Saved in:
Bibliographic Details
Main Authors: Lou, Hu, Gao, Yin-Jun, Zhang, Dong-Xiao, Du, Tai-Jiao, Zhang, Jun-Jie, Zhang, Jia-Rui
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2602.17952
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917283931095040
author Lou, Hu
Gao, Yin-Jun
Zhang, Dong-Xiao
Du, Tai-Jiao
Zhang, Jun-Jie
Zhang, Jia-Rui
author_facet Lou, Hu
Gao, Yin-Jun
Zhang, Dong-Xiao
Du, Tai-Jiao
Zhang, Jun-Jie
Zhang, Jia-Rui
contents One-dimensional function approximation is a fundamental problem in scientific computing and engineering applications. While neural networks possess powerful universal approximation capabilities, their optimization process is often hindered by flat loss landscapes induced by parameter-space symmetries, leading to slow convergence and poor generalization, particularly for high-frequency components. Inspired by the principle of \emph{symmetry breaking} in physics, this paper proposes a hardware-friendly approach for function approximation through \emph{input-space expansion}. The core idea involves augmenting the original one-dimensional input (e.g., $x$) with constant values (e.g., $π$) to form a higher-dimensional vector (e.g., $[π, π, x, π, π]$), effectively breaking parameter symmetries without increasing the network's parameter count. We evaluate the method on ten representative one-dimensional functions, including smooth, discontinuous, high-frequency, and non-differentiable functions. Experimental results demonstrate that input-space expansion significantly accelerates training convergence (reducing LBFGS iterations by 12\% on average) and enhances approximation accuracy (reducing final MSE by 66.3\% for the optimal 5D expansion). Ablation studies further reveal the effects of different expansion dimensions and constant selections, with $π$ consistently outperforming other constants. Our work proposes a low-cost, efficient, and hardware-friendly technique for algorithm design.
format Preprint
id arxiv_https___arxiv_org_abs_2602_17952
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Hardware-Friendly Input Expansion for Accelerating Function Approximation
Lou, Hu
Gao, Yin-Jun
Zhang, Dong-Xiao
Du, Tai-Jiao
Zhang, Jun-Jie
Zhang, Jia-Rui
Machine Learning
One-dimensional function approximation is a fundamental problem in scientific computing and engineering applications. While neural networks possess powerful universal approximation capabilities, their optimization process is often hindered by flat loss landscapes induced by parameter-space symmetries, leading to slow convergence and poor generalization, particularly for high-frequency components. Inspired by the principle of \emph{symmetry breaking} in physics, this paper proposes a hardware-friendly approach for function approximation through \emph{input-space expansion}. The core idea involves augmenting the original one-dimensional input (e.g., $x$) with constant values (e.g., $π$) to form a higher-dimensional vector (e.g., $[π, π, x, π, π]$), effectively breaking parameter symmetries without increasing the network's parameter count. We evaluate the method on ten representative one-dimensional functions, including smooth, discontinuous, high-frequency, and non-differentiable functions. Experimental results demonstrate that input-space expansion significantly accelerates training convergence (reducing LBFGS iterations by 12\% on average) and enhances approximation accuracy (reducing final MSE by 66.3\% for the optimal 5D expansion). Ablation studies further reveal the effects of different expansion dimensions and constant selections, with $π$ consistently outperforming other constants. Our work proposes a low-cost, efficient, and hardware-friendly technique for algorithm design.
title Hardware-Friendly Input Expansion for Accelerating Function Approximation
topic Machine Learning
url https://arxiv.org/abs/2602.17952