Adaptive Multi-Agent Assisting Framework for a Personal Teaching Environment

This Research to Practice Work in Progress presents the first steps and ideas of a framework to address the problem of suggesting the most suitable recommendations for instructors (for teaching and assessing), by designing an intelligent multi-agent recommender system that uses data analysis methods. The paper addresses the whole framework, specifically focusing on the assessment part. The framework proposed takes into consideration the heterogeneous personalities and teaching/assessing styles of different instructors to personalize and customize their experience. It provides immediate and customize instructions and feedback to help instructors improve their educational tasks. The dataset contains data collected from the engagement of the instructor with the agents, to predict their teaching/assessing style from their behavior. The agent system has the ability to recommend methods and tools against a topic. It tries to build different instructors’ profiles, to generalize the most common practices toward an activity for future use. The agent system is like a personal assistant that helps teachers with finding information, and it gives them recommendations.