Saved in:
Bibliographic Details
Main Authors: Tsipinakis, Nick, Tigkas, Panagiotis, Parpas, Panos
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2305.08742
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908865656782848
author Tsipinakis, Nick
Tigkas, Panagiotis
Parpas, Panos
author_facet Tsipinakis, Nick
Tigkas, Panagiotis
Parpas, Panos
contents Newton's method may exhibit slower convergence than vanilla Gradient Descent in its initial phase on strongly convex problems. Classical Newton-type multilevel methods mitigate this but, like Gradient Descent, achieve only linear convergence near the minimizer. We introduce an adaptive multilevel Newton-type method with a principled automatic switch to full Newton once its quadratic phase is reached. The local quadratic convergence for strongly convex functions with Lipschitz continuous Hessians and for self-concordant functions is established and confirmed empirically. Although per-iteration cost can exceed that of classical multilevel schemes, the method is efficient and consistently outperforms Newton's method, Gradient Descent, and the multilevel Newton method, indicating that second-order methods can outperform first-order methods even when Newton's method is initially slow. The promising empirical results open new avenues for designing reduced-cost second- and high-order methods with extremely fast convergence rates.
format Preprint
id arxiv_https___arxiv_org_abs_2305_08742
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Adaptive Multilevel Newton: A Quadratically Convergent Optimization Method
Tsipinakis, Nick
Tigkas, Panagiotis
Parpas, Panos
Optimization and Control
Newton's method may exhibit slower convergence than vanilla Gradient Descent in its initial phase on strongly convex problems. Classical Newton-type multilevel methods mitigate this but, like Gradient Descent, achieve only linear convergence near the minimizer. We introduce an adaptive multilevel Newton-type method with a principled automatic switch to full Newton once its quadratic phase is reached. The local quadratic convergence for strongly convex functions with Lipschitz continuous Hessians and for self-concordant functions is established and confirmed empirically. Although per-iteration cost can exceed that of classical multilevel schemes, the method is efficient and consistently outperforms Newton's method, Gradient Descent, and the multilevel Newton method, indicating that second-order methods can outperform first-order methods even when Newton's method is initially slow. The promising empirical results open new avenues for designing reduced-cost second- and high-order methods with extremely fast convergence rates.
title Adaptive Multilevel Newton: A Quadratically Convergent Optimization Method
topic Optimization and Control
url https://arxiv.org/abs/2305.08742