Optimal First-Order Algorithms as a Function of Inequalities
From MaRDI portal
Publication:6504730
arXiv2110.11035MaRDI QIDQ6504730FDOQ6504730
Authors: Chan-Woo Park, Ernest K. Ryu 
Abstract: In this work, we present a novel algorithm design methodology that finds the optimal algorithm as a function of inequalities. Specifically, we restrict convergence analyses of algorithms to use a prespecified subset of inequalities, rather than utilizing all true inequalities, and find the optimal algorithm subject to this restriction. This methodology allows us to design algorithms with certain desired characteristics. As concrete demonstrations of this methodology, we find new state-of-the-art accelerated first-order gradient methods using randomized coordinate updates and backtracking line searches.
Has companion code repository: https://github.com/chanwoo-park-official/a-star-map
This page was built for publication: Optimal First-Order Algorithms as a Function of Inequalities
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6504730)
