Google Professional Machine Learning Engineer
Get started today
Ultimate access to all questions.
You have developed a BigQuery ML model that predicts customer churn based on feature values such as customer activity and transaction history. This model is deployed to Vertex AI Endpoints for serving predictions. To maintain the model's accuracy over time, you need to automate the retraining process triggered by changes in model feature values, while minimizing the training costs. What approach should you use to achieve this with minimal additional code?
You have developed a BigQuery ML model that predicts customer churn based on feature values such as customer activity and transaction history. This model is deployed to Vertex AI Endpoints for serving predictions. To maintain the model's accuracy over time, you need to automate the retraining process triggered by changes in model feature values, while minimizing the training costs. What approach should you use to achieve this with minimal additional code?
Explanation:
Option D is the correct answer because it directly addresses the need to monitor training/serving skew, which pertains to discrepancies between the training data and the data being served during production. This is crucial when model feature values change over time. Option D outlines a comprehensive approach by setting up a Vertex AI Model Monitoring job to detect training/serving skew, configuring alerts through Pub/Sub to minimize unnecessary retraining, and utilizing Cloud Functions to automate the retraining process in BigQuery. This ensures minimal additional code is required and retraining is triggered only when necessary, thus optimizing training costs.