Toward Optimal Pedagogical Action Patterns by Means of Partially Observable Markov Decision Process

摘要: Good pedagogical actions are key components in all learning-teaching schemes. Automate that is an important Intelligent Tutoring Systems objective. We propose apply Partially Observable Markov Decision Process (POMDP) order to obtain automatic and optimal recommended action patterns benefit of human students, the context System. To achieve goal, we need previously create efficient POMDP solver framework with ability work real world tutoring cases. At present time, there several Web available open tool solvers, but their capacity limited, as experiments showed this paper exhibit. In work, describe discuss design ideas toward solver, useful our problem domain.