activity
20152024
most citedNewton Sketch: A Linear-time Optimization Algorithm with Linear-Quadratic Convergence

38 citations · 76 across the 23 of their papers we have counts for

collaborators

30 papers

cs.LG20212 cited

The Convex Geometry of Backpropagation: Neural Network Gradient Flows Converge to Extreme Points of the Dual Convex Program

Yifei Wang, Mert Pilanci

We study non-convex subgradient flows for training two-layer ReLU neural networks from a convex geometry and duality perspective. We characterize the implicit bias of unregularized…

cs.IT2021

Computational Polarization: An Information-theoretic Method for Resilient Computing

Mert Pilanci

We introduce an error resilient distributed computing method based on an extension of the channel polarization phenomenon to distributed algorithms. The method leverages an algorit…

math.OC20216 cited

Newton-LESS: Sparsification without Trade-offs for the Sketched Newton Update

Michał Dereziński, Jonathan Lacotte, Mert Pilanci +1

In second-order optimization, a potential bottleneck can be computing the Hessian matrix of the optimized function at every iteration. Randomized sketching has emerged as a powerfu…

math.OC20211 cited

Adaptive Newton Sketch: Linear-time Optimization with Quadratic Convergence and Effective Hessian Dimensionality

Jonathan Lacotte, Yifei Wang, Mert Pilanci

We propose a randomized algorithm with quadratic convergence rate for convex optimization problems with a self-concordant, composite, strongly convex objective function. Our method…

cs.LG2021

Training Quantized Neural Networks to Global Optimality via Semidefinite Programming

Burak Bartan, Mert Pilanci

Neural networks (NNs) have been extremely successful across many tasks in machine learning. Quantization of NN weights has become an important topic due to its impact on their ener…

cs.LG2021

Fast Convex Quadratic Optimization Solvers with Adaptive Sketching-based Preconditioners

Jonathan Lacotte, Mert Pilanci

We consider least-squares problems with quadratic regularization and propose novel sketching-based iterative methods with an adaptive sketch size. The sketch size can be as small a…