AIF-C01 Question 206
Select 3A data scientist is building a machine learning model using Amazon SageMaker and wants to ensure the origins of the training data are well-documented for future auditing and reproducibility. Which of the following AWS services or features should the data scientist use to achieve this?
- A
SageMaker Model Cards
- B
AWS Glue Data Catalog
- C
Amazon CloudWatch Logs
- D
SageMaker Feature Store
- E
AWS Cost Explorer
Show answer and explanation
Correct answers: A, B, D
Explanation
To ensure data origins are well-documented, tools like SageMaker Model Cards, AWS Glue Data Catalog, and SageMaker Feature Store provide capabilities for tracking data lineage and organizing metadata. These tools are specifically designed to enhance transparency and reproducibility in machine learning workflows. CloudWatch Logs and AWS Cost Explorer do not serve this purpose.
- A. Correct.
SageMaker Model Cards are used to document critical details about machine learning models, including data origins, training processes, and performance metrics, which are vital for transparency and compliance.
- B. Correct.
AWS Glue Data Catalog provides a centralized metadata repository for datasets, making it easier to track and document data origins and lineage.
- C. Incorrect.
Amazon CloudWatch Logs is designed for monitoring and logging system activities but does not provide functionality for tracking data origins or documenting data lineage.
- D. Correct.
SageMaker Feature Store helps manage and store features used in machine learning models and can track the lineage of these features, contributing to data origin documentation.
- E. Incorrect.
AWS Cost Explorer focuses on cost management and billing insights and is unrelated to documenting data origins or lineage.