Model governance: reducing the anarchy of production ML

Abstract

As the influence of machine learning grows over decisions in businesses and human life, so grows the need for Model Governance. In this paper, we motivate the need for, define the problem of, and propose a solution for Model Governance in production ML. We show that through our approach one can meaningfully track and understand the who, where, what, when, and how an ML prediction came to be. To the best of our knowledge, this is the first work providing a comprehensive framework for production Model Governance, building upon previous work in developer-focused Model Management.