This self-guided course presents basic concepts underlying AI/ML as applied to health care data. NOTE: Enrollment is ongoing. currently interactive office hours via Zoom are being offered with the course developers to answer question and review materials as needed.
The course content is organized as a series of self-contained python notebooks, each with an accompanying recorded tutorial and example datasets. The notebooks provide a written narrative of the python libraries that are used to clean/build training sets, define AI model architecture, and evaluate model performance. Help sessions with some live presentations will also be provided. NOTE: The course uses Google Colab to work with the notebooks and you will need a google account. It is possible to use them in another Jupyter notebook environments but we cannot provide support for if their are problems.
The content is intended for those who are able to write basic code in a python notebook, but do not have a background in statistical learning or experience implementing basic machine learning libraries. Although there are no required prerequisites to enroll in the course, participants will benefit from having some basic python skills and a willingness to learn more. Designed to meet the level of Masters Degree graduate students, it is also suitable for motivated undergraduates and independent learning.
Course Objectives
Participants will learn to apply basic machine learning algorithms to example datasets, and extend those methods to their own data. This includes...
- Formatting data into appropriate formats for model training and validation.
- Training and evaluating predictive classification models
- how to tune and improve model performance
Course Structure
The course has five modules:
- Introduction to Classification
- Unsupervised Learning: K-means Clustering
- Handling Non-linearity, Model Complexity, and Regularization
- Model Selection, Hyperparameter Tuning, and Evaluation
- Introduction to Neural Networks
Instrumental Persons
We would like to recognize and express our gratitude to the following people who shared in the creation and implementation of this course:
- Matthew McCoy, PhD, Georgetown University
- Samir Gupta, PhD, Georgetown University
- Aidan Hoyal, MSIS, MA, Communications Hub, AIM-AHEAD
The content is intended for those who are able to write basic code in a python notebook, but do not have a background in statistical learning or experience implementing basic machine learning libraries. Although there are no required prerequisites to enroll in the course, participants will benefit from having some basic python skills and a willingness to learn more. Designed to meet the level of Masters Degree graduate students, it is also suitable for motivated undergraduates and independent learning.
Screenshot from a sample notebook:
