Converting Conceptualizations into Executables: Commentary on Web-based Adaptive Education and Collaborative Problem Solving
Joseph M. Scandura
High-level descriptions for web-based adaptive education and collaborative problem solving proposed by Kennedy et al and by Eccles and Groth (this issue) provide useful starting points. The challenge is to build on and to convert such analyses into executable systems. This commentary shows how the proposed frameworks may be automated by representing knowledge in terms of Abstract Syntax Trees (ASTs) and associated Structural Learning Theory (SLT). A wide range of questions can be asked both at conceptually high levels of abstraction as well as regarding implementation.
ASTs provide an intuitive basis for accomplishing this goal by making it possible to represent knowledge at multiple levels of abstraction. The relatively complex inferential nature of associated tutorial logic also is shown to have important implications for how web-based adaptive education is actually implemented. Supporting collaborating learners will require representing the knowledge associated with multiple learners, and specifying inferences between multiple (or shared) learner models as learning progresses. In cooperative problem solving, problem solvers receive the same or complementary problems. In adversarial situations, problem solvers, human or automated, effectively have different problems (e.g., different goals), as well as different rule sets (i.e., knowledge).