• Journal of the Korea Computer Graphics Society
• /
• v.26 no.3
• /
• pp.61-68
• /
• 2020

In this paper, we propose a deep learning framework for predicting a depth map of a 360° panorama image. Previous works use synthetic 360° panorama datasets to train networks due to the lack of realistic datasets. However, the synthetic nature of the datasets induces features extracted by the networks to differ from those of real 360° panorama images, which inevitably leads previous methods to fail in depth prediction of real 360° panorama images. To address this gap, we use domain adaptation to learn features shared by real and synthetic panorama images. Experimental results show that our approach can greatly improve the accuracy of depth estimation on real panorama images while achieving the state-of-the-art performance on synthetic images.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.33 no.10
• /
• pp.17-25
• /
• 2005

A grid adaptation method within systems of chimera and patched grids is presented. Problem domains are divided into near-body and off-body fields. Near-body field is filled with curvilinear body-fitted grids that extend only a short distance from body surfaces and connected to other grid systems via chimera domain connectivity method. Off-body field is filled with patched uniform cartesian grids of varying levels of refinement. This method gives flexibility in grid generation and efficient adaptation capability. Several numerical experiments including 2D store separation were performed to show the performance of the proposed adaptation method.

• Proceedings of the Korean Society for Cognitive Science Conference
• /
• 2000.06a
• /
• pp.163-168
• /
• 2000

기계번역에서 좋은 품질의 번역 결과를 얻기 위해서는 대상으로 하고 잇는 전문 영역에 맞게 시스템의 번역 지식을 조정해야 한다. 본 연구에서는 대상 영역 코퍼스를 이용하여 기계번역 시스템의 특정 영역화를 지원하는 워크벤치를 설계하고 구현한다. 워크벤치는 대상 영역의 코퍼스에서 대상 영역의 지식을 추출하는 영역 지식 추출기와, 추출된 지식을 사용자에게 제시하여 사용자가 사전을 편집할 수 있는 환경을 제공하는 영역 지식 검색기와 사전 편집기로 구성된다. 구혀된 워크벤치를 이용하여 일반 영역 사전을 군사 정보 영역으로 특정 영역화를 해 본 결과, 효율성과 정확성에서의 향상이 있었다.

• Proceedings of the Korean Information Science Society Conference
• /
• 1998.10c
• /
• pp.75-77
• /
• 1998

Perhaps one of the most versatile approaches to learning in practical domains lies in case based reasoning. To date, however, most case based reasoning systems have tended to focus on relatively simple domains. The current study involves the development of a decision support system for a complex production process with a limited database. This paper presents a set of critical issues underlying CBR, then explores their consequences for a complex domain. Finally, the performance of the system is examined for resolving various types of quality control problems.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.12 no.1
• /
• pp.77-85
• /
• 2012

This paper describes a novel approach to adaptive courseware generation. This approach adopts its structure from existing intelligent tutoring systems and introduces a new component called pedagogical scenario model to support pedagogical flexibility in the adaptation process of courseware generation system. The adaptation is carried out using Dynamic Constraint Satisfaction Problem framework, which is a variant of classical Constraint Satisfaction Problem, to deliver courseware tailored to individual learner. Such a framework provides a high level of expressiveness to deal with the particular characteristics of courseware generation problem. Further, it automatically designs a sound courseware satisfying the design constraints imposed by the domain, the pedagogical scenario and learner models.

• Journal of Korean Clinical Nursing Research
• /
• v.22 no.3
• /
• pp.294-304
• /
• 2016

Purpose: This study was done to develop an evidence-based external ventricular drainage (EVD) nursing practice guideline in order to provide standardized nursing and prevent EVD related complications. Methods: We used the standardized methodology for nursing practice guideline adaptation developed by Korean Hospital Nurses Association for the guideline adaptation process in this study. Results: The newly developed EVD nursing practice guideline was adapted to the American Association of Neuroscience Nurses (AANN)'s clinical practice guideline which is 'Care of the patient undergoing intra-cranial pressure monitoring/external ventricular drainage of lumbar drainage.' There were 61 recommendations documented in the preliminary guideline all evaluated by 9 experts based on acceptability and applicability. The final practice guideline was composed of 3 domains with 57 recommendations. The three domains of nursing were the insertion, maintenance, and removal of the EVD. The number of recommendations in each domain was 8 in EVD insertions, 39 in EVD maintenance, and 10 in EVD removals. Of the 57 recommendations 3.5% were level 1, 31.5% were level 2, and 65% were level 3. Conclusion: The standardized practice guideline can improve nurses' performance and accuracy. It can also be used as the foundation for effective communication between all medical staff.

• Proceedings of the Korean Information Science Society Conference
• /
• 2010.06a
• /
• pp.98-99
• /
• 2010

• Journal of Information Processing Systems
• /
• v.17 no.5
• /
• pp.867-878
• /
• 2021

In context awareness and user intention tasks, dataset construction is expensive because specific domain data are required. Although pretraining with a large corpus can effectively resolve the issue of lack of data, it ignores domain knowledge. Herein, we concentrate on data domain knowledge while addressing data scarcity and accordingly propose a multi-channel long short-term memory (LSTM). Because multi-channel LSTM integrates pretrained vectors such as task and general knowledge, it effectively prevents catastrophic forgetting between vectors of task and general knowledge to represent the context as a set of features. To evaluate the proposed model with reference to the baseline model, which is a single-channel LSTM, we performed two tasks: voice phishing with context awareness and movie review sentiment classification. The results verified that multi-channel LSTM outperforms single-channel LSTM in both tasks. We further experimented on different multi-channel LSTMs depending on the domain and data size of general knowledge in the model and confirmed that the effect of multi-channel LSTM integrating the two types of knowledge from downstream task data and raw data to overcome the lack of data.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.11 no.3
• /
• pp.600-607
• /
• 2007

In the next generation mobile multimedia communications, the broad band shot-burst transmissions are used to reduce end-to-end transmission delay, and to limit the time variation of wireless channels over a burst. However, training overhead is very significant for such short burst formats. So, the availability of the short training sequence and the fast converging adaptive algorithm is essential in the system adopting the symbol-by-symbol adaptive equalizer. In this paper, we propose an adaptive equalizer using the DWMT (discrete multi-wavelet transform) and LMS (least mean square) adaptation. The proposed equalizer has a faster convergence rate than that of the existing transform-domain equalizers, while the increase of computational complexity is very small.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.5
• /
• pp.117-122
• /
• 2024

Many studies on software defect prediction have been conducted, but it has been difficult to use them due to a lack of training data. Cross-project defect prediction is a technique to solve this problem, where a prediction model learned with sufficient training data from existing source project is used to predict defects in the target project. Before learning, domain adaptation techniques, a type of transfer learning, are used to minimize the difference in data distribution between the two projects. In this paper, we produced new prediction models using W-BDA and MEDA and compared their performance with existing models using TCA and BDA. As a result of the evaluation experiment, MEDA showed irregular and poor performance compared to other models, but BDA showed better performance than TCA, and W-BDA showed slightly better performance than BDA.