Course materials including lectures and problem sets for CMU 10-742: Machine Learning in Healthcare, taught in Fall 2024. Course Website
Machine learning (ML) is experiencing explosive growth in healthcare, and is now top of mind for leaders at hospitals, insurance companies, and pharmaceutical firms. This course offers a survey of ML in healthcare today. Students will gain firsthand experience working with electronic health records, time-series medical data, health insurance ("administrative") data, and many other healthcare data sources. The course will cover how ML (and AI more generally) is impacting healthcare financing, operations, and care delivery, with select 'deep dives' into specific verticals such as radiology, pathology, and ophthalmology. Students will learn how to apply ML methods to varied problems in healthcare, such as predicting disease onset and forecasting how long a patient will remain in the hospital. The course will address the challenges of working responsibly with healthcare data, including potential biases and inconsistencies and confounders, and provide strategies for identifying and mitigating these issues.
The course assumes a strong competency in Python/pandas/jupyter, and hands-on experience building models such as xgboost, logistic regression, and neural networks. It also requires a mathematical maturity that includes college-level probability, statistics, and discrete math. No background in healthcare is expected. The class is open to graduate students in SCS. It is also open to qualified, motivated undergraduates from all majors in SCS, and to other students who fulfill the above requirements.
- Basic knowledge of machine learning concepts
- Programming experience in Python
- Familiarity with statistics and probability
- Background in linear algebra (recommended)
-
Lectures: lecture slides in ppt or pdf format
-
Problem Sets: Jupyter notebooks containing the five problem sets assigned during the course.
By the end of this course, students will be able to:
- Apply appropriate ML techniques to various healthcare problems
- Understand the unique challenges of working with medical data
- Implement and evaluate clinical prediction models
- Critically assess ML applications in medical literature
You can contact the course instructor at adam.berger@gmail.com
For technical issues with course materials, please open an issue in this repository. (Although I don't monitor it diligently.)
While this repository primarily serves as an archive of course materials, suggestions for improvements are welcome. Please feel free to:
- Report errors or typos via issues
- Submit corrections through pull requests
- Share feedback on course content
Special thanks to the teaching assistants and students of Fall 2024 who contributed to the development and refinement of these materials.
This course material is licensed under CC BY 4.0. When using these materials, please cite:
Adam Berger, 10-742 Machine Learning in Healthcare, Carnegie Mellon University School of Computer Science, Fall 2024.
Course website: https://adamleeberger.github.io/ML-in-healthcare-course/