Google Professional Machine Learning Engineer Question 440
Select 2Google Cloud PlatformYou are working on a machine learning project to predict customer churn. Your team has decided to use Vertex AI Pipelines to automate the end-to-end workflow. To better track and analyze the lineage of ML artifacts, you want to use Vertex ML Metadata. Which of the following actions can you perform using Vertex ML Metadata?
- A
Track the lineage of datasets, models, and metrics used in the pipeline
- B
Visualize the performance metrics of your ML model over multiple pipeline runs
- C
Manage access control for datasets stored in Google Cloud Storage
- D
Query metadata to identify which dataset version was used for a specific ML model
- E
Automatically deploy the best-performing model to an endpoint
Show answer and explanation
Correct answers: A, D
Explanation
Vertex ML Metadata enables tracking, querying, and analyzing the lineage of datasets, models, and metrics in your ML pipelines. This helps with reproducibility and auditing. However, it does not handle visualization, access control, or automated deployment, as these functionalities fall under other services or features within the Google Cloud ecosystem.
- A. Correct.
Correct. Vertex ML Metadata is designed to track the lineage of datasets, models, and metrics, making it easier to understand and audit the data flow in ML pipelines.
- B. Incorrect.
Incorrect. While you can store and query performance metrics in Vertex ML Metadata, it does not provide a direct visualization feature. Visualization would need to be handled through other tools or custom solutions.
- C. Incorrect.
Incorrect. Vertex ML Metadata does not handle access control for datasets; this is managed separately, such as through IAM policies or Cloud Storage permissions.
- D. Correct.
Correct. Vertex ML Metadata allows you to query metadata to trace which dataset version was used for a specific model, providing better reproducibility and debugging capabilities.
- E. Incorrect.
Incorrect. Vertex ML Metadata focuses on metadata tracking and query capabilities. Deploying models is handled through Vertex AI's deployment features, not Metadata.