Hindi Text Summarization: Using BERT

This paper points out a vacuum in the literature on text summarization: not enough attention has been paid to methods for Indian languages. This research investigates the current state-of-the-art techniques for summarizing Indian languages. The review reveals a bias in current research towards extractive summarization, which involves choosing important lines, whereas abstractive summary, which involves rewriting text, is more common. The paper advocates for a more balanced examination of both abstractive and extractive summarizing strategies specifically adapted for Indian languages by examining current methodologies and identifying their advantages and disadvantages. In certain situations, the suggested system outperforms models such as SummaRuNNer, Lead-3, NeuSumHD. These improvements improve the resilience of the model as well as the quality of the generated summaries with a ROUGE-1 score of 41.43 and ROUGE-2 score of 24.54.