Sharing our latest differential privacy milestones and advancements- Google Developers Blog

What this post added

Announces the largest application of differential privacy in the world (spanning close to three billion devices) to improve products like Google Home, Google Search on Android, and Messages. Details the 'shuffler' model research that enabled this scale. Releases PipelineDP for Java Virtual Machine (JVM) called PipelineDP4j, an evolution of joint work with OpenMined, to reduce the barrier for Java developers. Introduces new differential privacy algorithms powering Google Trends for greater insights into low-volume locales. Releases DP-Auditorium, a library for auditing differentially private algorithms without requiring access to internal properties of the application, which outperforms existing black-box access testers.