Identifying and Visualizing Trends in Science, Technology, and Innovation Using SciBERT

Identification of science, technology, and innovation trends is a critical topic both for the scientific community and for companies that develop technologies, work on science and technology policy or invest in high tech. In this research authors demonstrate a novel approach implemented in iFORA system (developed by National Research University Higher School of Economics) using the SciBERT. The aim of this article is to determine the capabilities of modern text mining with a focus on the possibility of using them to construct understandable and easy to grasp visualizations. The article determines that the current level of modern language models development can be used for trendspotting. In the result, authors show the ways to present the conclusions of trendspotting in the form of a simple visualization: The trend-matrix with the distribution of the key 100 trends of the analyzed areas in four corners formed by the intersection of the axes “total amount of research papers” and “mean publication year of research papers”. The approach is critical for policy makers, high tech companies, researchers, and consultants and helps to improve both the objectivity of the trendspotting and its practical applicability for management of science and technology.