• Journal of the Korean Society of Marine Environment & Safety
• /
• v.19 no.2
• /
• pp.207-212
• /
• 2013

In this study, distributed the ship's Centralized Air Conditioning System the way an individual to replace the air conditioning system by using Air-cooled air conditioner. Research results, Individually separated air conditioning system complement the heat source control and thermal efficiency problems and improves the efficiency of the device was confirmed. In addition, under the same conditions refrigeration capacity and coefficient of performance of the device, an average of about 3 %, 23 ~ 26 %, higher, Chilled Water Plants Compressor power consumption is about 12 % lower. Also while heating under the same conditions, power consumption is about 33.5 % lower. Therefore Individually Separated Air Conditioning System greatly contributed to the improved performance of the device and living spaces for comfortable temperature and humidity control as well as heating source could be obtained.

• Journal of Internet Computing and Services
• /
• v.25 no.2
• /
• pp.1-9
• /
• 2024

Although federated learning is designed to be safer than centralized methods in terms of security and privacy, it still has many vulnerabilities. An attacker performing an adversarial attack intentionally manipulates the deep learning model by injecting carefully crafted input data, that is, adversarial examples, into the client's training data to induce misclassification. A common defense strategy against this is so-called adversarial training, which involves preemptively learning the characteristics of adversarial examples into the model. Existing research assumes a scenario where all clients are under adversarial attack, but considering the number of clients in federated learning is very large, this is far from reality. In this paper, we experimentally examine aspects of adversarial training in a scenario where some of the clients are under attack. Through experiments, we found that there is a trade-off relationship in which the classification accuracy for normal samples decreases as the classification accuracy for adversarial examples increases. In order to effectively utilize this trade-off relationship, we present a method to perform adversarial training by adaptively selecting a loss function depending on whether the client is attacked.

• Asia pacific journal of information systems
• /
• v.15 no.1
• /
• pp.115-133
• /
• 2005

This study is to develop an information security management model(ISMM) for small and medium sized enterprises(SMEs). Based on extensive literature review, a five-pillar twelve-component reference ISMM is developed. The five pillars of SME's information security are: centralized decision making, ease of management, flexibility, agility and expandability. Twelve components are: scope & organization, security policy, resource assessment, risk assessment, implementation planning, control development, awareness training, monitoring, change management, auditing, maintenance and accident management. Subsequent survey designed and administered to expose experts' perception on the importance of these twelve components revealed that five out of tweleve components require relatively immediate attention than others, especially in SME's context. These five components are: scope and organization, resource assessment, auditing, change management, and incident management. Other seven components are policy, risk assessment, implementation planning, control development, awareness training, monitoring, and maintenance. It seems that resource limitation of SMEs directs their attention to ISMM activities that may not require a lot of resources. On the basis of these findings, a three-phase approach is developed and proposed here as an SME ISMM. Three phases are (1) foundation and promotion, (2) management and expansion, and (3) maturity. Implications of the model are discussed and suggestions are made for further research.

• Asia-Pacific Journal of Business
• /
• v.3 no.2
• /
• pp.23-33
• /
• 2012

Sustainable competitive advantage depends on formulation and implementation of appropriate human resource management (HRM) strategies and deployment of their competencies effectively in dynamic and complex environment. Competencies deployment is an approach to make decisions on the intentions and plans of organization concerning employment relationship and its recruitment, training, development, performance appraisal, reward and employee relations policies and practice. The improvement of organizational effectiveness is an overall objective of people management in organizations. In light of these, this study explores the present practices of HRM like: HR departments, HR policies and strategies, HR planning, recruitment selection and placement, training and development, performance appraisal, compensation and benefit, employee relations and communications. Based on the survey in 204 Nepalese organizations as a unit of analysis, the study concludes that HR practices in few organizations have action program for minorities, ethnic group, older employees and people with disabilities. Due to centralized organizational structure, most of HR decisions are taken into central office and line managers being involved highly in planning and implementing HR policies. In a nutshell, though HRM practices are not developed like developed countries, Nepalese organizations are realizing the significance of people management at work and changing their practices in the present dynamic environment.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.43 no.4
• /
• pp.543-552
• /
• 1994

This paper presents a decentralized quadratic regulation architecture with feedforward neural networks for the control problem of complex systems. In this method, the decentralized technique was used to treat several simple subsystems instead of a full complex system in order to reduce training time of neural networks, and the neural networks' nonlinear mapping ability is exploited to handle the nonlinear interaction variables between subsystems. The decentralized regulating architecture is composed of local neuro-controllers, local neuro-identifiers and an overall interaction neuro-identifier. With the interaction neuro-identifier that catches interaction characteristics, a local neuro-identifier is trained to simulate a subsystem dynamics. A local neuro-controller is trained to learn how to control the subsystem by using generalized Backprogation Through Time(BTT) algorithm. The proposed neural network based decentralized regulating scheme is applied in the power System Stabilization(PSS) control problem for an imterconnected power system, and compared with that by a conventional centralized LQ regulator for the power system.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.05a
• /
• pp.447-450
• /
• 2021

시뮬레이션을 비롯한 많은 다중 에이전트 환경에서는 중앙 집중 훈련 및 분산 수행(centralized training with decentralized execution; CTDE) 방식이 활용되고 있다. CTDE 방식 하에서 중앙 집중 훈련 및 분산 수행 환경에서의 다중 에이전트 학습을 위한 상태 행동 가치 기반(state-action value; Q-value) 다중 에이전트 알고리즘들에 대한 많은 연구가 이루어졌다. 이러한 알고리즘들은 Independent Q-learning (IQL)이라는 강력한 벤치 마크 알고리즘에서 파생되어 다중 에이전트의 공동의 상태 행동 가치의 분해(Decomposition) 문제에 대해 집중적으로 연구되었다. 본 논문에서는 앞선 연구들에 관한 알고리즘들에 대한 분석과 실용적이고 일반적인 도메인에서의 실험 분석을 통해 검증한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.05a
• /
• pp.228-230
• /
• 2020

The emergence of collaborative learning to the public is to tackle the user's privacy issue in centralized learning by bringing the AI models to the data source or client device for training Collaborative learning employs computing and storage resources on the client's device. Thus, it is privacy preserved by design. In harmony, blockchain is also prominent since it does not require an intermediary to process a transaction. However, these approaches are not yet fully ripe to be implemented in the real world, especially for the complex system (several challenges need to be addressed). In this work, we present the performance of collaborative learning and potential use case of blockchain. Further, we discuss privacy issues in the system.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.11a
• /
• pp.377-379
• /
• 2022

POI refers to the point of Interest in Location-Based Social Networks (LBSNs). With the rapid development of mobile devices, GPS, and the Web (web2.0 and 3.0), LBSNs have attracted many users to share their information, physical location (real-time location), and interesting places. The tremendous demand of the user in LBSNs leads the recommendation systems (RSs) to become more widespread attention. Recommendation systems assist users in discovering interesting local attractions or facilities and help social network service (SNS) providers based on user locations. Therefore, it plays a vital role in LBSNs, namely POI recommendation system. In the machine learning model, most of the training data are stored in the centralized data storage, so information that belongs to the user will store in the centralized storage, and users may face privacy issues. Moreover, sharing the information may have safety concerns because of uploading or sharing their real-time location with others through social network media. According to the privacy concern issue, the paper proposes a recommendation model to prevent user privacy and eliminate traditional RS problems such as cold-start and data sparsity.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.21 no.4
• /
• pp.409-420
• /
• 2015

In this study, I researched, analysed and compared the examples of civilian maritime search and rescue auxiliaries of world leading countries' such as Japan and America, and the Korean volunteer fire fighting team. Through this, I suggest that the decentralized civilian maritime search and rescue forces of Maritime Rescue and Salvage Association and KCG Civilian Auxiliary should be united into unified search and rescue system and establish legal basis for stabilized support and development. It seem that we should organize laws and regulations for the government to have centralized control of rescue mission as in the cases of America and Canada and elivate rescue mission capability with systemized education and training entrusted to specialized external training organization. I proposed to establish financial support such as fund for the stabilized status of civilian auxiliary, domestic and overseas training session for the civilian auxiliary to inspire integrity and sense of duty as a part of maritime search and rescue forces.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.4
• /
• pp.143-152
• /
• 2021

It is very important to learn behavioral policies that allow multiple agents to work together organically for common goals in various real-world applications. In this multi-agent reinforcement learning (MARL) environment, most existing studies have adopted centralized training with decentralized execution (CTDE) methods as in effect standard frameworks. However, this multi-agent reinforcement learning method is difficult to effectively cope with in a dynamic environment in which new environmental changes that are not experienced during training time may constantly occur in real life situations. In order to effectively cope with this dynamic environment, this paper proposes a novel multi-agent reinforcement learning system, C-COMA. C-COMA is a continual learning model that assumes actual situations from the beginning and continuously learns the cooperative behavior policies of agents without dividing the training time and execution time of the agents separately. In this paper, we demonstrate the effectiveness and excellence of the proposed model C-COMA by implementing a dynamic mini-game based on Starcraft II, a representative real-time strategy game, and conducting various experiments using this environment.