On the convergence of the gradient descent method with stochastic fixed-point rounding errors under the Polyak–Łojasiewicz inequality

Lu Xia (), Stefano Massei () and Michiel E. Hochstenbach ()
Additional contact information
Lu Xia: Eindhoven University of Technology
Stefano Massei: Università di Pisa
Michiel E. Hochstenbach: Eindhoven University of Technology

Computational Optimization and Applications, 2025, vol. 90, issue 3, No 6, 753-799

Abstract: Abstract In the training of neural networks with low-precision computation and fixed-point arithmetic, rounding errors often cause stagnation or are detrimental to the convergence of the optimizers. This study provides insights into the choice of appropriate stochastic rounding strategies to mitigate the adverse impact of roundoff errors on the convergence of the gradient descent method, for problems satisfying the Polyak–Łojasiewicz inequality. Within this context, we show that a biased stochastic rounding strategy may be even beneficial in so far as it eliminates the vanishing gradient problem and forces the expected roundoff error in a descent direction. Furthermore, we obtain a bound on the convergence rate that is stricter than the one achieved by unbiased stochastic rounding. The theoretical analysis is validated by comparing the performances of various rounding strategies when optimizing several examples using low-precision fixed-point arithmetic.

Keywords: Fixed-point arithmetic; Rounding error analysis; Gradient descent; Low-precision; Stochastic rounding; Polyak–Łojasiewicz inequality; 97N20; 62J02; 65G30; 65G50; 68T01 (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
http://link.springer.com/10.1007/s10589-025-00656-1 Abstract (text/html)
Access to the full text of the articles in this series is restricted.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:coopap:v:90:y:2025:i:3:d:10.1007_s10589-025-00656-1

Ordering information: This journal article can be ordered from
http://www.springer.com/math/journal/10589

DOI: 10.1007/s10589-025-00656-1

Access Statistics for this article

Computational Optimization and Applications is currently edited by William W. Hager

More articles in Computational Optimization and Applications from Springer
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().