Scientists from the Moscow State Psychological and Pedagogical University (MGPPU) created a unique computer game that allows one to assess the speed of decision-making by a player and his ability to predict the situation. The development will also serve as a specialized trainer for these skills, they said. This was reported by the press service of the university.
Modern methods of studying cognitive abilities imply not only the measurement of numeracy and logical thinking skills, according to the MGPPU scientists: the system they have developed is also capable of assessing the quality of forecasts and decisions made in rapidly changing conditions.
In the virtual space created by university specialists, the player is confronted by many active opponents. Their behavior is controlled by artificial intelligence based on a complex mathematical algorithm. The player's goal is to "eliminate" all opponents in a limited playing field.
"Based on the theory of Markov random processes and methods of nonlinear optimization, we have developed procedures that allow us not only to record the correctness of the subject's actions when solving problems, but also to analyze his behavior," explained the head of the laboratory of quantitative psychology at the Center for Information Technologies for Psychological Research of the Faculty of Information Technologies, Moscow State University of Psychology and Education Pavel Dumin.
According to him, artificial intelligence adapts the gameplay for a specific subject, observing his strategy and generating situations of such complexity that will give the most detailed idea of the real level of development of his skills.
The authors note that the game is in fact also a test tool and simulator of the abilities required by operators of complex automated systems, such as sapper robots or aerial monitoring drones.
"The difficulty in the development of such psychodiagnostic tools is the accumulation of large empirical data for the correct adjustment of adaptive procedures. So far, we compensate for the missing amount of data on testing people belonging to representative groups with the help of mathematical optimization," said Pavel Dumin.
Further tasks of the research team are to prepare the system for implementation in educational practice, as well as to develop software elements based on the created artificial intelligence that increase the viability of automated systems in emergency situations.