Abstract
In this research work, a differential evolution-based method has been used to compress the deep neural network architectures. The compression is achieved by selecting the most dominant filters/nodes during the training of the model. The usefulness of filters is established by the test accuracy of the model. The experimental results demonstrate that the performance of proposed model compares fairly well with other state of art model compression techniques. Moreover, the compression achieved with VGG16 on MNIST, CIFAR-10 and CIFAR-100 datasets was 98.32, 98.5 and 93.54%, respectively. The corresponding compression achieved on ResNet50 was 85.24, 85.38 and 79.37%, while SqueezeNet which is already compressed model could also be compressed by 72.94, 73.77 and 44.59%, respectively. MobileNet, which is already a compact model developed for mobile applications, could also be compressed by 93.04, 93.74 and 76.37% on MNIST, CIFAR-10 and CIFAR-100 datasets. The loss in accuracy in compressed models turns out to be less than 2%. Further, the compressed models report acceleration in inference time being 80.79% on VGG16, 74.14% on ResNet50, 42.96% on MobileNet and 11.79% on SqueezeNet.
Similar content being viewed by others
Data Availability
The datasets generated during and/or analyzed during the current study can be requested from the corresponding author.
References
Kornblith S, Shlens J, Le QV (2019) Do better imagenet models transfer better? In Proceedings of the IEEE conference on computer vision and pattern recognition, pp 2661–2671
Li H, Ota K, Dong M (2018) Learning iot in edge: deep learning for the internet of things with edge computing. IEEE Netw 32(1):96–101
Wang X, Han Y, Leung VCM, Niyato D, Yan X, Chen X (2020) A comprehensive survey. IEEE Commun Surv Tutorials Converg Edge Comput Deep Learn
Li H, Kadav A, Durdanovic I, Samet H, Graf HP (2016) Pruning filters for efficient convnets. arXiv preprint arXiv:1608.08710
Hu Y, Sun S, Li J, Wang X, Gu Q (2018) A novel channel pruning method for deep neural network compression. arXiv preprint arXiv:1805.11394
Gong Y, Liu L, Yang M, Bourdev L (2014) Compressing deep convolutional networks using vector quantization. arXiv preprint arXiv:1412.6115
Li T, Wu B, Yang Y, Fan Y, Zhang Y, Liu W (2019) Compressing convolutional neural networks via factorized convolutional filters. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 3977–3986
Wen W, Wu C, Wang Y, Chen Y, Li H (2016) Learning structured sparsity in deep neural networks. Adv Neural Inf Process Syst 29
Cheng Y, Wang D, Zhou P, Zhang T (2017) A survey of model compression and acceleration for deep neural networks. arXiv preprint arXiv:1710.09282
Yang C, An Z, Li C, Diao B, Xu Y (2019) Multi-objective pruning for cnns using genetic algorithm. In: International conference on artificial neural networks, pp 299–305. Springer
Abotaleb AM, Elwakil AT, Hadhoud M (2019) Hybrid genetic based algorithm for cnn ultra compression. In 2019 31st International conference on microelectronics (ICM), pp 199–202. IEEE
Wang Z, Li F, Shi G, Xie X, Wang F (2020) Network pruning using sparse learning and genetic algorithm. Neurocomputing
Samala RK, Chan HP, Hadjiiski LM, Helvie MA, Richter C, Cha K (2018) Evolutionary pruning of transfer learned deep convolutional neural network for breast cancer diagnosis in digital breast tomosynthesis. Phys Med Biol 63(9):095005
Zhou Y, Yen GG, Yi Z (2019) A knee-guided evolutionary algorithm for compressing deep neural networks. IEEE Trans Cybernet 51(3):1626–1638
Fernandes Jr FE, Yen GG (2021) Pruning deep convolutional neural networks architectures with evolution strategy. Inf Sci 552:29–47
Karaboga N, Cetinkaya B (2005) Performance comparison of genetic and differential evolution algorithms for digital fir filter design. In: Advances in information systems: third international conference, ADVIS 2004, Izmir, Turkey, 20--22 Oct2004. Proceedings 3, pp 482–488. Springer
Das S, Abraham A, Konar A (2008) Particle swarm optimization and differential evolution algorithms: technical analysis, applications and hybridization perspectives. In: Advances of computational intelligence in industrial systems, pp 1–38. Springer
Babu BV, Jehan MML (2003) Differential evolution for multi-objective optimization. In: The 2003 congress on evolutionary computation, 2003. CEC’03., vol 4, pp 2696–2703. IEEE
Paterlini S, Krink T (2006) Differential evolution and particle swarm optimisation in partitional clustering. Comput Stat Data Anal 50(5):1220–1247
Quang BF, Perov VL (1993) New evolutionary genetic algorithms for np-complete combinatorial optimization problems. Biol Cybernet 69(3):229–234
Panchal G, Panchal D (2015) Solving np hard problems using genetic algorithm. Transportation 106:6–2
LeCun Y, Denker J, Solla S (1989) Optimal brain damage. Adv Neural Inf Process Syst 2
Han S, Pool J, Tran J, Dally W (2015) Learning both weights and connections for efficient neural network. Adv Neural Inf Process Syst 28
Han S, Mao H, Dally WJ (2015) Deep compression: compressing deep neural networks with pruning, trained quantization and huffman coding. arXiv preprint arXiv:1510.00149
Luo JH, Wu J, Lin W (2017) Thinet: a filter level pruning method for deep neural network compression. In Proceedings of the IEEE international conference on computer vision, pp 5058–5066
He Y, Lin J, Liu Z, Wang H, Li LJ, Han S (2018) Amc: automl for model compression and acceleration on mobile devices. In: Proceedings of the European conference on computer vision (ECCV), pp 784–800
Krizhevsky A, Sutskever I, Hinton G (2012) Imagenet classification with deep convolutional neural networks. In: Advances in neural information processing systems, pp 1097–1105
Mittal D, Bhardwaj S, Khapra MM, Ravindran B (2019) Studying the plasticity in deep convolutional neural networks using random pruning. Mach Vis Appl 30(2):203–216
Das S, Abraham A, Konar A (2007) Automatic clustering using an improved differential evolution algorithm. IEEE Trans Syst Man Cybernet Part A Syst Humans 38(1):218–237
Deb, Roy JS, Gupta B (2014) Performance comparison of differential evolution, particle swarm optimization and genetic algorithm in the design of circularly polarized microstrip antennas. IEEE Trans Antennas Propag 62(8):3920–3928
Ramadas M, Abraham A, Kumar S (2019) FSDE-forced strategy differential evolution used for data clustering. J King Saud Univ Comput Inf Sci 31(1):52–61
Storn R, Price K (1997) Differential evolution-a simple and efficient heuristic for global optimization over continuous spaces. J Glob Opt 11(4):341–359
Simonyan K, Zisserman A (2014) Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556
He K, Zhang X, Ren S, Sun J (2016) Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp 770–778
Landola FN, Han S, Moskewicz MW, Ashraf K, Dally WJ, Keutzer K (2016) Squeezenet: alexnet-level accuracy with 50x fewer parameters and< 0.5 mb model size. arXiv preprint arXiv:1602.07360
Hughes D, Salathé M et al. (2015) An open access repository of images on plant health to enable the development of mobile disease diagnostics. arXiv preprint arXiv:1511.08060
Liu Z, Mu H, Zhang X, Guo Z, Yang X, Cheng KT, Sun J (2019) Metapruning: Meta learning for automatic neural network channel pruning. In Proceedings of the IEEE/CVF international conference on computer vision, pp 3296–3305
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Agarwal, M., Gupta, S.K. & Biswas, K.K. DECACNN: differential evolution-based approach to compress and accelerate the convolution neural network model. Neural Comput & Applic 36, 2665–2681 (2024). https://doi.org/10.1007/s00521-023-09166-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00521-023-09166-9