trust.auditable
Module Contents
Classes
A Trustworthy-AI interface that defines the capabilities of objects (e.g., Models, Graphs) |
- class trust.auditable.Auditable
A Trustworthy-AI interface that defines the capabilities of objects (e.g., Models, Graphs) that can self-record and report an audit trail of their activities and input/output artifacts. For example, an Auditable Model should be able to record and even retain its training data set, the training data scientist, the timestamp, model metrics, etc. available for query at any time in the future.
- property audit_trailing(self)
- audit_trail_get(self, key)
- audit_trail_set(self, key, value)
- audit_trail_delete(self, key)