
Accuracy on the Line: On the Strong Correlation Between OutofDistribution and InDistribution Generalization
For machine learning systems to be reliable, we must understand their pe...
read it

Never Go Full Batch (in Stochastic Convex Optimization)
We study the generalization performance of fullbatch optimization algor...
read it

Stochastic BiasReduced Gradient Methods
We develop a new primitive for stochastic optimization: a lowbias, low...
read it

Thinking Inside the Ball: NearOptimal Minimization of the Maximal Loss
We characterize the complexity of minimizing max_i∈[N] f_i(x) for convex...
read it

LargeScale Methods for Distributionally Robust Optimization
We propose and analyze algorithms for distributionally robust optimizati...
read it

Coordinate Methods for Matrix Games
We develop primaldual coordinate methods for solving bilinear saddlepo...
read it

SecondOrder Information in NonConvex Stochastic Optimization: Power and Limitations
We design an algorithm which finds an ϵapproximate stationary point (wi...
read it

Acceleration with a Ball Optimization Oracle
Consider an oracle which takes a point x and returns the minimizer of a ...
read it

Lower Bounds for NonConvex Stochastic Optimization
We lower bound the complexity of finding ϵstationary points (with gradi...
read it

Variance Reduction for Matrix Games
We present a randomized primaldual algorithm that solves the problem _x...
read it

Unlabeled Data Improves Adversarial Robustness
We demonstrate, theoretically and empirically, that adversarial robustne...
read it

A Rank1 Sketch for Matrix Multiplicative Weights
We show that a simple randomized sketch of the matrix multiplicative wei...
read it

No bad local minima: Data independent training error guarantees for multilayer neural networks
We use smoothed analysis techniques to provide guarantees on the trainin...
read it
Yair Carmon
is this you? claim profile