More Optimal Fractional-Order Stochastic Gradient Descent for Non-Convex Optimization Problems

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Partohaghighi, Mohammad, Marcia, Roummel, Chen, YangQuan
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866913822346838016
author Partohaghighi, Mohammad
Marcia, Roummel
Chen, YangQuan
author_facet Partohaghighi, Mohammad
Marcia, Roummel
Chen, YangQuan
contents Fractional-order stochastic gradient descent (FOSGD) leverages fractional exponents to capture long-memory effects in optimization. However, its utility is often limited by the difficulty of tuning and stabilizing these exponents. We propose 2SED Fractional-Order Stochastic Gradient Descent (2SEDFOSGD), which integrates the Two-Scale Effective Dimension (2SED) algorithm with FOSGD to adapt the fractional exponent in a data-driven manner. By tracking model sensitivity and effective dimensionality, 2SEDFOSGD dynamically modulates the exponent to mitigate oscillations and hasten convergence. Theoretically, for onoconvex optimization problems, this approach preserves the advantages of fractional memory without the sluggish or unstable behavior observed in naïve fractional SGD. Empirical evaluations in Gaussian and $α$-stable noise scenarios using an autoregressive (AR) model highlight faster convergence and more robust parameter estimates compared to baseline methods, underscoring the potential of dimension-aware fractional techniques for advanced modeling and estimation tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2505_02985
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle More Optimal Fractional-Order Stochastic Gradient Descent for Non-Convex Optimization Problems
Partohaghighi, Mohammad
Marcia, Roummel
Chen, YangQuan
Machine Learning
Optimization and Control
Fractional-order stochastic gradient descent (FOSGD) leverages fractional exponents to capture long-memory effects in optimization. However, its utility is often limited by the difficulty of tuning and stabilizing these exponents. We propose 2SED Fractional-Order Stochastic Gradient Descent (2SEDFOSGD), which integrates the Two-Scale Effective Dimension (2SED) algorithm with FOSGD to adapt the fractional exponent in a data-driven manner. By tracking model sensitivity and effective dimensionality, 2SEDFOSGD dynamically modulates the exponent to mitigate oscillations and hasten convergence. Theoretically, for onoconvex optimization problems, this approach preserves the advantages of fractional memory without the sluggish or unstable behavior observed in naïve fractional SGD. Empirical evaluations in Gaussian and $α$-stable noise scenarios using an autoregressive (AR) model highlight faster convergence and more robust parameter estimates compared to baseline methods, underscoring the potential of dimension-aware fractional techniques for advanced modeling and estimation tasks.
title More Optimal Fractional-Order Stochastic Gradient Descent for Non-Convex Optimization Problems
topic Machine Learning
Optimization and Control
url https://arxiv.org/abs/2505.02985