Google Professional Machine Learning Engineer Question 131
Select 3Google Cloud PlatformYou are building a machine learning pipeline on Google Cloud to process medical records containing protected health information (PHI). Your pipeline includes data ingestion, preprocessing, training, and serving. How can you ensure compliance with data privacy regulations such as HIPAA while minimizing the risk of exposing sensitive information?
- A
Use Cloud Data Loss Prevention (DLP) to identify and redact PHI before storing the data.
- B
Ensure that all data is stored and processed in encrypted storage systems such as Cloud Storage with Customer-Managed Encryption Keys (CMEK).
- C
Disable logging for all data processing pipelines to avoid recording sensitive information.
- D
Use Vertex AI's Explainable AI to validate predictions without exposing any sensitive data from the training set.
- E
Implement role-based access control (RBAC) and audit logging to restrict and monitor access to sensitive data.
Show answer and explanation
Correct answers: A, B, E
Explanation
When handling PHI and PII in machine learning pipelines, compliance with data privacy regulations such as HIPAA is crucial. Techniques such as using Cloud DLP for redacting sensitive data, encrypting data storage with CMEK, and implementing RBAC with audit logging help ensure data privacy and security. Disabling logging entirely is not recommended, as it can hinder operational monitoring and debugging, and Explainable AI does not directly address data privacy concerns.
- A. Correct.
Correct: Cloud Data Loss Prevention (DLP) helps identify and redact sensitive information, such as PHI, which is crucial for ensuring compliance with data privacy regulations.
- B. Correct.
Correct: Storing and processing data in encrypted storage, such as Cloud Storage with CMEK, ensures the confidentiality of sensitive information and aligns with best practices for handling PHI.
- C. Incorrect.
Incorrect: While disabling logging might prevent sensitive data from being recorded, it could also hinder debugging and monitoring efforts. Instead, sensitive information should be redacted before logging.
- D. Incorrect.
Incorrect: Vertex AI's Explainable AI is useful for understanding model predictions, but it does not directly address the handling or protection of PHI and PII in the pipeline.
- E. Correct.
Correct: Role-based access control (RBAC) and audit logging are essential for restricting access to sensitive data and maintaining compliance with regulations like HIPAA.