<abstract xmlns="http://www.w3.org/1999/xhtml">

<p>In this paper, genetic algorithm (GA) is used to optimally determine the architecture of a convolutional neural network (CNN) that is used to classify handwritten numbers. The CNN is a class of deep feed-forward network, which have seen major success in the field of visual image analysis. During training, a good CNN architecture is capable of extracting complex features from the given training data; however, at present, there is no standard way to determine the architecture of a CNN. Domain knowledge and human expertise are required in order to design a CNN architecture. Typically architectures, The GA determine the exact architecture of a CNN by evolving the various hyper parameters of the architecture for a given application. The proposed method was tested on the MNIST dataset. The results show that the genetic algorithm is capable of generating successful CNN architectures. The proposed method performs the entire process of architecture generation without any human intervention.</p>
</abstract>

In this paper, genetic algorithm (GA) is used to optimally determine the architecture of a convolutional neural network (CNN) that is used to classify handwritten numbers. The CNN is a class of deep feed-forward network, which have seen major success in the field of visual image analysis. During training, a good CNN architecture is capable of extracting complex features from the given training data; however, at present, there is no standard way to determine the architecture of a CNN. Domain knowledge and human expertise are required in order to design a CNN architecture. Typically architectures, The GA determine the exact architecture of a CNN by evolving the various hyper parameters of the architecture for a given application. The proposed method was tested on the MNIST dataset. The results show that the genetic algorithm is capable of generating successful CNN architectures. The proposed method performs the entire process of architecture generation without any human intervention.

Designing Convolutional Neural Network Architecture Using Genetic Algorithms

International Journal of Advanced Network, Monitoring and Controls

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

{"article-title":"Designing Convolutional Neural Network Architecture Using Genetic Algorithms"}

In this paper, genetic algorithm (GA) is used to optimally determine the architecture of a convolutional neural network (CNN) that is used to classify...

Parameters	Value
Tournament selection size	2
Crossover Probability	50%
Mutation probability	80%
Genes Mutated	10%

Exp. No.	Highest Fitness Value
1	0.984499992943
2	0.973899998105
3	0.988800008184
4	0.991900001359
5	0.947799991965
6	0.949000005102
7	0.983099997652
8	0.979799999475
9	0.956399999567
10	0.972350000068

Hyper parameter	Range
No. of Epoch	(0 – 127)
Batch Size	(0 – 256)
No. of Convolution Layers	(0 – 8)
No. of Filters at each Convo layer	(0 – 64)
Convo Filter Size at each Convo layer	(0 – 8)
Activations used at each Convo layer	(sigmoid, tanh, relu, linear)
Maxpool layer after each Convo layer	(true, false)
Maxpool Pool Size for each Maxpool layer	(0 – 8)
No. of Feed-Forward Hidden Layers	(0 – 8)
No. of Feed-Forward Hidden Neurons at each layer	(0 – 64)
Activations used at each Feed-Forward layer	(sigmoid, tanh, softmax, relu)
Optimizer	(Adagrad, Adadelta, RMS, SGD)

Designing Convolutional Neural Network Architecture Using Genetic Algorithms

Data publikacji: 22 lut 2021

Zakres stron: 26 - 35

DOI: https://doi.org/10.21307/ijanmc-2021-024

Słowa kluczowe
Convolutional Neural Network, Genetic Algorithm, MNIST Data

© 2021 Ashray Bhandare et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

Figure 7.

Figure 8.

Parameters of the genetic operations

Highest fitness values obtained during each of the 10 experiments

The various hyper parameters in CNN with their ranges

Designing Convolutional Neural Network Architecture Using Genetic Algorithms

Data publikacji: 22 lut 2021

Zakres stron: 26 - 35

DOI: https://doi.org/10.21307/ijanmc-2021-024

Słowa kluczoweConvolutional Neural Network, Genetic Algorithm, MNIST Data

© 2021 Ashray Bhandare et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

Figure 7.

Figure 8.

Parameters of the genetic operations

Highest fitness values obtained during each of the 10 experiments

The various hyper parameters in CNN with their ranges

Słowa kluczowe
Convolutional Neural Network, Genetic Algorithm, MNIST Data