Expanding Differentially Private Solutions: A Python Case Study
summary
PipelineDP is a Python library that implements end-to-end differentially private pipeline primitives (counts, sums, joins and per-user contributions) and a privacy accountant so engineers can run the same code locally or at scale on Spark and Beam. The talk shows concrete engineering choices: how contribution bounding and noise mechanisms are coded, privacy parameter handling, integration with data-access controls, and that the tool is open-source and deployed to reduce manual DP mistakes in real pipelines.
AI-written summary · about
related talks
13mUtility Analysis for Differentially-Private Pipelines
13mEvaluating the Usability of Differential Privacy Tools with Data Practitioners
27mYou've got trust issues, we've got solutions: Differential Privacy
14mEstablishing Privacy Metrics for Genomic Data Analysis
01h 14mTutorial: Privacy, Part II
35mDifferential Privacy Made Practical
21m"SECRETS ARE LIES, SHARING IS CARING, PRIVACY IS THEFT." - A Dive into Privacy Preserving
31mScaling privacy embedded distributed ML systems