• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.22 no.6
• /
• pp.75-81
• /
• 2022

Autonomous vehicles, which can dramatically solve the lack of parking spaces, are making great progress through deep reinforcement learning. Activation functions are used for deep reinforcement learning, and various activation functions have been proposed, but their performance deviations were large depending on the application environment. Therefore, finding the optimal activation function depending on the environment is important for effective learning. This paper analyzes 12 functions mainly used in reinforcement learning to compare and evaluate which activation function is most effective when autonomous vehicles use deep reinforcement learning to learn parking. To this end, a performance evaluation environment was established, and the average reward of each activation function was compared with the success rate, episode length, and vehicle speed. As a result, the highest reward was the case of using GELU, and the ELU was the lowest. The reward difference between the two activation functions was 35.2%.

• Proceedings of the KIEE Conference
• /
• 1989.07a
• /
• pp.84-88
• /
• 1989

Recently electronic control of engine is used to satisfy exhaust gas restriction and to increase fuel economical efficiency and acceleration. In this paper, Smith-predictor method is applied to increase control performance for A/F command and disturbance. A/F control performance is simulated by A/F command and disturbance of Chevolet 5.7 liter engine and Toyota 4A-ELU engine. This method can be applied to A/F controller so that lion accurate and rapid A/F control is possible.

• 한국생물공학회:학술대회논문집
• /
• 2002.04a
• /
• pp.489-492
• /
• 2002

We have performed batch chromatography experiments to calculate parameters which are used to design SMB chromatography system. Isotherms of S- and R-ketoprofen enantiomers were gained from small amount loading experiments in a column, and flow rate of four zones of SMB chromatography were calculated using mass balance equations. As a results, diameter and length of columns were 10mm and 600mm, and flow rate of each zone were as follows; $Q_{Feed}$=1.00ml/min ,$Q_{Elu}$=2.81ml/min ,$Q_{Raf}$=1.81ml/min ,$Q_{Ext}$=2.00ml/min , and $Q_{Rec}^{TMB}$=10.75ml/min , From the Darcy's equation, the pressure drop in whole columns of SMB chromatography was 128bar.

• Microbiology and Biotechnology Letters
• /
• v.17 no.3
• /
• pp.269-272
• /
• 1989

In order to obtain monoclonal antibody from ascites fluid at sufficiently high purity using a single hydroxylapatite chromatography (HA) a further optimization on its operating variables was carried out. By adjusting the pH of the eluent, the sodium phosphate buffer, to 6.0 from 6.8 and adding CaCl$_2$to 1 mM at the column inlet, the elution molarities (M$_{elu}$) for the desired monoclonal antibody and contaminating proteins can be distinguished from each other with enough resolution. Previously these two groups of proteins co-eluted at the same time at pH 6.8 and without CaCl$_2$. This sin81e step hydroxylapatite chromatography yields the desired antibody pure enough for diagnostic use.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.18 no.10
• /
• pp.75-80
• /
• 2017

In this paper, we describe the effect of the regularization method and the network with identity mapping on the performance of the activation functions in deep convolutional neural networks. The activation functions act as nonlinear transformation. In early convolutional neural networks, a sigmoid function was used. To overcome the problem of the existing activation functions such as gradient vanishing, various activation functions were developed such as ReLU, Leaky ReLU, parametric ReLU, and ELU. To solve the overfitting problem, regularization methods such as dropout and batch normalization were developed on the sidelines of the activation functions. Additionally, data augmentation is usually applied to deep learning to avoid overfitting. The activation functions mentioned above have different characteristics, but the new regularization method and the network with identity mapping were validated only using ReLU. Therefore, we have experimentally shown the effect of the regularization method and the network with identity mapping on the performance of the activation functions. Through this analysis, we have presented the tendency of the performance of activation functions according to regularization and identity mapping. These results will reduce the number of training trials to find the best activation function.

• Journal of the Korea Convergence Society
• /
• v.11 no.12
• /
• pp.39-47
• /
• 2020

In this paper, we proposed a deep learning-based detection method using Saliency Attention to detect salient objects in images. The salient object detection separates the object where the human eye is focused from the background, and determines the highly relevant part of the image. It is usefully used in various fields such as object tracking, detection, and recognition. Existing deep learning-based methods are mostly Autoencoder structures, and many feature losses occur in encoders that compress and extract features and decoders that decompress and extend the extracted features. These losses cause the salient object area to be lost or detect the background as an object. In the proposed method, Saliency Attention is proposed to reduce the feature loss and suppress the background region in the Autoencoder structure. The influence of the feature values was determined using the ELU activation function, and Attention was performed on the feature values in the normalized negative and positive regions, respectively. Through this Attention method, the background area was suppressed and the projected object area was emphasized. Experimental results showed improved detection results compared to existing deep learning methods.

• Korean Journal of Remote Sensing
• /
• v.33 no.4
• /
• pp.423-436
• /
• 2017

This paper presents a deep learning-based road segmentation framework from very high-resolution orthophotos. The proposed method uses Deep Convolutional Autoencoders for end-to-end mapping of orthophotos to road segmentations. In addition, a set of post-processing steps were applied to make the model outputs GIS-ready data that could be useful for various applications. The optimization of the model's parameters is explained which was conducted via grid search method. The model was trained and implemented in Keras, a high-level deep learning framework run on top of Tensorflow. The results show that the proposed model with the best-obtained hyperparameters could segment road objects from orthophotos at an average accuracy of 88.5%. The results of optimization revealed that the best optimization algorithm and activation function for the studied task are Stochastic Gradient Descent (SGD) and Exponential Linear Unit (ELU), respectively. In addition, the best numbers of convolutional filters were found to be 8 for the first and second layers and 128 for the third and fourth layers of the proposed network architecture. Moreover, the analysis on the time complexity of the model showed that the model could be trained in 4 hours and 50 minutes on 1024 high-resolution images of size $106{\times}106pixels$, and segment road objects from similar size and resolution images in around 14 minutes. The results show that the deep learning models such as Convolutional Autoencoders could be a best alternative to traditional machine learning models for road segmentation from aerial photographs.

• Journal of Korea Multimedia Society
• /
• v.23 no.4
• /
• pp.525-538
• /
• 2020

A practical animal face classification system that classifies animals in image and video data is considered as a pivotal topic in machine learning. In this research, we are proposing a novel method of fully connected dual Deep Convolutional Neural Network (DCNN), which extracts and analyzes image features on a large scale. With the inclusion of the state of the art Batch Normalization layer and Exponential Linear Unit (ELU) layer, our proposed DCNN has gained the capability of analyzing a large amount of dataset as well as extracting more features than before. For this research, we have built our dataset containing ten thousand animal faces of ten animal classes and a dual DCNN. The significance of our network is that it has four sets of convolutional functions that work laterally with each other. We used a relatively small amount of batch size and a large number of iteration to mitigate overfitting during the training session. We have also used image augmentation to vary the shapes of the training images for the better learning process. The results demonstrate that, with an accuracy rate of 92.0%, the proposed DCNN outruns its counterparts while causing less computing costs.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2022.05a
• /
• pp.43-43
• /
• 2022

유해 조류 대발생은 전국 각지의 인공호소나 하천에서 다발적으로 발생하며, 경관을 해치고 수질을 오염시키는 등 수자원에 부정적인 영향을 미친다. 본 연구에서는 인공호소에서 발생하는 유해 조류 대발생을 예측하기 위해 심층학습 기법을 이용하여 예측 모델을 개발하고자 하였다. 대상 지점은 대청호의 추동 지점으로 선정하였다. 대청호는 금강유역 중류에 위치한 댐으로, 약 150만명에 달하는 급수 인구수를 유지 중이기에 유해 남조 대발생 관리가 매우 중요한 장소이다. 학습용 데이터 구축은 대청호의 2011년 1월부터 2019년 12월까지 측정된 수질, 기상, 수문 자료를 입력 자료를 이용하였다. 수질 예측 모델의 구조는 다중 레이어 퍼셉트론(Multiple Layer Perceptron; MLP)으로, 입력과 한 개 이상의 은닉층, 그리고 출력층으로 구성된 인공신경망이다. 본 연구에서는 인공신경망의 은닉층 개수(1~3개)와 각각의 레이어에 적용되는 은닉 노드 개수(11~30개), 활성함수 5종(Linear, sigmoid, hyperbolic tangent, Rectified Linear Unit, Exponential Linear Unit)을 각각 하이퍼파라미터로 정하고, 모델의 성능을 최대로 발휘할 수 있는 조건을 찾고자 하였다. 하이퍼파라미터 최적화 도구는 Tensorflow에서 배포하는 Keras Tuner를 사용하였다. 모델은 총 3000 학습 epoch 가 진행되는 동안 최적의 가중치를 계산하도록 설계하였고, 이 결과를 매 반복마다 저장장치에 기록하였다. 모델 성능의 타당성은 예측과 실측 데이터 간의 상관관계를 R2, NSE, RMSE를 통해 산출하여 검증하였다. 모델 최적화 결과, 적합한 하이퍼파라미터는 최적화 횟수 총 300회에서 256 번째 반복 결과인 은닉층 개수 3개, 은닉 노드 수 각각 25개, 22개, 14개가 가장 적합하였고, 이에 따른 활성함수는 ELU, ReLU, Hyperbolic tangent, Linear 순서대로 사용되었다. 최적화된 하이퍼파라미터를 이용하여 모델 학습 및 검증을 수행한 결과, R2는 학습 0.68, 검증 0.61이었고 NSE는 학습 0.85, 검증 0.81, RMSE는 학습 0.82, 검증 0.92로 나타났다.

• Journal of Bio-Environment Control
• /
• v.32 no.4
• /
• pp.434-441
• /
• 2023

Determining the size or area of a plant's leaves is an important factor in predicting plant growth and improving the productivity of indoor farms. In this study, we developed a convolutional neural network (CNN)-based model to accurately predict the length and width of lettuce leaves using photographs of the leaves. A callback function was applied to overcome data limitations and overfitting problems, and K-fold cross-validation was used to improve the generalization ability of the model. In addition, ImageDataGenerator function was used to increase the diversity of training data through data augmentation. To compare model performance, we evaluated pre-trained models such as VGG16, Resnet152, and NASNetMobile. As a result, NASNetMobile showed the highest performance, especially in width prediction, with an R_squared value of 0.9436, and RMSE of 0.5659. In length prediction, the R_squared value was 0.9537, and RMSE of 0.8713. The optimized model adopted the NASNetMobile architecture, the RMSprop optimization tool, the MSE loss functions, and the ELU activation functions. The training time of the model averaged 73 minutes per Epoch, and it took the model an average of 0.29 seconds to process a single lettuce leaf photo. In this study, we developed a CNN-based model to predict the leaf length and leaf width of plants in indoor farms, which is expected to enable rapid and accurate assessment of plant growth status by simply taking images. It is also expected to contribute to increasing the productivity and resource efficiency of farms by taking appropriate agricultural measures such as adjusting nutrient solution in real time.