Google Professional Machine Learning Engineer exam dumps

Google Professional Machine Learning Engineer practice question 488 of 522

Professional Machine Learning Engineer. Professional level, Google Cloud. Free question with the correct answer and a full explanation.

Google Professional Machine Learning Engineer Question 488

Select 3Google Cloud Platform

You have deployed a machine learning model to production using Vertex AI. To ensure the model remains accurate over time, you decide to set up continuous evaluation. Which of the following steps are required to establish continuous evaluation with Vertex AI Model Monitoring?

  1. A

    Enable model monitoring and define the target metrics to monitor (e.g., accuracy, precision).

  2. B

    Define a skew detection threshold for features and labels based on historical training data statistics.

  3. C

    Deploy a custom container for monitoring that sends alerts when model performance degrades.

  4. D

    Configure automated data labeling to generate ground truth data for evaluating predictions.

  5. E

    Specify a BigQuery table or Cloud Storage bucket for storing prediction and ground truth data.

Show answer and explanation

Correct answers: A, B, E

Explanation

To establish continuous evaluation with Vertex AI Model Monitoring, you need to enable model monitoring, configure the metrics to monitor, and define thresholds for feature or label skew. You also need to specify a storage location (e.g., BigQuery or Cloud Storage) for logging prediction and ground truth data. These steps ensure the model is continuously evaluated for performance issues or data distribution changes over time.

  • A. Correct.

    Correct. Enabling model monitoring and defining the target metrics is a necessary step to establish continuous evaluation using Vertex AI Model Monitoring.

  • B. Correct.

    Correct. Skew detection thresholds help monitor for data distribution changes that may impact model performance, which is a critical part of continuous evaluation.

  • C. Incorrect.

    Incorrect. Deploying a custom container is not required for Vertex AI Model Monitoring. It provides built-in capabilities to monitor model performance.

  • D. Incorrect.

    Incorrect. Automated data labeling is not a built-in requirement for Vertex AI Model Monitoring; ground truth data is typically gathered through other processes.

  • E. Correct.

    Correct. Specifying a BigQuery table or Cloud Storage bucket is required to store the input data, predictions, and ground truth data for evaluation.

Timed practice exam

Take a Google Professional Machine Learning Engineer practice test under exam conditions

60 questions in 120 minutes, drawn from this bank, with a score report and a per-question review when you finish.

Start timed exam