Quantitative Convergences of Lie Group Momentum Optimizers

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Kong, Lingkai, Tao, Molei
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866910464699531264
author Kong, Lingkai
Tao, Molei
author_facet Kong, Lingkai
Tao, Molei
contents Explicit, momentum-based dynamics that optimize functions defined on Lie groups can be constructed via variational optimization and momentum trivialization. Structure preserving time discretizations can then turn this dynamics into optimization algorithms. This article investigates two types of discretization, Lie Heavy-Ball, which is a known splitting scheme, and Lie NAG-SC, which is newly proposed. Their convergence rates are explicitly quantified under $L$-smoothness and local strong convexity assumptions. Lie NAG-SC provides acceleration over the momentumless case, i.e. Riemannian gradient descent, but Lie Heavy-Ball does not. When compared to existing accelerated optimizers for general manifolds, both Lie Heavy-Ball and Lie NAG-SC are computationally cheaper and easier to implement, thanks to their utilization of group structure. Only gradient oracle and exponential map are required, but not logarithm map or parallel transport which are computational costly.
format Preprint
id arxiv_https___arxiv_org_abs_2405_20390
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Quantitative Convergences of Lie Group Momentum Optimizers
Kong, Lingkai
Tao, Molei
Machine Learning
Numerical Analysis
Optimization and Control
Explicit, momentum-based dynamics that optimize functions defined on Lie groups can be constructed via variational optimization and momentum trivialization. Structure preserving time discretizations can then turn this dynamics into optimization algorithms. This article investigates two types of discretization, Lie Heavy-Ball, which is a known splitting scheme, and Lie NAG-SC, which is newly proposed. Their convergence rates are explicitly quantified under $L$-smoothness and local strong convexity assumptions. Lie NAG-SC provides acceleration over the momentumless case, i.e. Riemannian gradient descent, but Lie Heavy-Ball does not. When compared to existing accelerated optimizers for general manifolds, both Lie Heavy-Ball and Lie NAG-SC are computationally cheaper and easier to implement, thanks to their utilization of group structure. Only gradient oracle and exponential map are required, but not logarithm map or parallel transport which are computational costly.
title Quantitative Convergences of Lie Group Momentum Optimizers
topic Machine Learning
Numerical Analysis
Optimization and Control
url https://arxiv.org/abs/2405.20390