Describe Azure Machine Learning Capabilities AI-900 Practice Questions (Page 5)

Questions 21–25

1. 21foundation · easy

   Which benefit does automated machine learning (AutoML) provide to data scientists?

   • AIt guarantees the model will be 100% accurate on new data.
   • BIt automatically deploys the best model to a production endpoint.
   • CIt eliminates the need for data cleaning and feature engineering.
   • DIt reduces the time and expertise needed to experiment with many models and hyperparameters.
   Reveal Solution Discussion Select an answer first
2. 22application · medium

   A healthcare startup wants to use Automated ML to build a classification model that predicts patient readmission risk. They have a labeled dataset in a CSV file stored in Azure Blob Storage. The team needs to track the experiment, compare multiple runs, and register the best model for future use. What is the correct sequence of steps?

   • AUse Azure Data Factory to copy data to Azure SQL Database, then run Automated ML directly on the SQL database without creating a dataset.
   • BCreate a compute instance, upload data to the compute instance's local disk, run Automated ML, and then delete the compute instance.
   • CCreate a workspace, register the model first, then run Automated ML on the registered model to improve it.
   • DCreate a workspace, create a compute cluster, upload data to a datastore, create an Automated ML job, review the best model, and register it.
   Reveal Solution Discussion Select an answer first
4. 23foundation · easy

   Which configuration setting is required when setting up an automated machine learning (AutoML) experiment in Azure Machine Learning?

   • AThe model explanation method
   • BThe target column to predict
   • CThe data drift monitoring schedule
   • DThe deployment target (e.g., AKS or ACI)
   Reveal Solution Discussion Select an answer first
5. 24expert · hard

   A team needs to run automated ML experiments that may take up to 24 hours. They also need to deploy the best model to a production endpoint with autoscaling. Which compute strategy should they use?

   • AUse a dedicated compute cluster for training and AKS for deployment
   • BUse a compute instance for training and AKS for deployment
   • CUse a low-priority cluster for training and AKS for deployment
   • DUse a low-priority cluster for training and ACI for deployment
   Reveal Solution Discussion Select an answer first
6. 25application · medium

   A model deployed to an AKS endpoint is returning high latency. The team wants to identify the cause. What should they use?

   • ADatastore
   • BModel registry
   • CPipeline
   • DApplication Insights
   Reveal Solution Discussion Select an answer first