Can AI Tell Emerging Technologies: Evaluating the Importance of Quantitative Features of Technology

Abstract

Many researchers and organizations have been archiving and analyzing vast documents and data for technology evaluations and emerging technology mining. Korea Institute of Science and Technology Information (KISTI) as one of them, has been collecting various technological data from technical literatures, such as patents and papers, and developing techniques to analyze and retrieve various quantitative features from it. Lately, the demand of utilizing our resources, data and technologies, for developing an intelligent technology information system which output is objective, consistent, and explainable, has been increase. By applying the latest advanced artificial intelligent techniques, deep learning, to our data and system, it is possible to improve our capability of evaluating technology and mining future emerging technology. For this reason, it is necessary to investigate and evaluate the effectiveness of each quantitative features of technology which are retrieved from technical literature analysis. In this paper, we present the results of our study of testing the effectiveness of various quantitative features of technology, which are being referred by human experts in technology evaluation and future emerging technology mining process, in both empirical and statistical ways. In the empirical approach, an artificial intelligent model is built to simulate the human expert group for emerging technology mining and the change of the performance is observed while the training features are changed. In the statistical approach, the relations between the basic distribution variables of data and the decision making is analyzed.