An Exploratory Assessment on Text Summarization

Text summarization involves creating a short, concise summary that captures the salient points of the source text. We'll demonstrate a synthesis process based on the use of trainable machine learning algorithms with a collection of features taken straight from the source text. There are two groups of these features: statistics-also referred to as extractive summarization-based on how frequently different textual components occur and linguistics also known as abstractive summarization - based on morphological and syntactic analysis of documents. In this paper, we show how BERT can be usefully applied in text summarization and provide a common framework for abstraction models. We also demonstrate that a fine-tuning method can further improve the quality of the generated summaries.