Non-Euclidean High-Order Smooth Convex Optimization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Contreras, Juan Pablo, Guzmán, Cristóbal, Martínez-Rubio, David
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916599823335424
author Contreras, Juan Pablo
Guzmán, Cristóbal
Martínez-Rubio, David
author_facet Contreras, Juan Pablo
Guzmán, Cristóbal
Martínez-Rubio, David
contents We develop algorithms for the optimization of convex objectives that have Hölder continuous $q$-th derivatives by using a $q$-th order oracle, for any $q \geq 1$. Our algorithms work for general norms under mild conditions, including the $\ell_p$-settings for $1\leq p\leq \infty$. We can also optimize structured functions that allow for inexactly implementing a non-Euclidean ball optimization oracle. We do this by developing a non-Euclidean inexact accelerated proximal point method that makes use of an \emph{inexact uniformly convex regularizer}. We show a lower bound for general norms that demonstrates our algorithms are nearly optimal in high-dimensions in the black-box oracle model for $\ell_p$-settings and all $q \geq 1$, even in randomized and parallel settings. This new lower bound, when applied to the first-order smooth case, resolves an open question in parallel convex optimization.
format Preprint
id arxiv_https___arxiv_org_abs_2411_08987
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Non-Euclidean High-Order Smooth Convex Optimization
Contreras, Juan Pablo
Guzmán, Cristóbal
Martínez-Rubio, David
Optimization and Control
Data Structures and Algorithms
Machine Learning
We develop algorithms for the optimization of convex objectives that have Hölder continuous $q$-th derivatives by using a $q$-th order oracle, for any $q \geq 1$. Our algorithms work for general norms under mild conditions, including the $\ell_p$-settings for $1\leq p\leq \infty$. We can also optimize structured functions that allow for inexactly implementing a non-Euclidean ball optimization oracle. We do this by developing a non-Euclidean inexact accelerated proximal point method that makes use of an \emph{inexact uniformly convex regularizer}. We show a lower bound for general norms that demonstrates our algorithms are nearly optimal in high-dimensions in the black-box oracle model for $\ell_p$-settings and all $q \geq 1$, even in randomized and parallel settings. This new lower bound, when applied to the first-order smooth case, resolves an open question in parallel convex optimization.
title Non-Euclidean High-Order Smooth Convex Optimization
topic Optimization and Control
Data Structures and Algorithms
Machine Learning
url https://arxiv.org/abs/2411.08987