1 code implementation • NeurIPS 2023 • Gaspard Beugnot, Julien Mairal, Alessandro Rudi
We present a novel approach to non-convex optimization with certificates, which handles smooth functions on the hypercube or on the torus.
no code implementations • 28 Feb 2022 • Gaspard Beugnot, Julien Mairal, Alessandro Rudi
This paper studies an intriguing phenomenon related to the good generalization performance of estimators obtained by using large learning rates within gradient descent algorithms.
no code implementations • NeurIPS 2021 • Gaspard Beugnot, Julien Mairal, Alessandro Rudi
The theory of spectral filtering is a remarkable tool to understand the statistical properties of learning with kernels.
no code implementations • NeurIPS 2021 • Gaspard Beugnot, Julien Mairal, Alessandro Rudi
The theory of spectral filtering is a remarkable tool to understand the statistical properties of learning with kernels.
no code implementations • 25 Feb 2021 • Gaspard Beugnot, Aude Genevay, Kristjan Greenewald, Justin Solomon
Optimal transport (OT) is a popular tool in machine learning to compare probability measures geometrically, but it comes with substantial computational burden.