Module Overview

In this module, we discuss the role of disclosure in ML for healthcare and public health contexts. Disclosure practices within the realm of ML refer to deliberate communication of an ML model's origins, including (but not limited to) its known development limitations, performance metrics, intended uses, and training and testing dataset composition. The development of robust disclosure methods and mediums have been proposed as a central strategy in the pursuit of achieving algorithmic transparency in healthcare and public health contexts.

Disclosures offered by ML practitioners serve a variety of purposes. We have broadly categorized several of them into two (overlapping) groups:

1. Disclosure for methods replicability and reproducibility

2. Disclosure for model generalization

We cover both of these topics in this module and end with a brief primer on model and dataset disclosure methods mediums.