Using Neural Networks to Identify Instructional Activities in Elementary Classrooms to Support Evaluation of Instruction

• In Event: AERA Poster Session 7
  In Poster Session: Technology, Instruction, Cognition and Learning SIG Poster Session 2 (San Diego)

Abstract

During the past 15 years, there has been a significant increase in the use of video for preparing teachers and studying teaching quality. In spite of this, key challenges remain to employing videos at scale including the time and financial cost involved in scoring them. Recent advances in computer vision, machine learning, and deep learning may provide solutions to these challenges. This study reports on efforts to train neural networks to identify complex instructional activities in elementary videos. We found that two specific neural networks used for deep learning can be utilized to assess student engagement levels and to indicate whether whole group, small group, or individual student activities were occurring based on occurrence of multiple activities and objects temporally.

Authors

• Peter A. Youngs, University of Virginia

• Ginger S. Watson, University of Virginia

• Scott Acton, University of Virginia

• Samarth Singh, University of Virginia

• Matthew Korban, University of Virginia

• Jonathan K. Foster, University of Virginia