U-M BIOMEDE 487: A.I. in BME

AI and machine learning algorithms have had a significant impact on biomedical science in the past decade. AI algorithms can learn patterns from biomedical data sets to provide actionable insights on disease diagnosis or treatment. 

The intention behind launching BIOMEDE 487 was to offer a comprehensive introduction to the world of artificial intelligence within the biomedical engineering sphere. “The original idea was that I wanted to teach it for seniors, first-year graduate students, and Ph.D. students,” said Sriram Chandrasekaran, associate professor, biomedical engineering. 

The course, now nearing its fourth year, has successfully lived up to its intent. ”I hope this course ignites students’ interest in AI and they can incorporate it in their research by understanding AI applications in biomedical engineering,” he explained. 

BIOMEDE 487 offers a wide-ranging overview of AI applications in biomedicine. Dr. Chandrasekaran designs it as an “umbrella” course, which provides an overview of a wide range of machine-learning tools (clustering, regression, decision trees, random forests and neural networks), biomedical data sets (imaging, omics and and health records) and diseases (cancer, infectious-, cardiovascular- and neurological-). “We look at a broad range of biomedical applications and talk about how to develop machine learning models that can tackle challenges in different subfields of biomedical engineering,” he said.

The course focuses on practical applications of AI in BME with hands-on tutorials, which form the core of the learning process. “Except for the very first lecture, which provides a course overview, every other week is hands-on,” Dr. Chandrasekaran said. Students begin with machinery learning concepts and immediately apply them to specific biomedical data. The content includes various programming languages—MATLAB, Python, and R —to cater to the diverse coding backgrounds of students. This flexibility, Dr. Chandrasekaran believes, significantly reduces the barrier to entry for those coming from diverse educational experiences and interests.

The course is structured around practical, real-world data challenges. “There are no exams; students really like that,” Dr. Chandrasekaran said. ”There’s a project that they work on throughout the course, making up 50% of the course grade. These projects have ranged from predicting early mortality during childbirth to identifying diabetic and stroke risks from patient health records. During the pandemic years, the focus shifted significantly, with students exploring COVID-19 transmission models and microbiome impacts on infections—highlighting the course’s adaptability to address pressing global concerns.”

Adding to the hands-on approach, Dr. Chandrasekaran invites guest lecturers to provide students with diverse perspectives. Speakers have included experts from the Mayo Clinic, Michigan State University, and various U-M departments. “These sessions offer a real-world viewpoint, making the course more engaging and informative,” Dr. Chandrasekaran said.

BIOMEDE 487 has been immensely popular across several departments. “I’ve seen students from nuclear engineering, industrial engineering, the medical school, pharmaceutical sciences, and chemical engineering taking this course,” Dr. Chandrasekaran added. “Even those well-versed in computer science enroll to understand the biomedical aspect of AI.” The team projects require cross-disciplinary collaboration, enhancing the learning experience.

The course contributes to students’ professional growth. Dr. Chandrasekaran noted: “Students use AI knowledge from this course in their Ph.D. research or data science roles in academia and industry. I’ve seen students incorporate AI algorithms into their thesis projects.”

Reflecting on the course’s success, Dr. Chandrasekaran acknowledges the challenge of accessibility due to its high demand. “The course fills up super fast. If someone wants to take it, I’d recommend signing up quickly as it fills up as soon as the window opens. The next opportunity will be in Fall 2025,” he said.

With each iteration of BIOMEDE 487, Dr. Chandrasekaran continues to refine the curriculum. His commitment to making AI accessible and applicable across the biomedical engineering spectrum echoes U-M’s dedication to producing graduates rooted in interdisciplinary technologies.