Entropy Error-based Convolutional Neural Network Sparse Optimization and Application
DOI:
https://doi.org/10.54097/rxbm7088Keywords:
Convolution Neural Network (CNN), Smooth L1 Regularization, Cross-Entropy Loss Function, Classification ProblemsAbstract
Convolutional Neural Networks (CNNs), as quintessential representatives of deep neural networks, have found widespread applications in numerous domains such as image recognition, object detection, and image generation, owing to their robust nonlinear mapping capabilities. However, the practical deployment of CNNs faces challenges such as low learning efficiency and subpar accuracy due to the intricacies in the network structure. This paper introduces the smoothing L1 regularization term atop the conventional entropy loss function, effectively enhancing both the learning efficiency and algorithmic precision of Convolutional Neural Networks. The efficacy of the improved algorithm is substantiated through numerical simulations.
Downloads
References
Leon. CNN: A Vision of Complexity, International Journal of Bifurcation and Chaos, 7 (1997) :2219-2425.
B. T. Li. Learning deep neural networks for node classification, Expert systems with applications, 137(2019): 324-334.
F. Ghasemi. Deep neural network in qsar studies using deep belief network, Applied Softcomputing, 62 (2018): 251-258.
Y. S. Xu. Convergence of deep convolutional neural networks, Neural Networks, 153 (2022): 553-563.
S. F. Wen, Translation analysis of English address image recognition based on image recognition, image and video processing, 11 (2019): 1-9.
M.Y.Park, T.Hastie, L1-Regularization Path Algorithm for Generalized Linear Model, Journal of the Royal Statistical Society Series B: Statistical Methodology, 69 (2007): 659–677.
C.Yang. Structured pruning of convolutional neural networks via L1 Regularization, IEEE Access, 7 (2019): 106385-106394.
L. Li. Approximating the Gradient of Cross-Entropy Loss Function, IEEE Access, 2020 (8): 111626-111635.
T.Wiatowski.A mathematical theory of deep convolutional neural networks for feature extraction, IEEE Transactions On Information Theory, 64(2017): 1845-1866.
P.Y.Wang. Differentially private sgd with non-smooth losses, Applied and Computational Harmonic Analysis, 56 (2022): 306-336.
Downloads
Published
Issue
Section
License
Copyright (c) 2024 Frontiers in Computing and Intelligent Systems
This work is licensed under a Creative Commons Attribution 4.0 International License.