• Proceedings of the Korea Information Processing Society Conference
• /
• 2009.04a
• /
• pp.50-52
• /
• 2009

Ontology maintenance is a time consuming and costly process which requires special skill and knowledge. It requires joint effort of both ontology engineer and domain specialist to properly maintain ontology and update knowledge in it. This is specially true for medical domain which is highly specialized domain. This paper proposes a novel approach for maintenance and update of existing ontologies in a medical domain. The proposed approach is based on modified Ontology Maturing Process which was originally developed for web domain. The proposed approach provides way to populate medical ontology with new knowledge obtained from medical documents. This is achieved through use of natural language processing techniques and highly specialized medical knowledge bases such as Unified Medical Language System.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.918-919
• /
• 2023

딥러닝 기술이 발전함에 따라 모델의 복잡성 역시 증가하고 있다. 본 연구에서는 모델 경량화를 위해 Knowledge Distillation 기법을 Crowd Counting Model에 적용했다. M-SFANet을 Teacher 모델로, 파라미터수가 적은 MCNN 모델을 Student 모델로 채택해 Knowledge Distillation을 적용한 결과, 기존의 MCNN 모델보다 성능을 향상했다. 이는 정확도와 메모리 효율성 측면에서 많은 개선을 이루어 컴퓨팅 리소스가 부족한 기기에서도 본 모델을 실행할 수 있어 많은 활용이 가능할 것이다.

• Korean Journal of Construction Engineering and Management
• /
• v.12 no.4
• /
• pp.3-10
• /
• 2011

Knowledge has been one of the key factors influencing the companies' competitive power. Domestic construction industries have introduced Knowledge Management from 1990s around major construction companies. However, construction organizations have recently give a skeptical response about Knowledge Management. The existing Knowledge Management in construction industries does not satisfy the demands of workers. Furthermore it makes workers do double business. Therefore, this study investigates the workers' satisfaction of Knowledge Management, and becomes the preliminary study to suggest how to improve Construction Knowledge Management.

• Korean Journal of Computational Design and Engineering
• /
• v.12 no.3
• /
• pp.153-162
• /
• 2007

In this paper, we propose a knowledge-based method for generating CNC torch path for laser cutting of the outlines of planar shapes. The proposed method consists of two main phases: laser cutting knowledge construction and CNC torch path generation using the knowledge. In the first phase, cutting experiments are conducted on various operating parameters, and then empirical data are stored and analyzed to make up the knowledge of laser cutting. With this knowledge, we can inquire what a kerf width is for specific operating parameters. In the second phase, using the knowledge of laser cutting, CNC torch path is generated for cutting the outlines of the given planar shapes. This phase is basically based on the offset generation of each outline by a sequence of arc splines, where the offset distance is the same as the half of the kerf width determined from the constructed knowledge. The proposed method based on laser cutting knowledge makes full use of arc interpolators in CNC torch path generation. The method can efficiently reduce the number of path segments while keeping the torch path within the desired accuracy.

• Journal of Korea Multimedia Society
• /
• v.24 no.12
• /
• pp.1598-1605
• /
• 2021

Recently, convolutional neural networks (CNNs) have been widely used with excellent performance in various computer vision fields, including super-resolution (SR). However, CNN is computationally intensive and requires a lot of memory, making it difficult to apply to limited hardware resources such as mobile or Internet of Things devices. To solve these limitations, network lightening studies have been actively conducted to reduce the depth or size of pre-trained deep CNN models while maintaining their performance as much as possible. This paper aims to lighten the SR CNN model, SRGAN, using the knowledge distillation among network lightening technologies; thus, it proposes four techniques with different methods of transferring the knowledge of the teacher network to the student network and presents experiments to compare and analyze the performance of each technique. In our experimental results, it was confirmed through quantitative and qualitative evaluation indicators that student networks with knowledge transfer performed better than those without knowledge transfer, and among the four knowledge transfer techniques, the technique of conducting adversarial learning after transferring knowledge from the teacher generator to the student generator showed the best performance.

• Proceedings of the Korea Institute of Fire Science and Engineering Conference
• /
• 2005.11a
• /
• pp.196-200
• /
• 2005

• Korean Journal of Computational Design and Engineering
• /
• v.11 no.2
• /
• pp.77-87
• /
• 2006

It is necessary to change the existing design process to cope with a short life-cycle product and various customer's demands. Also a frequent design change may delay the whole design process and it will increase the unit cost of the production. New alternatives or techniques have emerged to solve the existing design problems, such as a knowledge based engineering, an intelligent CAD, a function based design, and so on. In this paper, we propose a hybrid design system with a knowledge based design methodology and a function based design technique. The knowledge based design is good at a frequent design change and the function based design is effective to extract a core design behavior. In an early design process, the system utilizes a core design behavior through the function based design process. On the other hand, the system manages complicated design issues with the knowledge based design technique in the detailed design process. We conclude that the hybrid design system can bring fair effects on implementing an efficient design environment in aspect of time and expense.

• Journal of Information Technology Applications and Management
• /
• v.23 no.3
• /
• pp.13-23
• /
• 2016

In recent years, big data has usefully been deployed by organizations with the aim of getting a better prediction for the future. Moreover, knowledge management systems are being used by organizations to identify and create knowledge. Here, the output from analysis of big data and a knowledge management system are used to develop a new model with the goal of minimizing the cost of implementing new recognized processes including staff training, transferring and employment costs. Strategies are proposed from big data analysis and new processes are defined accordingly. The company requires various skills to execute the proposed processes. Organization's current experts and their skills are known through a pre-established knowledge management system. After a gap analysis, managers can make decisions about the expert arrangement, training programs and employment to bridge the gap and accomplish their goals. Finally, deduction graph is used to analyze the model.

• Journal of the Korean Society for Precision Engineering
• /
• v.22 no.1
• /
• pp.7-13
• /
• 2005

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2007.06a
• /
• pp.783-784
• /
• 2007