Silas Model Audit¶
N.B. This feature is currently restricted in the v0.8.7 version. Please download the v0.8.5 version to explore this feature.
The Silas Model Audit feature allows the user to specify certain constraints, criteria, requirements or specifications using logic and verify that the machine learning model is compliant with the user’s specifications. This features has two functionalities: model verification and enforcement learning. The former is formally verifies machine learning models against user specifications. The latter enforces user specifications in the learning process in order to generate compliant models.