Google Professional Machine Learning Engineer Question 59
Select 3Google Cloud PlatformYou are tasked with building a machine learning model to predict customer churn for a subscription-based service. Your organization has limited machine learning expertise and wants to minimize manual feature engineering. You decide to use Google Cloud AutoML Tables. Which of the following are key considerations to ensure successful model training using AutoML Tables?
- A
Ensure your dataset is properly pre-processed and cleaned before uploading to AutoML Tables.
- B
Provide a pre-trained TensorFlow model to AutoML Tables for transfer learning.
- C
Verify that your dataset is split into training, validation, and test sets before training begins.
- D
Understand that AutoML Tables automatically handles feature engineering and hyperparameter tuning.
- E
Ensure your dataset has at least 1,000 rows for effective training results.
Show answer and explanation
Correct answers: A, D, E
Explanation
When using AutoML Tables, it is crucial to provide clean and well-prepared data while leveraging its automated capabilities like feature engineering, hyperparameter tuning, and automatic dataset splitting. Additionally, ensuring a sufficient dataset size is important for effective training. These considerations help maximize the performance and usability of models trained using AutoML Tables.
- A. Correct.
AutoML Tables requires clean and well-prepared data for optimal performance. While AutoML handles feature engineering, it does not clean raw or inconsistent data automatically.
- B. Incorrect.
AutoML Tables does not allow you to upload a pre-trained model for transfer learning. It builds models from your dataset directly.
- C. Incorrect.
AutoML Tables does not require you to pre-split your dataset into training, validation, and test sets. It automatically handles dataset splitting during the training process.
- D. Correct.
AutoML Tables automates feature engineering and hyperparameter tuning, reducing the need for manual intervention.
- E. Correct.
For optimal results, AutoML Tables recommends datasets with at least 1,000 rows to ensure sufficient training data and better model generalization.