FGFP: A Fractional Gaussian Filter and Pruning for Deep Neural Networks Compression

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Tu, Kuan-Ting, Yu, Po-Hsien, Tseng, Yu-Syuan, Chien, Shao-Yi
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909712264462336
author Tu, Kuan-Ting
Yu, Po-Hsien
Tseng, Yu-Syuan
Chien, Shao-Yi
author_facet Tu, Kuan-Ting
Yu, Po-Hsien
Tseng, Yu-Syuan
Chien, Shao-Yi
contents Network compression techniques have become increasingly important in recent years because the loads of Deep Neural Networks (DNNs) are heavy for edge devices in real-world applications. While many methods compress neural network parameters, deploying these models on edge devices remains challenging. To address this, we propose the fractional Gaussian filter and pruning (FGFP) framework, which integrates fractional-order differential calculus and Gaussian function to construct fractional Gaussian filters (FGFs). To reduce the computational complexity of fractional-order differential operations, we introduce Grünwald-Letnikov fractional derivatives to approximate the fractional-order differential equation. The number of parameters for each kernel in FGF is minimized to only seven. Beyond the architecture of Fractional Gaussian Filters, our FGFP framework also incorporates Adaptive Unstructured Pruning (AUP) to achieve higher compression ratios. Experiments on various architectures and benchmarks show that our FGFP framework outperforms recent methods in accuracy and compression. On CIFAR-10, ResNet-20 achieves only a 1.52% drop in accuracy while reducing the model size by 85.2%. On ImageNet2012, ResNet-50 achieves only a 1.63% drop in accuracy while reducing the model size by 69.1%.
format Preprint
id arxiv_https___arxiv_org_abs_2507_22527
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle FGFP: A Fractional Gaussian Filter and Pruning for Deep Neural Networks Compression
Tu, Kuan-Ting
Yu, Po-Hsien
Tseng, Yu-Syuan
Chien, Shao-Yi
Machine Learning
Computer Vision and Pattern Recognition
Network compression techniques have become increasingly important in recent years because the loads of Deep Neural Networks (DNNs) are heavy for edge devices in real-world applications. While many methods compress neural network parameters, deploying these models on edge devices remains challenging. To address this, we propose the fractional Gaussian filter and pruning (FGFP) framework, which integrates fractional-order differential calculus and Gaussian function to construct fractional Gaussian filters (FGFs). To reduce the computational complexity of fractional-order differential operations, we introduce Grünwald-Letnikov fractional derivatives to approximate the fractional-order differential equation. The number of parameters for each kernel in FGF is minimized to only seven. Beyond the architecture of Fractional Gaussian Filters, our FGFP framework also incorporates Adaptive Unstructured Pruning (AUP) to achieve higher compression ratios. Experiments on various architectures and benchmarks show that our FGFP framework outperforms recent methods in accuracy and compression. On CIFAR-10, ResNet-20 achieves only a 1.52% drop in accuracy while reducing the model size by 85.2%. On ImageNet2012, ResNet-50 achieves only a 1.63% drop in accuracy while reducing the model size by 69.1%.
title FGFP: A Fractional Gaussian Filter and Pruning for Deep Neural Networks Compression
topic Machine Learning
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2507.22527