Databricks Machine Learning Associate exam dumps

Databricks Machine Learning Associate practice question 91 of 656

Databricks Certified Machine Learning Associate. Associate level, Databricks. Free question with the correct answer and a full explanation.

Databricks Machine Learning Associate Question 91

Select 3

You are working on a machine learning pipeline in Databricks and need to score a model using features from a registered feature store table. Which of the following steps are required to correctly retrieve the features and use them for scoring?

  1. A

    Use the FeatureStoreClient to load the feature table and join it with the input data.

  2. B

    Directly query the feature store table using SQL and use the result for scoring.

  3. C

    Ensure the primary keys in the input data match the primary keys in the feature store table.

  4. D

    Use the FeatureStoreClient to log the model before scoring.

  5. E

    Perform feature lookups using the FeatureStoreClient to retrieve the required feature values.

Show answer and explanation

Correct answers: A, C, E

Explanation

To correctly score a model using features from a feature store table in Databricks, it is crucial to use the FeatureStoreClient to handle feature lookups and joins programmatically. Additionally, the input data must have primary keys that match those in the feature table to ensure the correct features are retrieved. Direct SQL queries or unrelated actions like logging a model do not align with the recommended workflow for leveraging the feature store in Databricks.

  • A. Correct.

    Correct. The FeatureStoreClient is required to programmatically retrieve features from a feature store table. This ensures the features are retrieved in a consistent and scalable manner.

  • B. Incorrect.

    Incorrect. While querying the feature store table directly using SQL may seem feasible, it bypasses the feature store's built-in capabilities like versioning and lineage tracking, which are crucial in a production ML workflow.

  • C. Correct.

    Correct. The input data must have matching primary keys with the feature store table to ensure the correct feature values are retrieved for each input instance.

  • D. Incorrect.

    Incorrect. Logging a model using the FeatureStoreClient is part of model management workflows but is not needed for scoring a model using features from the feature store.

  • E. Correct.

    Correct. The FeatureStoreClient provides a method for performing feature lookups, ensuring the correct feature values are joined with the input data for scoring.

Timed practice exam

Take a Databricks Machine Learning Associate practice test under exam conditions

48 questions in 90 minutes, drawn from this bank, with a score report and a per-question review when you finish.

Start timed exam