Skip to content

Intelligent Textbooks Proposal

Title: University of Minnesota EECD Senior Seminar Project Proposal 2024

Project Name: Intelligent Textbooks for Electrical Engineering

Facilitator: Dan McCreary

Summary

This project will allow participating students to create a set of intelligent textbooks that embrace the modern use of generative AI. The output of this project will be the foundation of an intelligent textbook for each student who participates.  Students will gain a deep understanding of how to create high-quality prompts that extract precise knowledge from large-language models and store this information in an open format on GitHub with a textbook rendered on GitHub pages.  No prior programming experience is required, but students must work with faculty that are subject-matter experts in a specific course.

Details

We propose that one of the groups in the EECD Senior seminar build a series of "Intelligent Textbooks" using generative AI.  Last year, we had two groups participate.  One group built MicroSims and did hands-on work extending the AI Racing League to use the updated Raspberry Pi 4 computers.

Since last year, generative AI has become much more capable of building out not just single labs or writing small Python programs; it has also been used to generate the foundations of Intelligent Textbooks, including: 1. How to generate modern course descriptions built around moving from knowledge to content generation (Bloom's Taxonomy).

  1. How to generate a detailed enumerated list of Learning Concepts for a specific course or subject matter

  2. A dependency graph of these Learning Concepts indicating the order in which they should be taught to achieve specific objectives

  3. A taxonomy of Learning Concepts allowing the classification and grouping of similar concepts

  4. Tools to visualize, edit, and certify  these Learning Concepts with Subject-Matter Experts

  5. Hooks to allow lesson plans to be customized to the needs of a project, group, or individual student

  6. Frameworks to allow custom simulations to facilitate learning of complex concepts (MicroSims)

I propose to have each student in this project select one course at the University of Minnesota EECD department and work with a subject-matter expert to generate an intelligent textbook for this course. Note: Students will be required to use GitHub and GitHub pages to publish the content.  Although familiarity with GitHub is not required, it is necessary to complete these projects.

Why This is Important

Today, many courses still use costly printed textbooks.  These textbooks are rigid and force a learner's learning order.  Static textbooks don't promote the use of adaptive learning.  Building many advanced search and indexing tools using generative AI is now possible with just a few sample prompts.  We can now rate online courses on a scale of 1 to 5 and see if students prefer level 5 textbooks.

References