AWS Clean Rooms ML now supports redacted error log summaries

AWS Clean Rooms ML has introduced a new privacy feature for custom modeling that allows collaborators to share redacted error log summaries securely.

• Enables training and inference on custom ML models using collective datasets without sharing sensitive data
• New privacy control allows sending redacted error log summaries to specified collaboration members
• Error logs include exception type, error message, and code line with error
• Collaborators can configure which members receive summaries and control redaction of PII, numbers, and custom strings
• Maintains data and model privacy while generating predictive insights

This feature enhances collaborative ML model development by providing detailed error information while protecting sensitive data.