Summary

There are many attacks to be prepared for and vulnerabilities are discovered daily, so we must follow a framework that helps us keep up to date with current vulnerabilities and their mitigations where possible. The MITRE ATLAS framework is a great resource to get started as it is adapted to ML. We need to be aware of the 12 stages and multiple techniques per stage to protect our ML assets. However, as ML assets work with numerous other systems, the implementations we will see in the following chapters will include securing Azure Machine Learning and all its related services.

But before diving into those implementations, in the next chapter, we will learn about the security industry compliance standards we must adhere to and how to implement compliance controls together with responsible AI development practices.