Research Area:  Machine Learning

Abstract:

Nowadays, voluminous unstructured data is steaming on the Web/social media. It is not easy for individuals to find relevant information quickly from such a vast unstructured corpus. Text summarizing is the process of retrieving relevant information in brief without changing the fundamental meaning of information. Manual text summarization takes much time, cost and effort to summarise. Researchers worked on different machine learning approaches to summarise the text in the past but still lagging for better-summarised results. In this paper, initially, we used reinforced learning (with attention layer) as the base model. We analysed its performance analysis, and later we proposed a deep learning-based model and evaluated using Bilingual Evaluation Understudy (BLEU) value of 0.4 and Recall-Oriented Understudy for Gisting Evaluation (ROUGE) with a value of 0.6 on a large corpus at sentence level on a standard dataset. The proposed model shows promising results of accuracy and correctness as compared to the state-of-the-art.

Keywords:  
Extractive Text Summarization
Deep Learning
Recall-Oriented Understudy for Gisting Evaluation (ROUGE)
reinforced learning
Machine Learning

Author(s) Name:  Arun Kumar Yadav, Amit Singh, Mayank Dhiman, Vineet, Rishabh Kaundal, Ankit Verma & Divakar Yadav

Publisher name:  Springer

DOI:  10.1007/s41870-022-00863-7

Volume Information:  volume 14, pages2407–2415 (2022)

Paper Link:   https://link.springer.com/article/10.1007/s41870-022-00863-7