AutoML and ML Runtime MACHINE-LEARNING-ASSOCIATE Practice Questions (Page 4)

Questions 16–20

1. 16expert · hard

   A team is evaluating whether to adopt AutoML for their ML workflows. They have a team of experienced data scientists who are concerned about losing control over the model development process. What is the best way to address their concerns?

   • AUse AutoML to generate baseline models, then allow data scientists to refine them manually.
   • BUse AutoML only for simple problems and manual for complex ones.
   • CAvoid AutoML and continue with manual model development.
   • DUse AutoML exclusively and discourage manual intervention.
   Reveal Solution Discussion Select an answer first
2. 17foundation · easy

   A data scientist is setting up a new cluster for a machine learning project and wants to minimize the time spent on environment setup while ensuring that common ML libraries are available. Which Databricks feature directly addresses this need?

   • ADatabricks ML Runtime
   • BDatabricks Jobs
   • CDatabricks Repos
   • DDatabricks SQL Analytics
   Reveal Solution Discussion Select an answer first
4. 18application · medium

   A small company wants to enable its business analysts to create predictive models without deep programming knowledge. They also want to ensure that the models are built on a consistent, optimized environment. Which combination of Databricks features best supports this?

   • AUse Databricks AutoML for model development and ML Runtime for a pre-configured environment.
   • BUse Databricks SQL to build models directly from data.
   • CUse MLflow to manually track experiments and write custom training code.
   • DUse the standard Databricks Runtime and install libraries as needed.
   Reveal Solution Discussion Select an answer first
5. 19expert · hard

   A team is using AutoML to train models on a large dataset. They are concerned about the reproducibility of results across different runs and want to ensure that the same model is selected each time. What should they do?

   • AUse the standard Databricks Runtime instead of ML Runtime.
   • BSet a fixed seed for AutoML and use the same ML Runtime version for all runs.
   • CRun AutoML multiple times and manually select the most common model.
   • DDisable AutoML and manually train a single model with fixed parameters.
   Reveal Solution Discussion Select an answer first
6. 20expert · hard

   A data scientist is using AutoML to build a model for a dataset with many correlated features. They are concerned that AutoML's feature selection might not handle multicollinearity well. What should they do?

   • ALet AutoML handle feature selection, as it evaluates feature importance and may drop redundant features.
   • BIgnore the issue, as multicollinearity does not affect model performance.
   • CUse a regularization technique like Lasso to automatically handle multicollinearity.
   • DManually remove correlated features before running AutoML.
   Reveal Solution Discussion Select an answer first