Year: 2024
Type: article
Abstract: Text summarization plays a pivotal role in managing and distilling vast amounts of textual information into concise, coherent summaries.This survey explores the landscape of text summarization techniques, categorizing them into extractive and abstractive methods.Extractive techniques, such as graph-based algorithms and feature-based models like Text Rank and Lex Rank, focus on selecting salient sentences directly from the original text.In contrast, abstractive methods leverage advanced natural language processing (NLP) models, including neural networks and sequence-to-sequence architectures, to generate summaries that go beyond mere extraction by synthesizing new phrases.The article delves into evaluation metrics like ROUGE and BLEU, discussing their role in assessing summary quality, alongside challenges such as semantic coherence and scalability.Applications across domains like news media, academic research, and business are examined, highlighting the transformative impact of summarization on information retrieval and decision-making processes.Recent advances in deep learning, multimodal summarization, and ethical considerations in algorithmic design are also discussed, paving the way for future research directions.This survey consolidates current knowledge, offering insights into the evolving field of text summarization and its promising avenues for innovation.
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4404345423