2026 ASEE Annual Conference & Exposition

As machine learning (ML) and artificial intelligence (AI) become increasingly intertwined with engineering work, it is becoming increasingly important to develop engineers who are not only strong AI users, but also strong AI contributors. Therefore, it is important to help engineering students build expert-like thinking about ML/AI concepts so they are able to collaborate with others to develop and improve AI applications. One key step towards developing expert-like thinking is using a conceptual framework to organize and retrieve technical information.

One conceptual framework for teaching ML/AI consists of four categories that can be used to categorize concepts and be applied to ML/AI applications. The categories include data, task, algorithm, and assessment. Preliminary work showed how the framework could be used to teach ML and AI concepts in engineering courses, and this paper expands on that existing work to demonstrate and evaluate how the framework is applied by students to actual ML/AI applications.

After students were introduced to the ML/AI Conceptual Framework during a 1-credit ML for Engineering course, they were asked to apply the framework to an academic paper about an engineering application of ML/AI. Students were given a general application area (e.g. self-driving cars, power systems, manufacturing), and they were allowed to choose the academic paper they reviewed. They then wrote a group report where they were asked to describe each of the academic papers using the ML/AI Conceptual Framework. The instructor of the course also separately read and applied the framework to each paper. The instructor’s and students’ answers were compared, providing valuable insights about 1) how well the ML/AI Framework applies to a variety of engineering applications, and 2) a comparison of how the instructor and students understood and applied the framework. Takeaways for those using the framework and/or teaching ML/AI are presented.