Google Professional Machine Learning Engineer Question 140
Select 3Google Cloud PlatformYou are tasked with prototyping a machine learning model using Jupyter notebooks on Google Cloud. Which considerations should you prioritize to ensure the prototyping process is efficient and scalable?
- A
Use pre-built datasets from Google Cloud's public datasets to save time on data preparation.
- B
Ensure that the notebook is running on a local machine to reduce latency.
- C
Leverage Vertex AI Workbench to integrate the notebook with Google Cloud services for scalability.
- D
Perform hyperparameter tuning directly in the notebook to finalize the model during prototyping.
- E
Use BigQuery magic commands to query large datasets directly within the notebook.
Show answer and explanation
Correct answers: A, C, E
Explanation
When prototyping machine learning models using Jupyter notebooks on Google Cloud, it is important to leverage tools and services that enhance efficiency and scalability, such as pre-built datasets, Vertex AI Workbench, and BigQuery magic commands. These considerations help streamline the prototyping process and integrate seamlessly with Google Cloud's ecosystem. Running notebooks locally and performing hyperparameter tuning during prototyping are less efficient approaches.
- A. Correct.
Using pre-built datasets from Google Cloud's public datasets can save time and resources during the prototyping phase, especially when you're testing model functionality and performance.
- B. Incorrect.
Running the notebook on a local machine may introduce scalability and resource limitations. It is more efficient to use managed services when dealing with large-scale data.
- C. Correct.
Vertex AI Workbench provides a managed Jupyter notebook environment that integrates seamlessly with Google Cloud services, making it a scalable and efficient choice for prototyping.
- D. Incorrect.
Hyperparameter tuning is computationally expensive and is typically performed after the prototyping phase. During prototyping, the focus should be on model exploration and testing.
- E. Correct.
BigQuery magic commands allow you to query large datasets directly within the notebook, simplifying the workflow and enabling you to handle large-scale data efficiently during prototyping.