Chapter 11: ML Governance, Bias, Explainability, and Privacy

So far, you have successfully implemented a machine learning (ML) platform. At this point, you might be thinking that your job is done as an ML Solutions Architect (ML SA) and that the business is ready to deploy models into production. Well, it turns out that there are additional considerations. To put models into production, an organization also needs to put governance control in place to meet both the internal policy and external regulatory requirements. ML governance is usually not the responsibility of an ML SA; however, it is important for an ML SA to be familiar with the regulatory landscape and ML governance framework, especially in regulated industries, such as financial services. So, you should consider these requirements when you evaluate or build an ML solution.

In this chapter, we will provide an overview of the ML governance concept and some key components, such as model registry and model monitoring, in...