PROFESSIONAL-MACHINE-LEARNING-ENGINEER Question #241: Real Exam Question with Answer & Explanation

Question

You are training models in Vertex AI by using data that spans across multiple Google Cloud projects. You need to find, track, and compare the performance of the different versions of your models. Which Google Cloud services should you include in your ML workflow?

Options

• ADataplex, Vertex AI Feature Store, and Vertex AI TensorBoard
• BVertex AI Pipelines, Vertex AI Feature Store, and Vertex AI Experiments
• CDataplex, Vertex AI Experiments, and Vertex AI ML Metadata
• DVertex AI Pipelines, Vertex AI Experiments, and Vertex AI Metadata

Explanation

To track and compare the performance of different model versions across projects, Vertex AI Experiments and Vertex AI ML Metadata are essential for logging and tracking, while Dataplex can manage distributed data.

Common mistakes.

• A. Vertex AI Feature Store manages features, not model performance tracking; Vertex AI TensorBoard provides visualization, but not the core tracking and comparison mechanism required.
• B. Vertex AI Pipelines orchestrates ML workflows, but the core requirement is tracking and comparing model performance and versions, which is handled by Experiments and Metadata, not Pipelines or Feature Store directly.
• D. Vertex AI Pipelines orchestrates workflows, but is not the primary service for tracking and comparing model performance and versions; while Metadata is relevant, Vertex AI Experiments is key for comparison. Dataplex is needed for managing data across projects, which is missing here.

Concept tested. Model tracking, comparison, and metadata management

Reference. https://cloud.google.com/vertex-ai/docs/experiments/track-experiments

No community discussion yet for this question.