• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.1
• /
• pp.23-28
• /
• 2023

The installation rate of AMI(: Advanced Metering Infrastructure) capable of automatic electricity measurement is less than 43% nationwide and 10.5% in rural areas, which is very poor. Therefore, for the smart grid, automatic information recording of the watt-hour meter is required. For this purpose, it is necessary to develop a system capable of remote meter reading and use control by improving the existing watt-hour meter. In this paper, in order to enable the AMI function of the existing electricity meter, the remote meter reading and control technology of the existing electricity meter for AMI, the core of the smart grid, was developed using IoT and AI. The main research content was to recognize numbers using Tensorflow and Open-cv to convert it into a power meter sensing device for SG. We confirmed and checked the performance using the protyope system.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.33 no.4
• /
• pp.621-631
• /
• 2023

Adversarial training improves the robustness of deep neural networks for adversarial examples. However, the previous adversarial training method focuses only on the adversarial loss function, ignoring that even a small perturbation of the input layer causes a significant change in the hidden layer features. Consequently, the accuracy of a defended model is reduced for various untrained situations such as clean samples or other attack techniques. Therefore, an architectural perspective is necessary to improve feature representation power to solve this problem. In this paper, we apply an attention module that generates an attention map of an input image to a general model and performs PGD adversarial training upon the augmented model. In our experiments on the CIFAR-10 dataset, the attention augmented model showed higher accuracy than the general model regardless of the network structure. In particular, the robust accuracy of our approach was consistently higher for various attacks such as PGD, FGSM, and BIM and more powerful adversaries. By visualizing the attention map, we further confirmed that the attention module extracts features of the correct class even for adversarial examples.

• KNOM Review
• /
• v.22 no.3
• /
• pp.1-12
• /
• 2019

Recently, researching using big data and AI has emerged as a major issue in the ICT field. But, the size of big data for research is growing exponentially. In addition, users of data transmission of existing network method suggest that the problem the time taken to send and receive big data is slower than the time to copy and send the hard disk. Accordingly, researchers require dynamic and flexible network technology that can transmit data at high speed and accommodate various network structures. SDN/NFV technologies can be programming a network to provide a network suitable for the needs of users. It can easily solve the network's flexibility and security problems. Also, the problem with performing AI is that centralized data processing cannot guarantee real-time, and network delay occur when traffic increases. In order to solve this problem, the edge-computing technology, should be used which has moved away from the centralized method. In this paper, we investigate the concept and research trend of SDN, NFV, and edge-computing technologies, and analyze the trends of data central network technologies used by combining these three technologies.

• Journal of Information Technology Services
• /
• v.21 no.2
• /
• pp.61-84
• /
• 2022

With the recent application of deep learning to Natural Language Processing (NLP), the performance of NLP has improved significantly and NLP is emerging as a core competency of organizations. However, when encountering NLP use cases that are sporadically reported through various online and offline channels, it is often difficult to come up with a big picture of how to understand and interpret them or how to connect them to business. This study presents a framework for systematically analyzing NLP use cases, considering the characteristics of NLP techniques applicable to almost all industries and business functions, environmental changes in the era of the Fourth Industrial Revolution, and the effectiveness of adopting NLP reflecting all business functional areas. Through solving research questions based on the framework, the usefulness of it is validated. First, by accumulating NLP use cases and pivoting them around the business function dimension, we derive how NLP techniques are used in each business functional area. Next, by synthesizing related surveys and reports to the accumulated use cases, we draw implications for each business function and major NLP techniques. This work promotes the creation of innovative business scenarios and provides multilateral implications for the adoption of NLP by systematically viewing NLP techniques, industries, and business functional areas. The use case analysis framework proposed in this study presents a new perspective for research on new technology use cases. It also helps explore strategies that can dramatically improve organizational performance through a holistic approach that encompasses all business functional areas.

• Journal of Digital Convergence
• /
• v.20 no.5
• /
• pp.637-642
• /
• 2022

Self-authentication technology using electrocardiogram (ECG) signals is drawing attention as a self-authentication technology that can replace existing bio-authentication. A device that recognizes a digital electronic key can be mounted on a vehicle to wirelessly exchange data with a car, and a function that can lock or unlock a car door or start a car by using a smartphone can be controlled through a smartphone. However, smart keys are vulnerable to security, so smart keys applied with bio-authentication technology were studied to solve this problem and provide driver convenience. A personal authentication algorithm using electrocardiogram was mounted on a watch-type wearable device to authenticate bio, and when personal authentication was completed, it could function as a smart key of a car. The certification rate was 95 per cent achieved. Drivers do not need to have a smart key, and they propose a smart key as an alternative that can safely protect it from loss and hacking. Smart keys using personal authentication technology using electrocardiogram can be applied to various fields through personal authentication and will study methods that can be applied to identification devices using electrocardiogram in the future.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 2003.05a
• /
• pp.948-957
• /
• 2003

금형산업은 ETO(engineer to order)방식의 산업으로서 각 주문마다 견적가를 책정하는 것이 손익에 큰 영향을 미치게 된다. 현재, 대부분의 금형업체에서는 금형주문에 대한 견적을 금형전문가의 개별적인 지식에 의존하고 있는데 이러한 방식은 주관적이며, 체계적이지 못한 문제점을 갖고 있다. 따라서 보다 합리적이고 체계적인 기준에 의한 견적가 결정 방법이 필요하다. 본 연구에서는 먼저 고객의 요구사항을 QFD(Quality Function Deployment)를 사용함으로써 금형제작에 있어서 원가에 영향을 미치는 요소를 추줄하게 된다. 이렇게 얻어진 금형제작 정보와 고객의 주문정보를 바탕으로 인공신경망, 사례기반주론 등의 인공지능 기법을 사용하여, 사출금형의 견적가를 책정하는 사출금형의 견적지원시스템을 개발하고자 한다.

• The Pure and Applied Mathematics
• /
• v.15 no.4
• /
• pp.387-392
• /
• 2008

In this note, an alternative proof and extensions are provided for the following conclusions in [6, Theorem 1 and Theorem 3]: The functions $\frac1{x^2}-\frac{e^{-x}}{(1-e^{-x})^2}\;and\;\frac1{t}-\frac1{e^t-1}$ are decreasing in (0, ${\infty}$) and the function $\frac{t}{e^{at}-e^{(a-1)t}}$ for a $a{\in}\mathbb{R}\;and\;t\;{\in}\;(0,\;{\infty})$ is logarithmically concave.

• International journal of advanced smart convergence
• /
• v.13 no.2
• /
• pp.80-87
• /
• 2024

This paper focuses on improving accuracy in constrained computing settings by employing the ReLU (Rectified Linear Unit) activation function. The research conducted involves modifying parameters of the ReLU function and comparing performance in terms of accuracy and computational time. This paper specifically focuses on optimizing ReLU in the context of a Multilayer Perceptron (MLP) by determining the ideal values for features such as the dimensions of the linear layers and the learning rate (Ir). In order to optimize performance, the paper experiments with adjusting parameters like the size dimensions of linear layers and Ir values to induce the best performance outcomes. The experimental results show that using ReLU alone yielded the highest accuracy of 96.7% when the dimension sizes were 30 - 10 and the Ir value was 1. When combining ReLU with the Adam optimizer, the optimal model configuration had dimension sizes of 60 - 40 - 10, and an Ir value of 0.001, which resulted in the highest accuracy of 97.07%.

• Journal of Korea Multimedia Society
• /
• v.22 no.9
• /
• pp.1069-1079
• /
• 2019

In this study, we developed a $PM_{10}$ forecasting model using DNN and Membership Function, and improved the forecasting performance. The model predicts the $PM_{10}$ concentrations of the next 3 days in the Seoul area by using the weather and air quality observation data and forecast data. The best model(RM14)'s accuracy (82%, 76%, 69%) and false alarm rate(FAR:14%,33%,44%) are good. Probability of detection (POD: 79%, 50%, 53%), however, are not good performance. These are due to the lack of training data for high concentration $PM_{10}$ compared to low concentration. In addition, the model dose not reflect seasonal factors closely related to the generation of high concentration $PM_{10}$. To improve this, we propose Julian date membership function as inputs of the $PM_{10}$ forecasting model. The function express a given date in 12 factors to reflect seasonal characteristics closely related to high concentration $PM_{10}$. As a result, the accuracy (79%, 70%, 66%) and FAR (24%, 48%, 46%) are slightly reduced in performance, but the POD (79%, 75%, 71%) are up to 25% improved compared with those of the RM14 model. Hence, this shows that the proposed Julian forecast model is effective for high concentration $PM_{10}$ forecasts.

• Journal of Industrial Convergence
• /
• v.21 no.1
• /
• pp.175-185
• /
• 2023

Recently, novel viral infections such as COVID-19 have spread and pose a serious public health problem. In particular, these diseases have a fatal effect on the elderly, threatening life and causing serious social and economic losses. Accordingly, applications such as telemedicine, healthcare, and disease prevention using the Internet of Things (IoT) and artificial intelligence (AI) have been introduced in many industries to improve disease detection, monitoring, and quarantine performance. However, since existing technologies are not applied quickly and comprehensively to the sudden emergence of infectious diseases, they have not been able to prevent large-scale infection and the nationwide spread of infectious diseases in society. Therefore, in this paper, we try to predict the spread of infection by collecting various infection information with regional limitations through a virus disease information collector and performing AI analysis and severity matching through an AI broker. Finally, through the Korea Centers for Disease Control and Prevention, danger alerts are issued to the elderly, messages are sent to block the spread, and information on evacuation from infected areas is quickly provided. A realistic elderly support system compares the location information of the elderly with the information of the infected area and provides an intuitive danger area (infected area) avoidance function with an augmented reality-based smartphone application. When the elderly visit an infected area is confirmed, quarantine management services are provided automatically. In the future, the proposed system can be used as a method of preventing a crushing accident due to sudden crowd concentration in advance by identifying the location-based user density.