Custody

From The Foundation for Best Practices in Machine Learning

Custody[edit]

Control

Develop a system for documenting the chain of custody for Product(s) and the data, microservices, and applications that it is built on and with, that indicates: i) provenance ii) control iii) transfer, iv) analysis, and v) transformation.


Aim

To (a) ensure that the building blocks of the Product can be traced back to their origins; (b) allow for undesirable changes to be reverted; and (c) highlight associated risks that might occur in the Product Lifecycle.


Additional Information[edit]