Computing Optimal Regularizers for Online Linear Optimization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Gatmiry, Khashayar, Schneider, Jon, Jegelka, Stefanie
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913560073863168
author Gatmiry, Khashayar
Schneider, Jon
Jegelka, Stefanie
author_facet Gatmiry, Khashayar
Schneider, Jon
Jegelka, Stefanie
contents Follow-the-Regularized-Leader (FTRL) algorithms are a popular class of learning algorithms for online linear optimization (OLO) that guarantee sub-linear regret, but the choice of regularizer can significantly impact dimension-dependent factors in the regret bound. We present an algorithm that takes as input convex and symmetric action sets and loss sets for a specific OLO instance, and outputs a regularizer such that running FTRL with this regularizer guarantees regret within a universal constant factor of the best possible regret bound. In particular, for any choice of (convex, symmetric) action set and loss set we prove that there exists an instantiation of FTRL which achieves regret within a constant factor of the best possible learning algorithm, strengthening the universality result of Srebro et al., 2011. Our algorithm requires preprocessing time and space exponential in the dimension $d$ of the OLO instance, but can be run efficiently online assuming a membership and linear optimization oracle for the action and loss sets, respectively (and is fully polynomial time for the case of constant dimension $d$). We complement this with a lower bound showing that even deciding whether a given regularizer is $α$-strongly-convex with respect to a given norm is NP-hard.
format Preprint
id arxiv_https___arxiv_org_abs_2410_17336
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Computing Optimal Regularizers for Online Linear Optimization
Gatmiry, Khashayar
Schneider, Jon
Jegelka, Stefanie
Machine Learning
Data Structures and Algorithms
Computer Science and Game Theory
Statistics Theory
Follow-the-Regularized-Leader (FTRL) algorithms are a popular class of learning algorithms for online linear optimization (OLO) that guarantee sub-linear regret, but the choice of regularizer can significantly impact dimension-dependent factors in the regret bound. We present an algorithm that takes as input convex and symmetric action sets and loss sets for a specific OLO instance, and outputs a regularizer such that running FTRL with this regularizer guarantees regret within a universal constant factor of the best possible regret bound. In particular, for any choice of (convex, symmetric) action set and loss set we prove that there exists an instantiation of FTRL which achieves regret within a constant factor of the best possible learning algorithm, strengthening the universality result of Srebro et al., 2011. Our algorithm requires preprocessing time and space exponential in the dimension $d$ of the OLO instance, but can be run efficiently online assuming a membership and linear optimization oracle for the action and loss sets, respectively (and is fully polynomial time for the case of constant dimension $d$). We complement this with a lower bound showing that even deciding whether a given regularizer is $α$-strongly-convex with respect to a given norm is NP-hard.
title Computing Optimal Regularizers for Online Linear Optimization
topic Machine Learning
Data Structures and Algorithms
Computer Science and Game Theory
Statistics Theory
url https://arxiv.org/abs/2410.17336