Calibrating the Performance Assessment Mechanism in Virtual Laboratories with a Reinforcement Learning-Inspired Technique

Science universities strive to offer efficient lab training to their students and at the same time secure their safety and minimize the damages to the lab equipment. Thus, the development of distance learning tools for students to be trained virtually and safely in using the various lab instruments and performing experiments is necessary. Since the students are evaluated for their performance at the on-site labs, the assessment at the virtual labs is also needed and consequently, an embedded assessment mechanism for the evaluation of the user’s performance in the virtual lab is a necessary feature. For the assessment mechanism to be reliable and devoid of the designer’s bias, though, it may need calibration with Machine Learning. Hellenic Open University has developed its own virtual biology laboratory, Onlabs, which simulates its on-site one for its students to be trained and evaluated at. Considering the evaluation of the user’s performance in Onlabs, it is made with respect to particular experiments and is based on an embedded scoring algorithm. The latter is two-fold, measuring the extent to which the necessary steps have been made and the extent to which those steps were made in the correct order. Within the context of the experimental procedure of microscoping, the scoring algorithm has been recalibrated with the use of various Machine Learning techniques. In this paper, we propose the design of a Reinforcement Learning variant and recalibration of the scoring measure concerning the steps order. The results suggest that under specific parameters and Reinforcement Learning methods, a more efficient scoring mechanism may be achieved.