Winning Privately: The Differentially Private Lottery Ticket Mechanism

Abstract

We propose the differentially private lottery ticket mechanism (DPLTM). An end-to-end differentially private training paradigm based on the lottery ticket hypothesis. Using “high-quality winners”, selected via our custom score function, DPLTM significantly outperforms state-of-the-art. We show that DPLTM converges faster, allowing for early stopping with reduced privacy budget consumption. We further show that the tickets from DPLTM are transferable across datasets, domains, and architectures. Our extensive evaluation on several public datasets provides evidence to our claims.

Publication
Workshop on Machine Learning with Guarantees
33rd Conference on Neural Information Processing Systems (NeurIPS 2019).

This is a work in progress presented at the Workshop on Machine Learning with Guarantees, part of NeurIPS 2019, as a POSTER.

comments powered by Disqus