Application of Generative Adversarial Network Algorithm on Journaling System with Comics and Cognitive-Behavioral Theory in Social-Emotional Learning for Intermediate Elementary Students: A Pilot Study

Like any other skill, social-emotional learning (SEL) must be practiced to be automatic. To transform the knowledge of SEL into a practical and deliverable lesson, we propose an emotion-oriented journaling system with comics for students. The system transforms the cognitive-behavioral therapy into manageable steps using comics-based storytelling and uses a method combining the term frequency-inverse document frequency (TF-IDF) method and generative adversarial network (GAN) algorithm to implement a text-to-graphics conversion. The system was tested by 36 fifth-grade students for four weeks. The experimental result shows that 41.7% of students improved their abilities of emotional recognition and self-awareness after one month of journaling on the system. Consequently, the system provides a creative and effective way for social-emotional education while speeding up the generation of visual journaling content in less time.