Exploring Azure services involved in ML attacks
As you can see, attacks are multi-level and primarily based on the adversary’s goal. Since we do not know what that is, we can deploy multiple mitigation techniques to lessen the impact. As Azure Machine Learning is based on the Azure platform, we can deploy numerous tools to detect an incident, and by using automation, the platform will deploy mitigation steps before we are even aware that something has happened. Although we focused on ML attacks, attacks on related systems, virtual machines, and databases are still a concern. Let us look at associated services that can be used together with Azure Machine Learning.
Access
Microsoft Entra ID usually handles access in Azure. Microsoft Entra ID is Microsoft’s cloud-based identity and access management service. It provides a range of features and capabilities to manage user identities and secure access to various resources in the Azure cloud and other Microsoft services...
