Towards enhanced literature review efficiency powered by artificial intelligence

Abstract

Literature reviews for research and teaching tasks is increasingly facing new challenges, mainly due to the considerable growth of bibliographic material in all disciplines. Primarily, the search and selection of information become exhausting and overwhelming due to the lack of planning in the systematic collection of documents, as well as the improper handling of metadata and the high variability of the results offered by automatic search mechanisms. In this sense, artificial intelligence is transforming the landscape of literature review by employing systems capable of quickly analyzing and interpreting the content of text documents while providing very accurate results, higher than traditional search engines. This paper aims to describe recent advances in the intelligent integration process of literature review methods, referring to the use of machine and deep learning techniques associated with optimization stages of natural language processing. Finally, we address the implications of employing systems capable of making judgments and making decisions regarding the usefulness and relevance of the analyzed data.