Filter Independence-Aware Pruning: Efficient Neural Networks for On-Device AI

Wang, Jiali; Bie, Hongxia; Jing, Zhao; Zhi, Yichen; Fan, Yongkai; Ma, Wentao

doi:10.3390/electronics15040794

Open AccessArticle

Filter Independence-Aware Pruning: Efficient Neural Networks for On-Device AI

by

Jiali Wang

¹

,

Hongxia Bie

^1,*

,

Zhao Jing

¹,

Yichen Zhi

¹

,

Yongkai Fan

² and

Wentao Ma

¹

Intelligent Media Computing Center, School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing 100876, China

²

School of Computer and Cyber Sciences, Communication University of China, Beijing 100024, China

^*

Author to whom correspondence should be addressed.

Electronics 2026, 15(4), 794; https://doi.org/10.3390/electronics15040794

Submission received: 6 January 2026 / Revised: 4 February 2026 / Accepted: 10 February 2026 / Published: 12 February 2026

(This article belongs to the Section Artificial Intelligence)

Download

Browse Figures

Versions Notes

Abstract

Filter pruning is an effective approach for improving the inference efficiency of neural networks and is particularly attractive for on-device artificial intelligence (AI) applications. However, many existing methods fail to accurately identify redundant filters due to limited modeling of inter-filter dependencies. A filter pruning method based on nuclear norm analysis is proposed to quantify filter independence and guide structured pruning. By analyzing the layer-wise distribution of independence scores, a principled trade-off between pruning rate and accuracy preservation is achieved. In most evaluation scenarios, the proposed method achieves 75–95% parameter reduction and 70–80% FLOPs reduction, while substantially higher compression ratios (up to 99%) can be obtained for more redundant network architectures, with consistent performance trends observed across multiple accuracy-related metrics. Furthermore, deployment on an RK3588 neural processing unit (NPU) demonstrates substantial reductions in memory consumption and inference latency, confirming the practical effectiveness of the method for mobile and edge AI applications.

Keywords: filter pruning; structured pruning; nuclear norm; filter independence; efficient neural networks; neural processing unit

Share and Cite

MDPI and ACS Style

Wang, J.; Bie, H.; Jing, Z.; Zhi, Y.; Fan, Y.; Ma, W. Filter Independence-Aware Pruning: Efficient Neural Networks for On-Device AI. Electronics 2026, 15, 794. https://doi.org/10.3390/electronics15040794

AMA Style

Wang J, Bie H, Jing Z, Zhi Y, Fan Y, Ma W. Filter Independence-Aware Pruning: Efficient Neural Networks for On-Device AI. Electronics. 2026; 15(4):794. https://doi.org/10.3390/electronics15040794

Chicago/Turabian Style

Wang, Jiali, Hongxia Bie, Zhao Jing, Yichen Zhi, Yongkai Fan, and Wentao Ma. 2026. "Filter Independence-Aware Pruning: Efficient Neural Networks for On-Device AI" Electronics 15, no. 4: 794. https://doi.org/10.3390/electronics15040794

APA Style

Wang, J., Bie, H., Jing, Z., Zhi, Y., Fan, Y., & Ma, W. (2026). Filter Independence-Aware Pruning: Efficient Neural Networks for On-Device AI. Electronics, 15(4), 794. https://doi.org/10.3390/electronics15040794

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Menu

Filter Independence-Aware Pruning: Efficient Neural Networks for On-Device AI

Abstract

Share and Cite

Article Metrics

Article Access Statistics

Further Information

Guidelines

MDPI Initiatives

Follow MDPI