Inteligencia artificial basada en Deep Learning y Auto Reportes para elanálisis emocional del alumnado en el proceso de enseñanza-aprendizaje

Resumen

The development of automatic tools for detecting the emotional states of students can be key inmanaging their motivation and involvement in the teaching-learning process. Ten students of thesubject of Curricular Design and Digital Classrooms in Early Childhood Education, belonging to theMaster’s Degree in Early Childhood Education taught at the Faculty of Education of the Universityof Alicante took part in seven dynamics during the 2020-21 academic year, registering their faces tobe analyzed by an artificial intelligence system based on Deep Learning technology and trained inthe recognition of facial emotions. In parallel, the participants completed self-report forms on theiremotional state before, during and after the educational dynamics. The results obtained show a totalcorrelation between both systems, automatic and self-report, of 23%, partial of 60% and null in therest of the cases. However, the self-report shows a similar proportion of neutral, negative and positiveemotions with a slight preponderance of the latter, while the automatic system identifies neutral facialexpressions in a high percentage. Additionally, there does not seem to be a predominantly positiveemotional response to active dynamics compared to those where student participation is passive. Thesize of this sample, corresponding to a preliminary study, requires additional research to clarify orconfirm these findings.