Google Professional Machine Learning Engineer Question 442
Select 3Google Cloud PlatformYou are building a machine learning pipeline on Google Cloud, and you want to track the lineage of your datasets, models, and experiments to improve reproducibility and manage metadata effectively. Which features of Vertex ML Metadata would you use to achieve this?
- A
Track lineage of artifacts, such as datasets and models, across pipeline executions
- B
Store and query custom metadata related to experiments and model training
- C
Automatically deploy trained models to Vertex AI endpoints without additional steps
- D
Visualize the relationships between pipeline components, such as executions and artifacts
- E
Perform hyperparameter tuning using pre-configured metadata templates
Show answer and explanation
Correct answers: A, B, D
Explanation
Vertex ML Metadata is designed to help machine learning engineers track, store, and analyze metadata related to their ML workflows. Key features include tracking artifact lineage, storing and querying custom metadata, and visualizing relationships between pipeline components. These capabilities improve workflow reproducibility and provide better insights into the lifecycle of datasets, models, and experiments.
- A. Correct.
Correct. Vertex ML Metadata allows you to track the lineage of artifacts, such as datasets and models, to ensure reproducibility and better manage your ML workflows.
- B. Correct.
Correct. Vertex ML Metadata can store and query custom metadata about your experiments and training processes, enabling detailed tracking and analysis.
- C. Incorrect.
Incorrect. While Vertex AI can deploy models to endpoints, this is not a feature of Vertex ML Metadata itself. Metadata focuses on tracking and storing information, not deployment.
- D. Correct.
Correct. Vertex ML Metadata provides tools to visualize relationships between pipeline components, helping users understand the flow and dependencies within their ML workflows.
- E. Incorrect.
Incorrect. Hyperparameter tuning is not a feature of Vertex ML Metadata; it is managed separately in Vertex AI using tools like hyperparameter tuning jobs.