Q: 18
HOTSPOT You must use in Azure Data Science Virtual Machine (DSVM) as a compute target. You need to attach an existing DSVM to the workspace by using the Azure Machine Learning SDK for Python. How should you complete the following code segment? To answer, select the appropriate options in the answer area. NOTE: Each correct selection is worth one point. 
Your Answer
Discussion
attach_configuration for the first box, attach for the second. You use attach_configuration to set up how the DSVM connects, then attach to register it with your workspace. Pretty sure this is standard in AML SDK when adding existing compute. Let me know if anyone found a different method.
Actually, I think create_configuration for Box 1 and attach for Box 2 seems right. In some SDK patterns, you use create_configuration before registering existing compute. But this might mix up with AMLCompute, so not 100% confident. Anyone else run into this trap?
Error rate shift, centroid distance, then combine penalty data. Pretty sure that's right, penalty last is a common trap in these.
Evaluate error rate change, then analyze with centroid distance, finally combine penalty detection-saw something like this in a mock.
Be respectful. No spam.