Dr hab. Katarzyna Śledziewska’s project, funded in the competition for research projects at CESS

27 11 2024

The project of Dr hab. Katarzyna Śledziewska, prof. UW from the Faculty of Economic Sciences “Development of a methodology enabling partial automation of systematic literature review using tools from the fields of text mining and generative artificial intelligence” was funded in the competition for research projects at the Centre of Excellence in Social Sciences.

The project received funding in the amount of 65 161,00 PLN.

The project aims to explore how much of the scientific literature review process can be automated using a combination of AI tools, such as Large Language Models, alongside traditional text processing techniques like text mining. Systematic literature reviews are notoriously time-consuming, so our goal is to develop a methodology that leverages these tools to process large volumes of text while capturing semantic nuances, thereby accelerating the review process.

AI tools are already used in literature reviews, but they typically focus on the initial stages, such as selecting articles based on keywords or analyzing abstracts. Our project goes beyond this by proposing a method to analyze the full text of articles, facilitating final selection based on criteria set by researchers and extracting key information from those articles.

Additionally, we will develop methods to assess the quality of these automated reviews by comparing them with three human-conducted reviews. We will also create a Python-based tool that will be made available to other researchers to apply our methods. Initial findings suggest that AI models can not only match the performance of human researchers but also identify errors that may arise in manual reviews. In summary, the project seeks to make literature reviews faster and more accurate by employing various text processing tools.