Gespeichert in:
Bibliographische Detailangaben
1. Verfasser: Tyurin, Alexander
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2412.11773
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866915360540721152
author Tyurin, Alexander
author_facet Tyurin, Alexander
contents We study the classical optimization problem $\min_{x \in \mathbb{R}^d} f(x)$ and analyze the gradient descent (GD) method in both nonconvex and convex settings. It is well-known that, under the $L$-smoothness assumption ($\|\nabla^2 f(x)\| \leq L$), the optimal point minimizing the quadratic upper bound $f(x_k) + \langle\nabla f(x_k), x_{k+1} - x_k\rangle + \frac{L}{2} \|x_{k+1} - x_k\|^2$ is $x_{k+1} = x_k - γ_k \nabla f(x_k)$ with step size $γ_k = \frac{1}{L}$. Surprisingly, a similar result can be derived under the $\ell$-generalized smoothness assumption ($\|\nabla^2 f(x)\| \leq \ell(\|\nabla f(x)\|)$). In this case, we derive the step size $$γ_k = \int_{0}^{1} \frac{d v}{\ell(\|\nabla f(x_k)\| + \|\nabla f(x_k)\| v)}.$$ Using this step size rule, we improve upon existing theoretical convergence rates and obtain new results in several previously unexplored setups.
format Preprint
id arxiv_https___arxiv_org_abs_2412_11773
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Toward a Unified Theory of Gradient Descent under Generalized Smoothness
Tyurin, Alexander
Optimization and Control
We study the classical optimization problem $\min_{x \in \mathbb{R}^d} f(x)$ and analyze the gradient descent (GD) method in both nonconvex and convex settings. It is well-known that, under the $L$-smoothness assumption ($\|\nabla^2 f(x)\| \leq L$), the optimal point minimizing the quadratic upper bound $f(x_k) + \langle\nabla f(x_k), x_{k+1} - x_k\rangle + \frac{L}{2} \|x_{k+1} - x_k\|^2$ is $x_{k+1} = x_k - γ_k \nabla f(x_k)$ with step size $γ_k = \frac{1}{L}$. Surprisingly, a similar result can be derived under the $\ell$-generalized smoothness assumption ($\|\nabla^2 f(x)\| \leq \ell(\|\nabla f(x)\|)$). In this case, we derive the step size $$γ_k = \int_{0}^{1} \frac{d v}{\ell(\|\nabla f(x_k)\| + \|\nabla f(x_k)\| v)}.$$ Using this step size rule, we improve upon existing theoretical convergence rates and obtain new results in several previously unexplored setups.
title Toward a Unified Theory of Gradient Descent under Generalized Smoothness
topic Optimization and Control
url https://arxiv.org/abs/2412.11773