Year: 2007
Type: conference-paper
Abstract: This paper presents the architecture, structure of knowledge base, and the organization of the prototype of Leibniz pedagogical agent. This agent works as an educational assistant for Differential and Integral Calculus classes. Main innovative aspect present in this agent is the approach taken to model the interaction processes between students and teacher, which is based on Pedagogical Negotiation concepts. Another innovation is the inference mechanism used to detect divergences in the negotiation process. This mechanism allows probabilistic estimation of divergences in expected answers, combining a symbolic matcher of mathematical formulas with a bayesian networks inference engine. After an initial discussion about related work, the paper presents the theoretical foundations of Leibniz, shows the architecture of the agent, and the structure of its knowledge base. The organization of the prototype is also shown, including its inference process, interface facilities, and an example of use. The paper is finished with the presentation of some initial results.
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
SDG: Quality education
Open Access status: gold
OpenAlex ID: W4406908300
