MLA-C01 Question 360
Select 3A financial services company has deployed a machine learning model for credit risk assessment using Amazon SageMaker. The model's predictions are critical for regulatory compliance. The company wants to ensure that the model's performance does not degrade over time and that it continues to make accurate predictions, especially when the input data distribution changes. Which combination of actions should the company take to monitor and maintain the model effectively?
- A
Enable Amazon SageMaker Model Monitor to detect data drift and anomalies in the input features.
- B
Set up an Amazon CloudWatch alarm to trigger when the inference endpoint's latency exceeds a specified threshold.
- C
Periodically retrain the model using a pipeline in Amazon SageMaker Pipelines with fresh labeled data.
- D
Enable Amazon SageMaker Clarify to monitor feature importance changes in real-time.
- E
Use an Amazon EventBridge rule to trigger an AWS Lambda function for model retraining whenever data drift is detected.
Show answer and explanation
Correct answers: A, C, E
Explanation
To ensure the deployed model maintains its performance and compliance, the company needs to monitor data drift and take corrective actions when necessary. SageMaker Model Monitor enables automated detection of data drift, while periodic retraining with updated data ensures the model stays aligned with the latest patterns. Automating retraining workflows with EventBridge and Lambda adds efficiency and scalability to the maintenance process. Monitoring endpoint latency, while useful for operational health, does not address model performance issues. SageMaker Clarify is focused on bias detection and explainability rather than real-time feature monitoring.
- A. Correct.
This is correct. Amazon SageMaker Model Monitor can be used to detect data drift, anomalies in input data, and other issues that can impact model performance over time.
- B. Incorrect.
This option is not correct. While monitoring endpoint latency using CloudWatch is a best practice for operational health, it does not directly address model performance or data drift.
- C. Correct.
This is correct. Periodic retraining with fresh data can ensure the model adapts to evolving data patterns and maintains performance.
- D. Incorrect.
This option is not correct. Amazon SageMaker Clarify is primarily used for detecting biases and explaining model predictions, not for monitoring feature importance changes in real-time.
- E. Correct.
This is correct. An EventBridge rule combined with a Lambda function can automate retraining workflows when data drift is detected, enhancing the model's maintenance process.