Analytics & Data Science with Privacy at its Core
summary
Claims a platform-first solution that separates three data zones—production, analytics, and test/dev—and applies specific techniques to let analytics run on desensitized data without leaking PII. Describes secure multi-party computation, fully homomorphic encryption, and differential privacy as the core math-heavy tools, plus format-preserving masking, tokenization, data clean rooms, and ML-driven policy extraction. Names IBM Cloud Pak for Data and a HADES desensitization engine for auditability and risk assessment.
AI-written summary · about
chapters (9)
- 0:00Introduction to privacy analytics
- 2:11Privacy as a discipline
- 5:08Advanced privacy techniques
- 10:01Platform for data desensitization
- 12:32Three zones of data treatment
- 17:53Masking and tokenization
- 20:03ML-driven compliance and policies
- 23:16Risk assessment and auditability
- 25:47Closing and resources
related talks
01h 59mDeep Dive Workshop – Data Security and Data Protection
39mData Privacy Engineering to power Responsible Enterprise AI
01h 19mWhen Data Can’t Move but Insight Must
28mData Engineers: Privacy Is Your Problem
24mSafeguarding Sensitive Data Access At Scale with Privacy-Enhancing Technologies
40mEnsuring Data Security in the AI Revolution
30mSharing without Sharing: the Case for Data Clean Rooms Powered by Confidential Computing
22mData Privacy & Compliance in a integrated Hybrid Cloud Environment