Combining artificial intelligence and human problem solving: Proposing a new architecture for ITS's

Abstract

This paper shows a comparative study between the Artificial Intelligence Problem Solving and the Human Problem Solving. The study is based on the solution by many ways of problems proposed via multiple-choice questions. General techniques used by humans to solve this kind of problems are grouped in blocks and each block is divided in steps. A new architecture for ITS - Intelligent Tutoring System is proposed to support experts' knowledge representation and novices' activities. Problems are represented by a text and feasible answers with particular meaning and form, to be rigorously analyzed by the solver to find the right one. Paths through a conceptual space of states represent each right solution.