Characterizing the effect of seating arrangement in a classroom learning using neural networks

Abstract

A framework patterned from Eric Mazur's Peer Instruction Method is developed to probe the propagation of information in a classroom setting. Using actual experimentation, we demonstrate that a two layer neural network is capable of accurately characterizing the final spatial information profile of students that are allowed to interact. The resulting dynamics suggest that the spreading of correct information can be optimized by strategically assigning a seating arrangement based on the perceived intelligence level of each student. The results are expected to provide an impetus on new instructional structures/formats that inherently incorporate student interactions during lectures.