EPFL @ COLT 2022

charles-postiaux © 2022 COLT website

charles-postiaux © 2022 COLT website

6 EPFL papers were accepted to the 35th Annual Conference on Learning Theory (COLT).
COLT will take place in London, UK, July 2–5, 2022.
We take this opportunity to congratulate all the authors of the accepted EPFL papers.

List of the accepted EPFL papers:

Negative Curvature Obstructs Acceleration for Strongly Geodesically Convex Optimization, Even with Exact First-Order Oracles
Christopher Criscitiello; Nicolas Boumal

Trace Norm Regularization for Multi-Task Learning with Scarce Data
Etienne Boursier; Mikhail Konobeev; Nicolas Flammarion

Accelerated SGD for Non-Strongly-Convex Least Squares
Aditya Varre; Nicolas Flammarion

Label Noise (Stochastic) Gradient Descent Implicitly Solves the Lasso for Quadratic Parametrization
Loucas Vivien; Julien Reygner; Nicolas Flammarion

The Merged-Staircase Property: A Necessary and Nearly Sufficient Condition for SGD Learning of Sparse Functions on Two-Layer Neural Networks
Emmanuel Abbe; Enric Boix Adsera; Theodor Misiakiewicz

Approximate Cluster Recovery from Noisy Labels
Buddhima Gamlath; Silvio Lattanzi; Ashkan Norouzi-Fard; Ola Svensson