• 한국산학기술학회논문지
• /
• 제11권4호
• /
• pp.1285-1293
• /
• 2010

본 연구는 원격대학교 신입 성인학습자의 성적향상과 중도학습포기의 방지를 위하여 신입생에게 e-멘토링 활동을 참여시키고, 하나 이상의 역할모형을 설정하여 운영할 수 있는 e-멘토링 시스템의 설계 및 구축 방법을 제시하고자 한다. 이 시스템의 장점으로, 첫째 원격대학교 멘토링 참가자들이 역할모형을 설정하고 멘토-멘티 매칭 시 상호요구사항을 반영한 자동 매칭시스템을 구축하여 참가자들의 참여율을 높이고, 성적향상이 가능하도록 하였다. 둘째 멘토링 활동 종료 후 역할모형 및 활동내역을 평가하는 시스템을 구축하여 우수 참가자에는 보상체계를 수립하고, 우수 멘토에게는 연속참여를 유도하여 경험이 풍부한 멘토가 지속적으로 신입생 성인학습자의 학습부적응에 도움을 줄 수 있도록 하였다.

• 한국정보통신학회논문지
• /
• 제26권5호
• /
• pp.662-667
• /
• 2022

의미역 결정 작업은 서술어와 문장 내 행위자, 피행위자, 장소, 시간 등 서술어와 관련 있는 논항들을 추출하는 작업이다. 기존 의미역 결정 방법은 문장의 언어학적 특징 추출을 위한 파이프라인을 구축하는데, 파이프라인 내 각 추출 작업들의 오류가 의미역 결정 작업의 성능에 영향을 미치기 때문에 현재는 End-to-End 방법의 신경망 모델을 이용한 방법들이 제안되고 있다. 본 논문에서는 의미역 결정 작업을 위해 Biaffine Average Attention 구조를 이용한 신경망 모델을 제안한다. 제안하는 모델은 기존 연구에서 제안된 특정 시점에 대한 레이블 예측을 위해 주변 시점 정보를 이용하는 LSTM 모델 대신 문장 내 서술어와 논항의 거리에 상관없이 문장 전체 정보에 집중할 수 있는 Biaffine Average Attention 구조로 이루어져 있다. 제안하는 모델의 성능 평가를 위해 F1 점수를 이용하여 기존 연구에서 제안한 BERT 기반의 모델들과 비교하였으며, 76.21%의 성능으로 비교 모델보다 높은 성능을 보였음을 확인하였다.

• Korean Journal of Radiology
• /
• 제25권3호
• /
• pp.224-242
• /
• 2024

The emergence of Chat Generative Pre-trained Transformer (ChatGPT), a chatbot developed by OpenAI, has garnered interest in the application of generative artificial intelligence (AI) models in the medical field. This review summarizes different generative AI models and their potential applications in the field of medicine and explores the evolving landscape of Generative Adversarial Networks and diffusion models since the introduction of generative AI models. These models have made valuable contributions to the field of radiology. Furthermore, this review also explores the significance of synthetic data in addressing privacy concerns and augmenting data diversity and quality within the medical domain, in addition to emphasizing the role of inversion in the investigation of generative models and outlining an approach to replicate this process. We provide an overview of Large Language Models, such as GPTs and bidirectional encoder representations (BERTs), that focus on prominent representatives and discuss recent initiatives involving language-vision models in radiology, including innovative large language and vision assistant for biomedicine (LLaVa-Med), to illustrate their practical application. This comprehensive review offers insights into the wide-ranging applications of generative AI models in clinical research and emphasizes their transformative potential.

• 의학교육논단
• /
• 제24권1호
• /
• pp.46-55
• /
• 2022

This study aimed to analyze the subjects, situations, and reflection levels related to role modeling experienced by medical students during their clinical clerkship and their own reflections. This study intends to suggest ways of improving how residents and clinical faculty should treat and teach medical students. Written interviews were conducted regarding senior medical students' role modeling experiences during their clinical clerkships in 2018 and 2019. Content analysis was conducted for a total of 224 cases from 196 students. Content analysis revealed three types of role modeling content: subjects (faculty, residents, nurses, peer students), situations (clinical competence, personal qualities, teaching skills), and the level of reflection (critical reflection, reflection, thoughtful action, and habitual action) in each case. As role model subjects, faculty were found to be the paramount role model (n=142, 62.83%). Role modeling was the most frequently performed for clinical competence (n=103, 45.98%). Clinical competence was frequently shown in communication and empathic listening during rounds and outpatient relationships between the patient and doctor. Regarding the level of reflection for role modeling, the number of critical reflections was 86 (38.39%) and that of reflections was 80 (35.71%). In particular, negative role modeling showed a high level of critical reflection in relation to faculty (64.44%) and nurses (8.89%). In conclusion, role modeling of medical students participating in clinical clerkships occurs in situations that the role models are not aware of, with positive or negative effects on the formation of professional identity among medical students.

• Journal of Magnetics
• /
• 제20권3호
• /
• pp.284-289
• /
• 2015

This paper discusses general methods of modelling magnetic saturation in steady-state, two-axis (d & q) frame models of dual stator induction generators (DSIG). In particular, the important role of the magnetic coupling between the d-q axes (cross-magnetizing phenomenon) is demonstrated, with and without cross-saturation. For that purpose, two distinct models of DSIGs, with and without cross-saturation, are specified. These two models are verified by an application that is sensitive to the presence of cross-saturation, to prove the validity of these final methods and the equivalence between all developed models. Advantages of some of the models over the existing ones and their applicability are discussed. In addition, an alternative is given to evaluate all saturation factors (static and dynamic) by just calculating the static magnetizing inductance which is simply the magnitude of the ratio of the magnetizing flux to the current. The comparison between the simulation results of the proposed model with experimental results gives a good correspondence, especially at startup.

• 한국대기환경학회지
• /
• 제14권2호
• /
• pp.81-94
• /
• 1998

Six complex terrain dispersion models recommended by the U. S. Environmental Protection Agency were investigated using a hypothetical case in which a plume approaches complex terrain. The six models considered were Valley, CTSCREEN, COMPLEX 1, SHORTZ, RTDM, and CTDMPLUS, the latter four being closely studied. Highest concentrations were predicted for 48 receptors and plume behaviors were compared for stable and unstable meteorological conditions. Under stable conditions, ground-level concentrations were determined by the height of the plume centerline above the terrain. The concentrations estimated by SHORTZ and COMPLEX I were higher than those estimated by CTSCREEN, with CTDMPLUS predicting the lowest concentrations. In particular, the height of the lift midpoint, as well as the co.nterline of the plume, are important in the model calculation of CTDMPLUS. Under unstable conditions, the vertical dispersion plays a key role in determining ground -level concentrations. For this case, concentrations predicted by CTDMPLUS were the 'highest, whereas those predicted by SHORTZ were the lowest. Concentration distributions predicted by CTDMPLUS are quite similar to typical Gaussian distributions even on complex terrain, except for a slight shift of the plume centerline due to the of(tract of the geostrophic wind. In addition,24-hour average concentrations were estimated for comparison with results from the Valley model. Among the four models studied closely, CTDMPLUS predicted the lowest 24-hour average concentrations, but the concentrations estimated by Valley were lower than those estimated by CTDMPLUS.

• 한국조경학회지
• /
• 제31권6호
• /
• pp.53-63
• /
• 2004

Planting models for restoration forest on the seaside have been demanded because coastal reclaimed land has increased for habitation sites, industrial complexes and new towns on the west seaside of Korea. The planting models have to consider endurance for bad environmental conditions in order to make a role to protect the urban space against the extreme seaside environment. The dominant species, relative impotance value, individuals and species number were analysed in natural forests that were exposed to extreme seaside conditions in Deokjeok island and Younghung island, Incheon. The native species such as Pinus thunbergii and Pinus densiflora, which survive on the seaside, were mainly recommended because the coastal reclaimed land had extreme environmental conditions. Stable vegetation structures could be made by multi-layer planing by using these species. A diverse vegetation community could be made according to these planting models. The maritime forests made by these planting models might be more effective for environmental adaptation and a windbreak forest than alone tree, and the young trees below 3m height could easily adapt to these conditions.

• IMMUNE NETWORK
• /
• 제2권2호
• /
• pp.65-71
• /
• 2002

The immunological mechanism of the responses to ultraviolet (UV) B radiation in mouse models were investigated by the suppression of contact hypersensitivity (CHS) and delayed type hypersensitivity (DTH), and susceptibility to infection. However, there are some differences in immune suppression according to the different models as well as the irradiation protocols. Therefore, this review focused on the differences in the suppressive effects on CHS and DTH, and susceptibility to infection in relation to the different in vivo models. Recent advances in cytokine knockout mice experiments have the reexamination of the role of the critical cytokines in UVB-induced immune suppression, which was investigated previously by blocking antibodies. The characteristics of the suppressor cells responsible for UVB-induced tolerance were determined. The subcellular mechanism of UVB-induced immune suppression was also explained by the induction of apoptotic cells through the Fas and Fas-ligand interaction. The phagocytosis of the apoptotic cells is believed to induce the production of the immune suppressive cytokine like interleukin-10 by macrophages. Therefore, the therapeutic UVB response to a skin disease, such as psoriasis, by the depletion of infiltrating T cells could be considered in the extension line of apoptosis and immune suppression.

• 한국자동차공학회논문집
• /
• 제6권3호
• /
• pp.34-44
• /
• 1998

It is clear that when an automobile negotiates a curve the lateral acceleration causes an increase in tire normal load for the wheels on the outside of the curve and a decrease in load for the inside wheels. However, just how the details of the suspension linkages and the parameters of the springs and shock absorbers affect the dynamics of the load transfer os not easily understood. One even encounters the false idea that since it is the compression and extension of the main suspension springs spring body role which largely determines the changes in normal load, of roll could be reduced, the load transfer would also be reduced. Using free body diagrams, one can explain quite clearly how the load is transferred for steady state cornering, and, using complex multibody models of particular vehicles one can simulate in good fidelity how load transfer occurs dynamically. Here we adopt a middle ground by using the concept of roll center and using a series of half-car bond graph models to point out main effects. Since bond graph junction structures automatically and consistently constrain geometric and force variables simultaneously, they can be used to point out hidden assumptions of other simplified vehicle models.

• 약학회지
• /
• 제38권6호
• /
• pp.733-741
• /
• 1994

To detect anxiolytic activity from Sanguisorbae Radix we used various animal models of fear or anxiety that are sensitive to known anxiolytic drugs. While diazepam showed significant anxiolytic activities in all five animal models empolyed in this study, $5-HT_3$ antagonist ondansetron and ethylacetate fraction of Sanguisorbae Radix did show anti-anxiety effects in social interaction and two compartment exploration tests. Ethylacetate fraction of Sanguisorbae Radix and 5-HT related drugs like ondansetron and buspirone, however, seem to have merit over diazepam in terms of not causing drowsiness. Among ten subfractions obtained from ethylacetate fraction of Sanguisorbae Radix by silica gel chromatography, subfraction I showed higher anxiolytic activities than subfraction DEF in two animal models, social interaction and two compartment exploration tests. There is growing evidence for the role of 5-HT in the control of anxiety. We hope that new compound(s) will be found from the active fractions of Sanguisorbae Radix as a potential anxiolytic agent in the future.