Learning Agents for Tutoring Systems

Computerized training systems must adapt to individual learning styles, as required within Socratic dialogues; the Socratic dialogue has been well recognized as an ideal on-to-one teaching methodology. This dialogue is such that specific questions are directed to the individual student, eventually guiding him or her to enlightenment on a particular subject. In this manner, the student not only would experience the thinking necessary to reconstruct the subject domain but also make the training software adapt to his/her learning style for maximal learning performance. Here, we propose to investigate the applicability of software learning agents to tutoring systems. A software agent is a program that runs in the background and performs user-assigned tasks autonomously. A learning agent observes user's actions or other events in the background, finds repetitive pattersns, automates them upon approval, and gets smarter with user or environment feedback over time. We believe that learning agents can dramatically improve computerized training systems. In particular, we propose to review learning styles and artificial intelligence methods for tutoring systems, to investigate the applicability of software agent technology to tutoring systems, and analyze feasibility on a selected vocational training example. Anticipated Benefits: The successful completion of our Phase I plans have provided the foundation for a Phase II effort involving the development of a software agent prototype for a vocational training system. The successful completion of our Phase II plans will enable a commercialization effort for Phase III to market the developed software agent to the Fortune 1000 market. Under a Phase III development effort, we will focus on the commercial product development, launching, and marketing.