Fast and Memory Efficient Differentially Private-SGD via JL Projections

774 views

About this video

Hand-picked by a BrighterTube curator · found via round-3 back catalogue of a channel Lumen picks video by video · 82% confidence

Fast and Memory Efficient Differentially Private-SGD via JL Projections teaches optimization using Johnson-Lindenstrauss projections to reduce memory while maintaining privacy. Presents concrete algorithmic improvements with theoretical analysis.

A 54-minute video from Google TechTalks: “Fast and Memory Efficient Differentially Private-SGD via JL Projections”.

Video by Google TechTalks, played through YouTube's embedded player.

Part of Differential Privacy for ML