Search Results for author:
Found 11 papers, 5 papers with code
How to Make the Gradients Small Privately: Improved Rates for Differentially Private Non-Convex Optimization
We use this framework to obtain improved, and sometimes optimal, rates for several classes of non-convex loss functions.
Why Does Differential Privacy with Large Epsilon Defend Against Practical Membership Inference Attacks?
In practical applications, such a worst-case guarantee may be overkill: practical attackers may lack exact knowledge of (nearly all of) the private data, and our data set might be easier to defend, in some sense, than the worst-case data set.
Optimal Differentially Private Model Training with Public Data
We show that the optimal error rates can be attained (up to log factors) by either discarding private data and training a public model, or treating public data like it is private and using an optimal DP algorithm.
Stochastic Differentially Private and Fair Learning
However, existing algorithms for DP fair learning are either not guaranteed to converge or require full batch of data in each iteration of the algorithm to converge.
Private Stochastic Optimization With Large Worst-Case Lipschitz Parameter: Optimal Rates for (Non-Smooth) Convex Losses and Extension to Non-Convex Losses
To address these limitations, this work provides near-optimal excess risk bounds that do not depend on the uniform Lipschitz parameter of the loss.
Private Non-Convex Federated Learning Without a Trusted Server
silo data and two classes of Lipschitz continuous loss functions: First, we consider losses satisfying the Proximal Polyak-Lojasiewicz (PL) inequality, which is an extension of the classical PL condition to the constrained setting.
Private Federated Learning Without a Trusted Server: Optimal Algorithms for Convex Losses
This paper studies federated learning (FL)--especially cross-silo FL--with data from people who do not trust the server or other silos.
A Stochastic Optimization Framework for Fair Risk Minimization
We consider the problem of fair classification with discrete sensitive attributes and potentially large models and data sets, requiring stochastic solvers.
Output Perturbation for Differentially Private Convex Optimization with Improved Population Loss Bounds, Runtimes and Applications to Private Adversarial Training
Finally, we apply our theory to two learning frameworks: tilted ERM and adversarial learning.
Efficient Search of First-Order Nash Equilibria in Nonconvex-Concave Smooth Min-Max Problems
As a byproduct, the choice $\varepsilon_y = O(\varepsilon_x{}^2)$ allows for the $O(\varepsilon_x{}^{-3})$ complexity of finding an $\varepsilon_x$-stationary point for the standard Moreau envelope of the primal function.
Optimization and Control 90C06, 90C25, 90C26, 91A99
