Course Subject: | Applied Math in Data Science and Machine Learning |
Student Level: | Undergraduate and Graduate |
Number of Students: | 10-50 (depending on the course) |
Developed by: | He Wang, Associate Teaching Professor and Director of MS in Applied Mathematics and MS in Operations Research, College of Science |
What Students Did
In homework, computer labs, and the final project, students solved real-world math problems using their understanding of mathematical theory and concepts. They then prompted ChatGPT (or GitHub Copilot or Llama) to generate the code they needed to represent their models in either R, MATLAB, or Python, understand the code, and determine whether the response was correct.
Learning Goals and Purpose
Students use ChatGPT to help with coding and debugging so they can focus on the most important learning–mathematical concepts and the logic of the programming–rather than syntax. Correctly using ChatGPT is also part of the learning, and it will boost the learning of applied mathematics.
Assessment
Students are graded mainly on the math part of the assignments. They are not required to cite ChatGPT, since coding is the tool to do the numerical computations, but not a core learning outcome. Before ChatGPT was introduced into the courses, students used online resources to help with coding and debugging, which is not effective for beginners. The grading was the same now as it was then.
Faculty Reflections
It is very important for students to learn to work independently and to be lifelong learners, and this is what students practice in my courses.
Step-by-Step Directions for Students
Step 1 | Solve the real-world math problem through knowledge of mathematical theory and concepts. |
Step 2 | Develop a model of the solution using R, Python, or MATLAB, with the help of ChatGPT (or GitHub Copilot, or Llama). |
Step 3 | Determine whether the ChatGPT response is accurate and refine prompts as needed to generate the correct answer. |
Download PDF version of this assignment
Return to AI Gallery Home