In accordance with regulations and requirements, the editorial department's website domain has been changed to arocmag.cn. The original domain (arocmag.com) has been redirecting to new domain since Jan. 1st, 2025.
See More
App. Res. of Comp. Accepted Paper
Vol. 43, 2026 No. 2

Hybrid compression method via knowledge distillation and meta-gradient initialization-based pruning

Zhang Hongmei1,2
Pan Shoudeng1,2
Liu Kejia3
Huo Junjie1,2
Du Ruiyang1,2
Wei Hongyi1,2
1. School of Information and Communication, Guilin University of Electronic Technology, Guilin Guangxi 541010, China
2. Key Laboratory of Cognitive Radio and Information Processing, Ministry of Education, School of Information and Communication, Guilin University of Electronic Technology, Guilin Guangxi 541010, China
3. Guangxi Highest Good Development Co. , Ltd. , Nanning 530029, China
DOI: 10.19734/j.issn.1001-3695.2025.07.0246

Abstract

To address the issues of representation degradation and accuracy decline in deep neural networks during pruning at initialization, this paper proposes a Dual-Stage Homogeneous Distillation (DHD) framework for efficient sparse modeling. The approach consists of two stages: Homogeneous Self-Distillation (HSD) and Sparsity-Aware Temperature Coupling (SATC) . In the HSD stage, a parameter mirroring initialization strategy constructs a topologically symmetric student network, which enhances parameter smoothness and model performance through self-distillation optimization. The SATC stage dynamically adjusts the distillation temperature for the correct and incorrect classes in the teacher model, enabling adaptive alignment between knowledge transfer intensity and pruning ratio. Experiments on CIFAR-10, Tiny-ImageNet, and ImageNet demonstrate that the proposed method maintains high accuracy even at a pruning rate of 91.4%, significantly outperforming methods such as SNIP and GraSP. The DHD framework effectively mitigates representation degradation under high pruning ratios, improves the performance and robustness of sparse sub-networks, and offers a feasible model compression solution for deployment on edge devices.

Foundation Support

广西区重点研发计划项目:基于人工智能供水运营模型的智慧水务数据研究及系统开发应用示范(桂科AB23075114)
广西区自然科学基金重点项目:边缘环境下基于深度压缩的恶意软件检测(2020GXNSFDA238001)

Publish Information

DOI: 10.19734/j.issn.1001-3695.2025.07.0246
Publish at: Application Research of Computers Accepted Paper, Vol. 43, 2026 No. 2

Publish History

[2025-10-28] Accepted Paper

Cite This Article

张红梅, 潘守登, 刘坷嘉, 等. 基于知识蒸馏及元梯度初始化剪枝的混合压缩方法 [J]. 计算机应用研究, 2026, 43 (2). (2025-11-04). https://doi.org/10.19734/j.issn.1001-3695.2025.07.0246. (Zhang Hongmei, Pan Shoudeng, Liu Kejia, et al. Hybrid compression method via knowledge distillation and meta-gradient initialization-based pruning [J]. Application Research of Computers, 2026, 43 (2). (2025-11-04). https://doi.org/10.19734/j.issn.1001-3695.2025.07.0246. )

About the Journal

  • Application Research of Computers Monthly Journal
  • Journal ID ISSN 1001-3695
    CN  51-1196/TP

Application Research of Computers, founded in 1984, is an academic journal of computing technology sponsored by Sichuan Institute of Computer Sciences under the Science and Technology Department of Sichuan Province.

Aiming at the urgently needed cutting-edge technology in this discipline, Application Research of Computers reflects the mainstream technology, hot technology and the latest development trend of computer application research at home and abroad in a timely manner. The main contents of the journal include high-level academic papers in this discipline, the latest scientific research results and major application results. The contents of the columns involve new theories of computer discipline, basic computer theory, algorithm theory research, algorithm design and analysis, blockchain technology, system software and software engineering technology, pattern recognition and artificial intelligence, architecture, advanced computing, parallel processing, database technology, computer network and communication technology, information security technology, computer image graphics and its latest hot application technology.

Application Research of Computers has many high-level readers and authors, and its readers are mainly senior and middle-level researchers and engineers engaged in the field of computer science, as well as teachers and students majoring in computer science and related majors in colleges and universities. Over the years, the total citation frequency and Web download rate of Application Research of Computers have been ranked among the top of similar academic journals in this discipline, and the academic papers published are highly popular among the readers for their novelty, academics, foresight, orientation and practicality.

Indexed & Evaluation

  • The Second National Periodical Award 100 Key Journals
  • Double Effect Journal of China Journal Formation
  • the Core Journal of China (Peking University 2023 Edition)
  • the Core Journal for Science
  • Chinese Science Citation Database (CSCD) Source Journals
  • RCCSE Chinese Core Academic Journals
  • Journal of China Computer Federation
  • 2020-2022 The World Journal Clout Index (WJCI) Report of Scientific and Technological Periodicals
  • Full-text Source Journal of China Science and Technology Periodicals Database
  • Source Journal of China Academic Journals Comprehensive Evaluation Database
  • Source Journals of China Academic Journals (CD-ROM Version), China Journal Network
  • 2017-2019 China Outstanding Academic Journals with International Influence (Natural Science and Engineering Technology)
  • Source Journal of Top Academic Papers (F5000) Program of China's Excellent Science and Technology Journals
  • Source Journal of China Engineering Technology Electronic Information Network and Electronic Technology Literature Database
  • Source Journal of British Science Digest (INSPEC)
  • Japan Science and Technology Agency (JST) Source Journal
  • Russian Journal of Abstracts (AJ, VINITI) Source Journals
  • Full-text Journal of EBSCO, USA
  • Cambridge Scientific Abstracts (Natural Sciences) (CSA(NS)) core journals
  • Poland Copernicus Index (IC)
  • Ulrichsweb (USA)