Supporting Problem-Solving in Mathematics with a Conversational Agent Capable of Representing Gifted Students´ Knowledge

This paper describes a conversational agent designed to support problem solving in Mathematics. The agent´s knowledge base has been structured to represent gifted students´ problem solving strategies. These were students who won the Brazilian Mathematics Olympics for Public Schools, and the idea here has been to elicit and represent their formal and heuristic knowledge for problem solving. The paper describes the method for capturing the cognitive processes of gifted students in solving Math problems and the structuring of this knowledge for the conversational agent. The paper also presents the results achieved, showing that the students who used the agent to solve Math problems were fully engaged in the tasks proposed and had a better performance than when not using the agent.