Incorporation of ChatGPT and Other Large Language Models into a Graduate Level Computational Bioengineering Course

King, Michael R.

DOI: https://doi.org/10.1007/s12195-024-00793-3

2024-02-08

Cellular and Molecular Bioengineering

Abstract:The remarkable capabilities of generative artificial intelligence and large language models (LLMs) such as ChatGPT have delighted users around the world. Educators have regarded these tools as either a cause for great concern, an opportunity to educate students on cutting-edge technology, or often some combination of the two. Throughout the Fall 2023 semester, we explored the use of ChatGPT (and Bard, among other LLMs) in a graduate level numerical and statistical methods course for PhD-level bioengineers. In this article we share examples of this ChatGPT content, our observations on what worked best in our course, and speculate on how bioengineering students may be best served by this technology in the future.

engineering, biomedical,biophysics,cell & tissue engineering,cell biology

What problem does this paper attempt to address?

The paper attempts to address the issue of how to effectively integrate generative artificial intelligence (such as ChatGPT) and other large language models (LLMs) into graduate-level biomedical engineering computation courses. Specifically, the authors explored the application of ChatGPT and other LLMs in the course "BME 7410: Quantitative Methods in Biomedical Engineering" during the Fall 2023 semester and evaluated the effectiveness of these tools through 7 assignments.

### Main Issues:
1. **Effectiveness of Educational Tools**: Investigate the practical application effects of ChatGPT and other LLMs in education, particularly whether they can enhance students' understanding and mastery of numerical and statistical methods.
2. **Student Engagement**: Assess students' acceptance and engagement with these new technologies, as well as their performance when using these tools.
3. **Limitations of the Technology**: Identify and discuss the limitations and potential issues of these tools in actual teaching, such as generating erroneous code, reliance on Python instead of MATLAB, etc.
4. **Future Application Prospects**: Predict the potential uses and development directions of these tools in future biomedical engineering education and practice.

### Solutions:
- **Assignment Design**: Set specific LLM tasks in each assignment, requiring students to interact with ChatGPT or other LLMs to complete specific computational and programming tasks.
- **Evaluation Methods**: Evaluate the performance of these tools in actual teaching through students' assignment submissions and feedback.
- **Interaction Methods**: Encourage students to engage in multiple rounds of interaction with LLMs, including asking questions, critically thinking about the outputs of LLMs, and debugging errors.
- **Extended Activities**: Increase students' understanding and interest in LLMs by sharing relevant preprints, journal articles, tweets, and live demonstrations.

### Conclusion:
- **Positive Feedback**: Most students expressed interest and satisfaction with the LLM assignments, believing that these tools increased the course's fun and practicality.
- **Technological Progress**: During the course, the capabilities and behaviors of LLMs changed significantly, demonstrating the rapid development in this field.
- **Future Outlook**: It is recommended to introduce LLM-based final projects in future courses to provide a deeper learning experience.

Through these methods, the authors hope not only to enhance students' computational and programming skills but also to cultivate their critical thinking and application abilities regarding emerging technologies.