• Home
  • Search
  • Semi-Structured Sparsity Using Dynamic Percentile
  • https://doi.org/10.1109/icmisi65108.2025.11115469Copy DOI Icon

Semi-Structured Sparsity Using Dynamic Percentile

  • May 10, 2025
  • Omar Sabra +3 more
Show More
  • Abstract
  • Literature Map
  • References
  • Similar Papers
Abstract

The computational efficiency of deep neural networks is limited by hardware accelerators and sparsity integration. Leveraging sparsity can result in significant computational savings by reducing the number of operations and memory accesses required. In recent years, N:M sparsity, where N out of every M consecutive elements are zero, has emerged as a promising approach for optimizing neural networks. This semi-structured sparsity pattern facilitates high performance at substantial sparsity levels while enabling significant speedups on specialized hardware. However, existing methods often assume the minimum N values to be unimportant, regardless of their relative values. In this paper, we propose p:M, a novel framework that leverages dynamic percentiles in semi-structured sparsity, optimizing model accuracy for higher sparsities. Our approach reformulates the sparsity constraint, introducing dynamic percentiles for masking weights. This dynamic percentile process enables efficient pruning during training without additional computational overhead. Experimental results demonstrate that our method consistently improves upon traditional N:M sparsity techniques across various neural network architectures and datasets, achieving superior performance with reduced computational cost.

Similar Papers
  • Research Article

(Invited) Unconventional Computing with Memristive Devices and Arrays

  • Jul 23, 2018
  • Electrochemical Society Meeting Abstracts
  • Jun Yang
  • Conference Article
  • Citations65

Deep Learning for Fast Adaptive Beamforming

  • May 01, 2019
  • Ben Luijten +6
  • Conference Article
  • Citations8

INA: Incremental Network Approximation Algorithm for Limited Precision Deep Neural Networks

  • Nov 01, 2019
  • Zheyu Liu +5
  • Research Article
  • Citations1

Learning Sensor Sample-Reweighting for Dynamic Early-Exit Activity Recognition Via Meta Learning.

  • Jun 01, 2025
  • IEEE journal of biomedical and health informatics
  • Zenan Fu +5
  • Dissertation

Real-world applicable deep learning techniques

  • Jan 01, 2020
  • Ning Liu
  • Research Article

Hybrid Optimization for Computing-in-Memory Architectures: Addressing Device Variation and Stuck-At Faults with Progressive Quantization

  • Jan 01, 2025
  • IEEE Transactions on Components, Packaging and Manufacturing Technology
  • Rui Xiao +4
  • Research Article
  • Citations2

Обнаружение очагов возгорания на технологических объектах с использованием сверточной нейронной сети

  • Oct 25, 2023
  • Pozharovzryvobezopasnost/Fire and Explosion Safety
  • А А Евсиков +1
  • Research Article
  • Citations27

Tuning ANN Hyperparameters for Forecasting Drinking Water Demand

  • May 10, 2021
  • Applied Sciences
  • Andrea Menapace +2
  • Research Article

Comparative performance of backpropagation networks designed by genetic algorithms and heuristics

  • Jun 01, 1998
  • International Journal of Intelligent Systems in Accounting, Finance & Management
  • James V Hansen
  • Research Article

Comparing Application-Level Hardening Techniques for Neural Networks on GPUs

  • Mar 06, 2025
  • Electronics
  • Giuseppe Esposito +3
  • PDF
  • Research Article
  • Citations6

A novel approach to the development of neural network architecture based on metaheuristic protis approach

  • Aug 31, 2023
  • Eastern-European Journal of Enterprise Technologies
  • T Henny Febriana Harumy +3
  • Research Article

INTELLIGENT MODEL FOR CLASSIFYING HEMODYNAMIC PATTERNS OF BRAIN ACTIVATION TO IDENTIFY NEUROCOGNITIVE MECHANISMS OF SPATIAL-NUMERICAL ASSOCIATIONS

  • Jan 01, 2024
  • Vestnik komp'iuternykh i informatsionnykh tekhnologii
  • R G Asadullaev +1
  • Conference Article
  • Citations12

CondenseNeXt: An Ultra-Efficient Deep Neural Network for Embedded Systems

  • Jan 27, 2021
  • Priyank Kalgaonkar +1
  • Research Article
  • Citations126

Source localization using deep neural networks in a shallow water environment.

  • May 01, 2018
  • The Journal of the Acoustical Society of America
  • Zhaoqiong Huang +4
  • Research Article
  • Citations66

Multilevel image segmentation with adaptive image context based thresholding

  • Feb 01, 2010
  • Applied Soft Computing
  • Siddhartha Bhattacharyya +2
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.