Optimization of Hyper-parameter for CNN Model using Genetic Algorithm

Yoo, Jaewon; Yoon, Hyun-il; Kim, Hyeong-Gyun; Yoon, Hee-Seung; Han, Seung-Soo

doi:10.1109/icecie47765.2019.8974762

Cited by 22 publications

(14 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…From Table 3, the selection of individuals in current generation as the parents for the next generation is based on stochastic uniform, which provides the highest accuracy compared with other search strategies defined in GA [31]. Furthermore, the type of crossover parameter, which classifies two parents of the intermediate generation to build a new individual for the next generation, is 'scattered' as it provides higher flexibility compared with the other types of the crossover functions [32]. It can also be seen from Table 3 that the type of mutation function is constrained dependent.…”

Section: Determining the Coefficients Values Using Gamentioning

confidence: 99%

Developing a Mathematical Model for Wind Power Plant Siting and Sizing in Distribution Networks

et al. 2020

View full text Add to dashboard Cite

Wind Power Plants (WPPs) are generally located in remote areas with weak distribution connections. Hence, the value of Short Circuit Capacity (SCC), WPP size and the short circuit impedance angle ratio (X/R) are all very critical in the voltage stability of a distribution system connected WPP. This paper presents a new voltage stability model based on the mathematical relations between voltage, the level of wind power penetration, SCC and X/R at a given Point of Common Coupling (PCC) of a distribution network connected WPP. The proposed model introduces six equations based on the SCC and X/R values seen from a particular PCC point. The equations were developed for two common types of Wind Turbine Generators (WTGs), including: the Induction Generator (IG) and the Double Fed Induction Generator (DFIG). Taking advantage of the proposed equations, design engineers can predict how the steady-state PCC voltage will behave in response to different penetrations of IG- and DFIG-based WPPs. In addition, the proposed equations enable computing the maximum size of the WPP, ensuring grid code requirements at the given PCC without the need to carry out complex and time-consuming computational tasks or modelling of the system, which is a significant advantage over existing WPP sizing approaches.

show abstract

Section: Determining the Coefficients Values Using Gamentioning

confidence: 99%

Developing a Mathematical Model for Wind Power Plant Siting and Sizing in Distribution Networks

et al. 2020

View full text Add to dashboard Cite

show abstract

“…With the use of Learning rate we can determine how much weights will be modified in subsequent training iterations (Yoo et al, 2019). Beta_1 and Beta_2 are hyperparameters used for first-and second-order moment estimation, respectively.…”

Section: Convolutional Neural Networkmentioning

confidence: 99%

Implementation of a Convolutional Neural Network for Eye Blink Artifacts Removal From the Electroencephalography Signal

2022

View full text Add to dashboard Cite

Electroencephalography (EEG) signals are disrupted by technical and physiological artifacts. One of the most common artifacts is the natural activity that results from the movement of the eyes and the blinking of the subject. Eye blink artifacts (EB) spread across the entire head surface and make EEG signal analysis difficult. Methods for the elimination of electrooculography (EOG) artifacts, such as independent component analysis (ICA) and regression, are known. The aim of this article was to implement the convolutional neural network (CNN) to eliminate eye blink artifacts. To train the CNN, a method for augmenting EEG signals was proposed. The results obtained from the CNN were compared with the results of the ICA and regression methods for the generated and real EEG signals. The results obtained indicate a much better performance of the CNN in the task of removing eye-blink artifacts, in particular for the electrodes located in the central part of the head.

show abstract

“…Encoding is the primary problem for the genetic algorithm to optimize neural network hyperparameters [27]. That is the process of expressing the hyperparameter combination of the neural network into the search space that the genetic algorithm could handle.…”

Section: Figure 6 the Process Of Genetic Algorithm Optimizationmentioning

confidence: 99%

Embryo Evaluation Based on ResNet with AdaptiveGA-optimized Hyperparameters

Zhou¹,

Zhou²,

Han³

et al. 2022

Journal of Internet Technology

View full text Add to dashboard Cite

<p>In vitro fertilization (IVF) embryo evaluation based on morphology is an effective method to improve the success rate of transplantation. Although convolutional neural networks (CNNs) have made great achievements in many image classifications, there are still great challenges in accurately classifying embryos due to the insufficient samples, interference of exfoliated cells, and inappropriate hyperparameter configuration in the classification network. In this paper, a residual neural network optimized by the adaptive genetic algorithm is proposed to evaluate embryos. Firstly, a novel algorithm for extracting the region of interest (ROI) is embedded in the preprocessing part of the model to eliminate exfoliated cells close to the embryo. Secondly, several kinds of specific transformation methods are established to expand the dataset based on the symmetry of embryos. In addition, an adaptive genetic algorithm is adopted to search for optimal hyperparameters. Experiments on the data set provided by Shanghai General Hospital show that the algorithm has an excellent performance in embryo evaluation. The accuracy of our model is 86.4%, the recall is 88.4%, and the AUC is 0.93. Our results indicated that the proposed model can effectively improve the classification performance of ResNet, and thus achieve the clinic requirements of embryo evaluation.</p> <p> </p>

show abstract

Optimization of Hyper-parameter for CNN Model using Genetic Algorithm

Cited by 22 publications

References 7 publications

Developing a Mathematical Model for Wind Power Plant Siting and Sizing in Distribution Networks

Developing a Mathematical Model for Wind Power Plant Siting and Sizing in Distribution Networks

Implementation of a Convolutional Neural Network for Eye Blink Artifacts Removal From the Electroencephalography Signal

Embryo Evaluation Based on ResNet with AdaptiveGA-optimized Hyperparameters

Contact Info

Product

Resources

About