Abstract

The number of papers is such that it’s impossible to read them all. Everybody who carries out a literature review is aware of this. For each study you have to determine its aim, the methods it employed, the data and metrics it used, the limitations or omissions it had, and what the authors believe should happen next. This process is very time-consuming. Currently, most AI tools are only capable of helping with locating papers, summarising them, or answering questions. They barely deal with the task of comparing studies in an organised manner or identifying areas where further research is needed. In this seminar we present an AI-powered Research Decision Support System. The system combines Natural Language Processing, Large Language Models, and intelligent document analysis and examines a number of papers together. It extracts the key points from each paper and presents them in a structured format. It then compares the findings of the studies and compiles the evidence so that limitations which recur frequently and trends that are just beginning to appear become obvious. It also attempts to identify possible research gaps and rank them. The criteria used are intended to be transparent and based on evidence. Our aim is to reduce the amount of manual work involved in a literature review and to provide researchers with a well-organised resource on which to base their planning and decision-making.

Keywords

How to Cite This Article

Rights