Text Summarization for Kannada Text Documents: A Review

In recent times, we have observed the emergence of Automatic Text Summarization (ATS) systems, which are now being developed not only for the English language but also for low-resource languages. The growing quantity of textual information on the Internet, including vast archives of news articles, scientific papers, legal documents, and more, has made ATS increasingly essential. Manual text summarization is known to be time-consuming, costly, and often impractical when dealing with the massive volume of text available. Since the 1980s, research scientists have been dedicated to enhancing ATS techniques. These approaches are typically categorized as extractive, abstractive, or hybrid. The extractive methodology involves selecting the most crucial sentences from the input document(s) and concatenating them to create a summary. In contrast, the abstractive methodology rephrases the content of the input document(s) into an intermediate representation, generating a summary with sentences that differ from the original text. The hybrid methodology integrates extractive and abstractive methodologies. Despite the ongoing development of various methods, it is worth noting that the summaries generated by these systems still fall short of human-generated summaries. Most research efforts have concentrated on the extractive approach, but there is a pressing need to shift more attention towards abstractive and hybrid methodologies. This research endeavour aims to offer a comprehensive survey for researcher scientists by covering various aspects of ATS, including methodologies, techniques, building blocks, methodologies, datasets, evaluation methods, and the specific challenges that need to be addressed in the context of Kannada language text.