This chapter covers Azure Machine Learning, a platform for building custom ML models. AZ-900 tests this under objective 2.2 at a conceptual level.
Jump to a section
A simple way to picture Azure Machine Learning Overview
Cognitive Services are pre-built AI tools. Azure Machine Learning is a workshop where you build custom AI models — it provides tools, data management, and training infrastructure, but you design and train the models yourself based on your specific data and needs.
What Azure ML does
Machine Learning provides tools for building, training, and deploying custom ML models — useful when Cognitive Services' pre-trained models don't fit specific needs.
Data and training infrastructure
Azure ML provides data management, experiment tracking, and compute resources for training models.
When to use Azure ML vs. Cognitive Services
Cognitive Services fit pre-trained scenarios; Azure ML fits when you need to build a custom model from your own data.
Prepare data
Training data is loaded and prepared.
Choose an algorithm
An appropriate ML algorithm is selected.
Train the model
The model is trained on the data.
Deploy
The trained model is deployed as a service.
An organization trains a custom model on its historical data to predict customer churn, more tailored than a generic pre-trained model.
Objective 2.2 expects conceptual understanding that Azure ML is for building custom models, distinct from pre-trained Cognitive Services.
Common wrong answer: confusing Azure ML with Cognitive Services — Cognitive Services are pre-trained; Azure ML is for custom model building.
Stable terms: Azure Machine Learning, custom model. Memory trick: Azure ML = build custom models; Cognitive Services = use pre-trained models.
Azure ML is a platform for building custom ML models.
Provides data management, training infrastructure, experiment tracking.
Use when pre-trained models (Cognitive Services) don't fit your needs.
Mistake
Azure ML and Cognitive Services do the same thing.
Correct
Cognitive Services are pre-trained; Azure ML is for building custom models.
It's a platform for building, training, and deploying custom ML models from your own data, useful when pre-trained services don't fit specific needs.
You've just covered Azure Machine Learning Overview — now see how well it sticks with free AZ-900 practice questions. Full explanations included, no account needed.
Done with this chapter?