Probing the Transition to Dataset-Level Privacy in ML Models Using an Output-Specific and Data-Resolved Privacy Profile

no code implementations • 27 Jun 2023 • Tyler LeBlond, Joseph Munoz, Fred Lu, Maya Fuchs, Elliott Zaresky-Williams, Edward Raff, Brian Testa

Differential privacy (DP) is the prevailing technique for protecting user data in machine learning models.

A General Framework for Auditing Differentially Private Machine Learning

no code implementations • 16 Oct 2022 • Fred Lu, Joseph Munoz, Maya Fuchs, Tyler LeBlond, Elliott Zaresky-Williams, Edward Raff, Francis Ferraro, Brian Testa

We present a framework to statistically audit the privacy guarantee conferred by a differentially private machine learner in practice.

An Algorithm for Approximating Continuous Functions on Compact Subsets with a Neural Network with one Hidden Layer

no code implementations • 10 Feb 2019 • Elliott Zaresky-Williams

George Cybenko's landmark 1989 paper showed that there exists a feedforward neural network, with exactly one hidden layer (and a finite number of neurons), that can arbitrarily approximate a given continuous function $f$ on the unit hypercube.