• Communications of the Korean Institute of Information Scientists and Engineers
• /
• v.34 no.3
• /
• pp.31-34
• /
• 2016

This article reviews Apache SINGA, a general distributed deep learning (DL) platform. The system components and its architecture are presented, as well as how to configure and run SINGA for different types of distributed training using model/data partitioning. Besides, several features and performance are compared with other popular DL tools.

• Journal of the Korean Society of Industry Convergence
• /
• v.23 no.2_2
• /
• pp.263-270
• /
• 2020

In this paper, In this paper, we describe the UAV system using image processing for autonomous quadcopters, where they can apply logistics, rescue work etc. we propose high-speed hovering height and posture control method based on state feedback control with CNN from camera because we can get image of the information only every 30ms. Finally, we show the advantages of proposed method by simulations and experiments.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2019.05a
• /
• pp.153-154
• /
• 2019

There are many jobs, not actually sleeping. Sleepy driving is one of the biggest problems in modern society. In this paper, we propose a system to control underwater guns by using deep learning (YOLO) to check eyes and to check drowsiness. So let your mind be clear.

• East Asian mathematical journal
• /
• v.38 no.1
• /
• pp.95-105
• /
• 2022

In this paper, we consider a quadratic matrix equation Q(X) = AX² + BX + C = 0 where A, B, C ∈ ℝ^n×n. A new approach is proposed to find solutions of Q(X), using the novel structure of the information processing system. We also present some numerical experimetns with Artificial Neural Network.

• Journal of Internet Computing and Services
• /
• v.22 no.6
• /
• pp.1-8
• /
• 2021

With the rapid growth of intelligent devices and communication technologies, 5G network environment has become more heterogeneous and complex in terms of service management and orchestration. 5G architecture requires supportive technologies to handle the existing challenges for improving the Quality of Service (QoS) and the Quality of Experience (QoE) performances. Among many challenges, traffic steering is one of the key elements which requires critically developing an optimal solution for smart guidance, control, and reliable system. Mobile edge computing (MEC), software-defined networking (SDN), network functions virtualization (NFV), and deep learning (DL) play essential roles to complementary develop a flexible computation and extensible flow rules management in this potential aspect. In this proposed system, an accurate flow recommendation, a centralized control, and a reliable distributed connectivity based on the inspection of packet condition are provided. With the system deployment, the packet is classified separately and recommended to request from the optimal destination with matched preferences and conditions. To evaluate the proposed scheme outperformance, a network simulator software was used to conduct and capture the end-to-end QoS performance metrics. SDN flow rules installation was experimented to illustrate the post control function corresponding to DL-based output. The intelligent steering for network communication traffic is cooperatively configured in SDN controller and NFV-orchestrator to lead a variety of beneficial factors for improving massive real-time Internet of Things (IoT) performance.

• Journal of Intelligence and Information Systems
• /
• v.22 no.2
• /
• pp.127-142
• /
• 2016

Deep learning model is a kind of neural networks that allows multiple hidden layers. There are various deep learning architectures such as convolutional neural networks, deep belief networks and recurrent neural networks. Those have been applied to fields like computer vision, automatic speech recognition, natural language processing, audio recognition and bioinformatics where they have been shown to produce state-of-the-art results on various tasks. Among those architectures, convolutional neural networks and recurrent neural networks are classified as the supervised learning model. And in recent years, those supervised learning models have gained more popularity than unsupervised learning models such as deep belief networks, because supervised learning models have shown fashionable applications in such fields mentioned above. Deep learning models can be trained with backpropagation algorithm. Backpropagation is an abbreviation for "backward propagation of errors" and a common method of training artificial neural networks used in conjunction with an optimization method such as gradient descent. The method calculates the gradient of an error function with respect to all the weights in the network. The gradient is fed to the optimization method which in turn uses it to update the weights, in an attempt to minimize the error function. Convolutional neural networks use a special architecture which is particularly well-adapted to classify images. Using this architecture makes convolutional networks fast to train. This, in turn, helps us train deep, muti-layer networks, which are very good at classifying images. These days, deep convolutional networks are used in most neural networks for image recognition. Convolutional neural networks use three basic ideas: local receptive fields, shared weights, and pooling. By local receptive fields, we mean that each neuron in the first(or any) hidden layer will be connected to a small region of the input(or previous layer's) neurons. Shared weights mean that we're going to use the same weights and bias for each of the local receptive field. This means that all the neurons in the hidden layer detect exactly the same feature, just at different locations in the input image. In addition to the convolutional layers just described, convolutional neural networks also contain pooling layers. Pooling layers are usually used immediately after convolutional layers. What the pooling layers do is to simplify the information in the output from the convolutional layer. Recent convolutional network architectures have 10 to 20 hidden layers and billions of connections between units. Training deep learning networks has taken weeks several years ago, but thanks to progress in GPU and algorithm enhancement, training time has reduced to several hours. Neural networks with time-varying behavior are known as recurrent neural networks or RNNs. A recurrent neural network is a class of artificial neural network where connections between units form a directed cycle. This creates an internal state of the network which allows it to exhibit dynamic temporal behavior. Unlike feedforward neural networks, RNNs can use their internal memory to process arbitrary sequences of inputs. Early RNN models turned out to be very difficult to train, harder even than deep feedforward networks. The reason is the unstable gradient problem such as vanishing gradient and exploding gradient. The gradient can get smaller and smaller as it is propagated back through layers. This makes learning in early layers extremely slow. The problem actually gets worse in RNNs, since gradients aren't just propagated backward through layers, they're propagated backward through time. If the network runs for a long time, that can make the gradient extremely unstable and hard to learn from. It has been possible to incorporate an idea known as long short-term memory units (LSTMs) into RNNs. LSTMs make it much easier to get good results when training RNNs, and many recent papers make use of LSTMs or related ideas.

• International Journal of Computer Science & Network Security
• /
• v.21 no.6
• /
• pp.17-22
• /
• 2021

The attack trend on end-users via mobile devices is increasing in both the danger level and the number of attacks. Especially, mobile devices using the Android operating system are being recognized as increasingly being exploited and attacked strongly. In addition, one of the recent attack methods on the Android operating system is to take advantage of Android Package Kit (APK) files. Therefore, the problem of early detecting and warning attacks on mobile devices using the Android operating system through the APK file is very necessary today. This paper proposes to use the method of analyzing abnormal behavior of APK files and use it as a basis to conclude about signs of malware attacking the Android operating system. In order to achieve this purpose, we propose 2 main tasks: i) analyzing and extracting abnormal behavior of APK files; ii) detecting malware in APK files based on behavior analysis techniques using machine learning or deep learning algorithms. The difference between our research and other related studies is that instead of focusing on analyzing and extracting typical features of APK files, we will try to analyze and enumerate all the features of the APK file as the basis for classifying malicious APK files and clean APK files.

• International Journal of Computer Science & Network Security
• /
• v.23 no.8
• /
• pp.9-16
• /
• 2023

Speech can actively elicit feelings and attitudes by using words. It is important for researchers to identify the emotional content contained in speech signals as well as the sort of emotion that resulted from the speech that was made. In this study, we studied the emotion recognition system using a database in Arabic, especially in the Saudi dialect, the database is from a YouTube channel called Telfaz11, The four emotions that were examined were anger, happiness, sadness, and neutral. In our experiments, we extracted features from audio signals, such as Mel Frequency Cepstral Coefficient (MFCC) and Zero-Crossing Rate (ZCR), then we classified emotions using many classification algorithms such as machine learning algorithms (Support Vector Machine (SVM) and K-Nearest Neighbor (KNN)) and deep learning algorithms such as (Convolution Neural Network (CNN) and Long Short-Term Memory (LSTM)). Our Experiments showed that the MFCC feature extraction method and CNN model obtained the best accuracy result with 95%, proving the effectiveness of this classification system in recognizing Arabic spoken emotions.

• Journal of information and communication convergence engineering
• /
• v.18 no.4
• /
• pp.267-277
• /
• 2020

It is important to understand the exact habitat distribution of endangered species because of their decreasing numbers. In this study, we build a system with a deep learning module that collects the image data of endangered animals, processes the data, and saves the data automatically. The system provides a more efficient way than human effort for classifying images and addresses two problems faced in previous studies. First, specious answers were suggested in those studies because the probability distributions of answer candidates were calculated even if the actual answer did not exist within the group. Second, when there were more than two entities in an image, only a single entity was focused on. We applied an object detection algorithm (YOLO) to resolve these problems. Our system has an average precision of 86.79%, a mean recall rate of 93.23%, and a processing speed of 13 frames per second.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.46 no.4
• /
• pp.312-320
• /
• 2023

Recently, there has been an increasing attempt to replace defect detection inspections in the manufacturing industry using deep learning techniques. However, obtaining substantial high-quality labeled data to enhance the performance of deep learning models entails economic and temporal constraints. As a solution for this problem, semi-supervised learning, using a limited amount of labeled data, has been gaining traction. This study assesses the effectiveness of semi-supervised learning in the defect detection process of manufacturing using the MixMatch algorithm. The MixMatch algorithm incorporates three dominant paradigms in the semi-supervised field: Consistency regularization, Entropy minimization, and Generic regularization. The performance of semi-supervised learning based on the MixMatch algorithm was compared with that of supervised learning using defect image data from the metal casting process. For the experiments, the ratio of labeled data was adjusted to 5%, 10%, 25%, and 50% of the total data. At a labeled data ratio of 5%, semi-supervised learning achieved a classification accuracy of 90.19%, outperforming supervised learning by approximately 22%p. At a 10% ratio, it surpassed supervised learning by around 8%p, achieving a 92.89% accuracy. These results demonstrate that semi-supervised learning can achieve significant outcomes even with a very limited amount of labeled data, suggesting its invaluable application in real-world research and industrial settings where labeled data is limited.