Neural Networks for Automated Classification of Classroom Video and Ratings of Teaching Quality
Abstract: Classroom video is a key element of teacher preparation and the evaluation of teaching quality, but employing it at scale is time-consuming and expensive. A large portion of the expense is due to the use of human raters to view and score videos. Automated classification of teaching video and ratings of teaching quality could dramatically reduce costs while also creating new opportunities for teacher training and evaluation. Recent work shows that artificial neural networks are capable of highly accurate video classification in general settings, but no work addresses the use of neural networks for the classification of classroom video and ratings of teaching quality. Results will shed light on the accuracy of neural networks and the feasibility of using these methods on a large scale. In addition to making significant gains in efficiency while significantly reducing costs, this work has the potential to yield new developments in teaching metrics and autonomous classroom simulations.