Saved in:
Bibliographic Details
Main Authors: Szente, Teodor Alexandru, Harrison, James, Zanfir, Mihai, Sminchisescu, Cristian
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2411.14855
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916492177571840
author Szente, Teodor Alexandru
Harrison, James
Zanfir, Mihai
Sminchisescu, Cristian
author_facet Szente, Teodor Alexandru
Harrison, James
Zanfir, Mihai
Sminchisescu, Cristian
contents Fractional gradient descent has been studied extensively, with a focus on its ability to extend traditional gradient descent methods by incorporating fractional-order derivatives. This approach allows for more flexibility in navigating complex optimization landscapes and offers advantages in certain types of problems, particularly those involving non-linearities and chaotic dynamics. Yet, the challenge of fine-tuning the fractional order parameters remains unsolved. In this work, we demonstrate that it is possible to train a neural network to predict the order of the gradient effectively.
format Preprint
id arxiv_https___arxiv_org_abs_2411_14855
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Applications of fractional calculus in learned optimization
Szente, Teodor Alexandru
Harrison, James
Zanfir, Mihai
Sminchisescu, Cristian
Machine Learning
Fractional gradient descent has been studied extensively, with a focus on its ability to extend traditional gradient descent methods by incorporating fractional-order derivatives. This approach allows for more flexibility in navigating complex optimization landscapes and offers advantages in certain types of problems, particularly those involving non-linearities and chaotic dynamics. Yet, the challenge of fine-tuning the fractional order parameters remains unsolved. In this work, we demonstrate that it is possible to train a neural network to predict the order of the gradient effectively.
title Applications of fractional calculus in learned optimization
topic Machine Learning
url https://arxiv.org/abs/2411.14855