Multimodal Classification of Teaching Activities from University Lecture Recordings

The way of understanding online higher education has greatly changed due to the worldwide pandemic situation. Teaching is undertaken remotely, and faculty incorporate lecture audio recordings as part teaching material. This new teaching–learning setting largely impacted university classes. While technology that enriches virtual classrooms been abundant over past two years, same not occurred in supporting students during learning. To overcome this limitation, our aim work toward enabling easily access piece lesson recording which teacher explains a theoretical concept, solves an exercise, or comments on organizational issues course. end, we present multimodal classification algorithm identifies type activity being carried out at any time by using transformer-based language model exploits features from file automated transcription. experimental results will show some academic activities are more identifiable with signal while resorting text transcription needed identify others. All all, contribution aims recognize lesson.