Sparsity in neural networks can improve their privacy

no code implementations • 20 Apr 2023 • Antoine Gonon, Léon Zheng, Clément Lalanne, Quoc-Tung Le, Guillaume Lauga, Can Pouliquen

This article measures how sparsity can make neural networks more robust to membership inference attacks.

Can sparsity improve the privacy of neural networks?

no code implementations • 11 Apr 2023 • Antoine Gonon, Léon Zheng, Clément Lalanne, Quoc-Tung Le, Guillaume Lauga, Can Pouliquen

Sparse neural networks are mainly motivated by ressource efficiency since they use fewer parameters than their dense counterparts but still reach comparable accuracies.