HILBERT & PATRIC: Hybrid Intelligent Agent Technology for Teaching Context-Independent Reasoning

There is a disturbing paradox at the heart of contemporary American education: As this education turns more ‘electronic,’ we are moving away from the one kind of learning that we know to be most effective, namely, one-on-one instruction. As the need for good teachers at the university level continues to grow, we see this paradox intensifying. A specific exacerbating force is that, as the data tells us, as educators, we are not producing students able to successfully employ context-independent reasoning (CIDR) in technical domains. Our future scientists and engineers have been shown to lack this fundamental tool of their trade. The fact is, teaching students to be good CIDR reasoners requires the professor to develop a one-on-one relationship with each student, with a keen eye on how each applies their own strategies for proof construction. With this in mind, we envision the automated logic instructor as adaptable, and fully available to each student, at all times. This is obviously not possible with human instruction, but our preliminary work suggests that our vision is capable of being realized in the digital domain: We are developing a suite of intelligent agents, including PATRIC and HILBERT that marries the cutting edge in AI-driven tutoring with the state-of-the-art in proof construction courseware.