advancing multilingual models. By integrating modern techniques and addressing key challenges, this study contributes to the development of more accurate and reliable automatic summarization systems.