• International journal of advanced smart convergence
• /
• v.12 no.4
• /
• pp.458-463
• /
• 2023

This study presents a paper on the implementation of a Raspberry Pi-based educational smart farm system. It confirms that in a real smart farm environment, the control of temperature, humidity, soil moisture, and light intensity can be smoothly managed. It also includes remote monitoring and control of sensor information through a web service. Additionally, information about intruders collected by the Pi camera is transmitted to the administrator. Although the cost of existing smart farms varies depending on the location, material, and type of installation, it costs 400 million won for polytunnel and 1.5 billion won for glass greenhouses when constructing 0.5ha (1,500 pyeong) on average. Nevertheless, among the problems of smart farms, there are lax locks, malfunctions to automation, and errors in smart farm sensors (power problems, etc.). We believe that this study can protect crops at low cost if it is complementarily used to improve the security and reliability of expensive smart farms. The cost of using this study is about 100,000 won, so it can be used inexpensively even when applied to the area. In addition, in the case of plant cultivators, cultivators with remote control functions are sold for more than 1 million won, so they can be used as low-cost plant cultivators.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.17 no.1
• /
• pp.47-52
• /
• 2024

Through the coronavirus pandemic, research on the use and advancement of blockchain-based decentralized identity authentication (Decentralized ID) technology is being actively conducted in various fields, centered on the central government, local governments, and private businesses. In this paper, we introduce the results of development based on Hyperledger Fabric to change the existing central server-based identity authentication to a decentralized one. These development results can strengthen the security and transparency of identity authentication systems for commercial purposes and provide stable services for user ID issuance, inquiry, and disposal. In addition, the decentralized identity authentication system verified performance results of DID creation of 262,000 rps and DID inquiry of 1,850 rps, DID VP creation of 200 rps, and DID VP inquiry of 220 rps or less through public authentication.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.2
• /
• pp.494-510
• /
• 2024

Internet users are exposed to sophisticated cyberattacks that intrusion detection systems have difficulty detecting. Therefore, research is increasing on intrusion detection methods that use artificial intelligence technology for detecting novel cyberattacks. Unsupervised learning-based methods are being researched that learn only from normal data and detect abnormal behaviors by finding patterns. This study developed an anomaly-detection method based on unsupervised machines and deep learning for a network intrusion detection system (NIDS). We present a hybrid anomaly detection approach based on unsupervised learning techniques using the autoencoder (AE), Isolation Forest (IF), and Local Outlier Factor (LOF) algorithms. An oversampling approach that increased the detection rate was also examined. A hybrid approach that combined deep learning algorithms and traditional machine learning algorithms was highly effective in setting the thresholds for anomalies without subjective human judgment. It achieved precision and recall rates respectively of 88.2% and 92.8% when combining two AEs, IF, and LOF while using an oversampling approach to learn more unknown normal data improved the detection accuracy. This approach achieved precision and recall rates respectively of 88.2% and 94.6%, further improving the detection accuracy compared with the hybrid method. Therefore, in NIDS the proposed approach provides high reliability for detecting cyberattacks.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.4
• /
• pp.903-921
• /
• 2024

Industry 5.0 puts forward higher requirements for smart cities, including low-carbon, sustainable, and people-oriented, which pose challenges to the design of smart cities. In response to the above challenges, this study introduces the cyber-physical-social system (CPSS) and parallel system theory into the design of smart cities, and constructs a smart city framework based on parallel system theory. On this basis, in order to enhance the security of smart cities, a sustainable patrol subsystem for smart cities has been established. The intelligent patrol system uses a drone platform, and the trajectory planning of the drone is a key problem that needs to be solved. Therefore, a mathematical model was established that considers various objectives, including minimizing carbon emissions, minimizing noise impact, and maximizing coverage area, while also taking into account the flight performance constraints of drones. In addition, an improved metaheuristic algorithm based on ant colony optimization (ACO) algorithm was designed for trajectory planning of patrol drones. Finally, a digital environmental map was established based on real urban scenes and simulation experiments were conducted. The results show that compared with the other three metaheuristic algorithms, the algorithm designed in this study has the best performance.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.6
• /
• pp.1518-1539
• /
• 2024

Artificial intelligence-enabled business models aim to improve decision-making, operational efficiency, innovation, and productivity. The presented systematic literature review is conducted to highlight elucidating the utilization of artificial intelligence (AI) methods and techniques within AI-enabled businesses, the significance and functions of AI-enabled organizational models and frameworks, and the design parameters employed in academic research studies within the AI-enabled business domain. We reviewed 39 empirical studies that were published between 2010 and 2023. The studies that were chosen are classified based on the artificial intelligence business technique, empirical research design, and SLR search protocol criteria. According to the findings, machine learning and artificial intelligence were reported as popular methods used for business process modelling in 19% of the studies. Healthcare was the most experimented business domain used for empirical evaluation in 28% of the primary research. The most common reason for using artificial intelligence in businesses was to improve business intelligence. 51% of main studies claimed to have been carried out as experiments. 53% of the research followed experimental guidelines and were repeatable. For the design of business process modelling, eighteen AI mythology were discovered, as well as seven types of AI modelling goals and principles for organisations. For AI-enabled business models, safety, security, and privacy are key concerns in society. The growth of AI is influencing novel forms of business.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.27 no.2
• /
• pp.256-264
• /
• 2024

In the last 10 years, AI has made rapid progress, and image classification, in particular, are showing excellent performance based on deep learning. Nevertheless, due to the nature of deep learning represented by a black box, it is difficult to actually use it in critical decision-making situations such as national defense, autonomous driving, medical care, and finance due to the lack of explainability of judgement results. In order to overcome these limitations, in this study, a model description algorithm capable of local interpretation was applied to the inception network-derived AI to analyze what grounds they made when classifying national defense data. Specifically, we conduct a comparative analysis of explainability based on confidence values by performing LIME analysis from the Inception v2_resnet model and verify the similarity between human interpretations and LIME explanations. Furthermore, by comparing the LIME explanation results through the Top1 output results for Inception v3, Inception v2_resnet, and Xception models, we confirm the feasibility of comparing the efficiency and availability of deep learning networks using XAI.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.20 no.1
• /
• pp.59-74
• /
• 2024

In order to improve safety consciousness due to the increase in disasters and safety accidents, safety education is necessary to prepare for disasters with interest in safety. This study is a board game design for disaster safety education for elementary school students based on Keller's learning motivation theory. By considering the school safety curriculum and the safety education contents of the School Safety Mutual Aid Association and the Ministry of Public Administration and Security, the content and goals of learning were derived and the order of learning was determined. When designing game content, the fun elements of the game were applied to Keller's learning motivation inducing factors such as attention concentration (A), relevance (R), confidence (C), satisfaction (S), and educational game design elements to induce the achievement of learning goals at the game planning stage. It is expected that the existing safety education focusing on lecture-style and audiovisual will be supplemented and used in the educational field.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2024.05a
• /
• pp.323-326
• /
• 2024

본 연구에서는 커널 퍼저인 Syzkaller 를 사용하여 리눅스 커널의 네트워크 서브시스템을 퍼징하고 그 결과를 분석하여, 높은 커버리지를 달성하기가 왜 어려운지 분석하고 이를 개선하기 위한 방법들을 제안한다. 첫 번째 실험에서는 TCP 및 IPv4 소켓과 관련된 시스템 콜 및 매개변수만 허용하여 리눅스 커널 네트워크 퍼징을 진행하고, 두 번째 실험에서는 Syzkaller 가 지원하는 모든 시스템 콜을 포함하도록 범위를 확장한다. 첫 번째 실험 결과, 퍼징 시작 약 55 시간만에 TCP 연결 수립에 성공하였다. 두 번째 실험 결과, 첫 번째 실험보다 전반적인 커버리지와 라우팅 서브시스템의 커버리지는 개선되었으나 TCP 연결 수립에는 실패하였다. TCP 연결 수립을 위해서는 서버의 IP 주소 및 포트번호를 클라이언트가 무작위 입력 생성을 통해 맞혀야 하는데, 이 과정에서 시간이 오래 걸리기 때문에 연결 수립이 쉽게 이루어지지 않는 것으로 분석된다. 추가적으로, 본 연구에서는 TCP 연결 수립을 쉽게 하기 위한 하이브리드 퍼징, IP 패킷 포워딩 허용, 패킷 description 없이 퍼징 등 Syzkaller 를 이용하여 리눅스 커널 네트워크 서브시스템을 더 효율적으로 퍼징할 수 있는 방법들을 제안한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2024.05a
• /
• pp.711-714
• /
• 2024

In the dynamic landscape of modern machine learning, Federated Learning (FL) has emerged as a compelling paradigm designed to enhance privacy by enabling participants to collaboratively train models without sharing their private data. Specifically, Distillation-based Federated Learning, like Federated Learning with Model Distillation (FedMD), Federated Gradient Encryption and Model Sharing (FedGEMS), and Differentially Secure Federated Learning (DS-FL), has arisen as a novel approach aimed at addressing Non-IID data challenges by leveraging Federated Learning. These methods refine the standard FL framework by distilling insights from public dataset predictions, securing data transmissions through gradient encryption, and applying differential privacy to mask individual contributions. Despite these innovations, our survey identifies persistent vulnerabilities, particularly concerning the susceptibility to logit inversion attacks where malicious actors could reconstruct private data from shared public predictions. This exploration reveals that even advanced Distillation-based Federated Learning systems harbor significant privacy risks, challenging the prevailing assumptions about their security and underscoring the need for continued advancements in secure Federated Learning methodologies.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2024.05a
• /
• pp.623-626
• /
• 2024

최근 거대 언어 모델의 발전으로 프로그램 합성 분야에서 활용되고 있는 코드 생성 언어 모델의 보안적 측면에 대한 중요성이 부각되고 있다. 그러나, 이를 위해 모델 전체를 재학습하기에는 많은 자원과 시간이 소모된다. 따라서, 본 연구에서는 효율적인 미세조정 방식 중 하나인 프롬프트 튜닝으로 코드 생성 언어 모델이 안전한 코드를 생성할 확률을 높이는 방법을 탐구한다. 또한 이에 따른 기능적 정확성 간의 상충 관계를 분석한다. 실험 결과를 통해 프롬프트 튜닝이 기존 방법에 비해 추가 파라미터를 크게 줄이면서도 보안률을 향상시킬 수 있음을 알 수 있었다. 미래 연구 방향으로는 새로운 조정 손실함수와 하이퍼파라미터 값을 조정하여 성능을 더욱 향상시킬 수 있는지 조사할 것이다. 이러한 연구는 보다 안전하고 신뢰할 수 있는 코드 생성을 위한 중요한 발전을 이끌 수 있을 것으로 기대된다.