Methods for pruning deep neural networks

Vadera, S; Ameen, SA

Abstract

This paper presents a survey of methods for pruning deep neural networks. It begins by
categorising over 150 studies based on the underlying approach used and then focuses on three categories:
methods that use magnitude based pruning, methods that utilise clustering to identify redundancy, and
methods that use sensitivity analysis to assess the effect of pruning. Some of the key influencing studies
within these categories are presented to highlight the underlying approaches and results achieved. Most
studies present results which are distributed in the literature as new architectures, algorithms and data
sets have developed with time, making comparison across different studied difficult. The paper therefore
provides a resource for the community that can be used to quickly compare the results from many different
methods on a variety of data sets, and a range of architectures, including AlexNet, ResNet, DenseNet and
VGG. The resource is illustrated by comparing the results published for pruning AlexNet and ResNet50 on
ImageNet and ResNet56 and VGG16 on the CIFAR10 data to reveal which pruning methods work well in
terms of retaining accuracy whilst achieving good compression rates. The paper concludes by identifying
some research gaps and promising directions for future research.

Citation

Vadera, S., & Ameen, S. (2022). Methods for pruning deep neural networks. IEEE Access, 63280- 63300. https://doi.org/10.1109/ACCESS.2022.3182659

• Journal Article Type: Article
• Acceptance Date: May 1, 2022
• Online Publication Date: Jun 13, 2022
• Publication Date: Jun 13, 2022
• Deposit Date: Jun 7, 2022
• Publicly Available Date: Aug 22, 2022
• Journal: IEEE Access
• Publisher: Institute of Electrical and Electronics Engineers
• Pages: 63280- 63300
• DOI: https://doi.org/10.1109/ACCESS.2022.3182659
• Publisher URL: http://dx.doi.org/10.1109/ACCESS.2022.3182659
• Related Public URLs: http://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=6287639