The Model & Dataset Disclosure for Healthcare and Public Health (MDSD4Health) curriculum website is a space to learn about how machine learning disclosure methods and mediums benefit the pursuit of algorithmic transparency in healthcare and public health contexts.
1.1: What is Machine Learning?
1.2: Bias & Fairness in Machine Learning
2.1: Model Replicability & Reproducibility
2.2: Model Generalization
2.3: A Primer on Model & Dataset Disclosures (MDSDs)
3.1 Datasheets for Datasets
3.2: Other Dataset Disclosures
4.1 Model Cards
4.2: Other Model Disclosures
Revise our exercises on GitHub
Our Guiding Ideas
Machine learning education for all
Everyone can be impacted by the use of machine learning in healthcare and public health contexts; therefore, everyone is entitled to information about how these methods function.
Transparency in health-related automation
Disclosures that enable transparency in model and dataset reporting are necessary in mitigating health disparities resulting from over-reliance on unchecked automation in healthcare and public health contexts.
Information as a public good
Information that serves to benefit the public is a public good.
All original and external materials curated for MDSD4Health.com have been selected to enable free access. All readings, exercises, and webpages on this website are available at no cost.
Note: Because much of our material is curated from external sources, the availability and access of these materials may change over time. If you find that an external link is broken or a resource is suddenly stuck behind a paywall, please let us know here!