MLM: Masked Language Modeling Using Deep Learning for Efficient Summarization of Unstructured Data

The exponential growth of unstructured data in various fields has led to an increasing demand for efficient summarization techniques. This research paper aims to examine the existing approaches to summarize unstructured data and their application areas. The act of summarizing a text to extract its key informational components is known as summarization. In recent years, a number of techniques have been put out to extract key passages from text sources and display them in a condensed manner. The first difficulty with these techniques is identifying the ideas that effectively represent the text’s core idea and extracting sentences that do so. The abstractive summarization of biological materials is the main topic of this work. In the field of biomedical research, there is a growing need for efficient and accurate summarization techniques that can process the vast amount of data generated. This research paper proposes the construction of a new corpus for the biomedical domain and aims to determine the significant features that can be used to generate a summary of biomedical transcripts. The paper discusses the methodology used for corpus construction and identifies the features that can be used for summarization. In this context, we have implemented masked language modeling (MLM) and perform various tests with the PubMed databases from BioMed Central by generating an overall ROUGE of 74.80. The findings obtained demonstrate that the suggested strategy allows for a better interpretation of the key ideas and sentences of biological papers for obtaining abstractive summarization.