• Journal of the Korea Society of Computer and Information
• /
• v.20 no.11
• /
• pp.119-126
• /
• 2015

CDSS provides clinical doctors with knowledge to be required when they diagnose or make decision about treatment strategy. Arden Syntax is one of the language with which we write MLM that is a component of CDSS. It was designated as a standard by HL7/ANSI. ArdenML is an XML version of Arden Syntax. In this paper we propose a tool which translates Arden Syntax MLMs into ArdenML MLM. To this end we first defines the corresponding relation between two languages. Next we presents a modified version of Arden Syntax grammar to improve performance of lexical analysis and minimize parsing conflicts. Finally we presents syntax and semantics gaps between the both languages, which are a structural representation problem, a data type problem, and a disrelation problem. Our translator resolves such issues and generates exact ArdenML codes for an arbitrary Arden Syntax MLM.

• Annual Conference on Human and Language Technology
• /
• 2020.10a
• /
• pp.9-13
• /
• 2020

관계 추출은 주어진 문장이나 문서에 존재하는 개체들 간의 의미적 관계를 찾아내는 작업을 말한다. 최근 문서 수준 관계 추출 말뭉치인 DocRED가 공개되면서 문서 수준 관계 추출에 대한 연구가 활발히 진행되고 있다. 또한 사전 학습된 Masked Language Model(MLM)이 자연어처리 분야 전체에 영향력을 보이면서 관계 추출에서도 MLM을 사용하는 연구가 진행되고 있다. 그러나 문서 수준의 관계 추출은 문서의 단위가 길기 때문에 Self-attention을 기반으로 하는 MLM을 사용하면 모델의 계산량이 증가하는 문제가 있다. 본 논문은 이 점을 보완하기 위해 관계 추출에 필요한 문장을 선별하는 간단한 전처리 방법을 제안한다. 또한 문서의 길이에 상관없이 관계 추출에 필요한 어휘 정보를 자동으로 습득 할 수 있는 Relation-Context Co-attention 방법을 제안한다. 제안 모델은 DocRED 말뭉치에서 Dev F1 62.01%, Test F1 59.90%로 높은 성능을 보였다.

• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.327-332
• /
• 2023

데이터의 품질과 다양성은 모델 성능에 지대한 영향을 끼친다. 본 연구에서는 Topic을 활용한 데이터 전처리와 BERT 기반 MLM, T5, Random Masking을 이용한 증강으로 데이터의 품질과 다양성을 높이고자 했으며, 이를 KoBERT 기반 에세이 자동 평가 모델에 적용했다. 데이터 전처리만 진행했을 때, Quadratic Weighted Kappa Score(QWK)를 기준으로 모델이 에세이의 모든 평가 항목에 대해 베이스라인보다 더욱 높은 일치도를 보였으며 평가항목별 일치도의 평균을 기준으로 0.5368029에서 0.5483064(+0.0115035)로 상승했다. 여기에 제안하는 증강 방식을 추가 할 경우 MLM, T5, Random Masking 모두 성능 향상 효과를 보였다. 특히, MLM 데이터 증강 방식을 추가로 적용하였을 때 최종적으로 0.5483064에서 0.55151645(+0.00321005)으로 상승해 가장 높은 일치도를 보였으며, 에세이 총점으로 QWK를 기준으로 성능을 평가하면 베이스라인 대비 0.4110809에서 0.4380132(+0.0269323)로의 성능 개선이 있었다.

• Journal of Korean Clinical Nursing Research
• /
• v.18 no.1
• /
• pp.126-135
• /
• 2012

Purpose: This study was aimed to develop a simulation training program of an advanced cardiac life support (ACLS) based on the mastery learning model (Simulation-MLM), and evaluate the effects of the program on critical care nurses. Methods: As an experimental pre-post test with a non-equivalent control group, the study employed convenience sampling of 38 critical care nurses. The experimental group received the Simulation-MLM including a theoretical lecture, formative evaluation, and simulation training, whereas only a theoretical lecture for the control group. The knowledge, self-efficacy, and performance degrees of respondents were measured to verify the effects of the Simulation-MLM. The statistical processing of the collected data utilized the SPSS WIN 17.0 program. Results: After receiving Simulation-MLM, the participants in the experimental group reported higher marks in the knowledge, self-efficacy and performance of ACLS compared with those in the control group. However, both experimental and control groups demonstrated no significant differences in knowledge, self-efficacy and performance. Conclusion: Despite of the limitation of a small sample size, this study was considered meaningful in a sense that it showed a venue for improving ACLS training efficiency. Future research with more distinct treatment differentiation and better adequate outcome variables was warranted in order to prove the effects of a theory-based simulation education.

• Journal of Distribution Research
• /
• v.15 no.3
• /
• pp.35-69
• /
• 2010

Successful management of distributors is crucial for the success of multi-level marketing (MLM) because MLM sells products and services through distributor networks. The present study examines the impact of the up-line distributors' transformational leadership on down-line distributors' entrepreneurship, market orientation, and performance. The proposed research questions were examined based on a survey of 135 down-line distributors working at the largest domestic MLM company. A mediation mechanism was found that the up-line distributors' transformational leadership increases the down-line distributors' entrepreneurship and market orientation, which in turn influences their performance. Based on the findings, the present study suggests the importance of transformational leadership in MLM. Practical implications for the roles of transformational leadership and its impact on corporate management in specific business areas are also discussed.

• Journal of Distribution Science
• /
• v.17 no.3
• /
• pp.5-19
• /
• 2019

Purpose - The purpose of this research is to investigate the factors contributing to sustainable competitive advantage for multi-level marketing (MLM) firms in Malaysia. The selected variables in this study are company image, product innovation, leadership, distributor rewards system and distributor training system. Research design, data, and methodology - Quantitative research method is employed with collected sample size of 398 respondents using judgmental sampling technique. Normality and reliability test were performed in the first stage utilizing SPSS 22 and Confirmatory Factory Analysis (CFA) and variance analysis were obtained in the subsequent stage, following up with the overall fit of the measurement model, Structural Equation Model (SEM) using AMOS 22 with maximum likelihood estimation to assess the internal consistency, convergent validity and discriminant validity. Results - The research findings show that company image, leadership, distributor rewards system and distributor training system were supported and are factors affecting the sustainable competitive advantage of MLM companies in Malaysia. However, in this study, product innovation was not supported but this result does not depict that it is trivial and inconsequential in maintain sustainable advantage. Conclusion - Companies can build sustainable competitive advantage by focusing on these contributing factors. Several other comments and implications were brought to light and discussed in the paper.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2000.11a
• /
• pp.249-255
• /
• 2000

The design and test of an magnetostrictive linear motor(MLM) that operates based on self-moving cell concept is presented. The moving cell is composed of Terfenol-D linear actuator and a ring structure, and a cell train is constructed by connecting two cells in series. Since this motor uses the stroke of Terfenol-D actuators and friction force of the cells, it can essentially produce long stroke and large force. The overall performance of the MLM was measured in terms of speed and force. The pushing force is directly related with the friction force. This work is a proof-of-concept stage and investigation is necessary for realistic application.

• Annual Conference on Human and Language Technology
• /
• 2021.10a
• /
• pp.486-489
• /
• 2021

본 논문에서는 KorBERT 한국어 언어모델에 기반하여 텍스트 분류문제를 빈칸채우기 문제로 변환하고 빈칸에 적합한 어휘를 예측하는 방식의 프롬프트기반 분류모델에 대해서 소개한다. [CLS] 토큰을 이용한 헤드기반 분류와 프롬프트기반 분류는 사전학습의 NSP모델과 MLM모델의 특성을 반영한 것으로, 텍스트의 의미/구조적 분석과 의미적 추론으로 구분되는 텍스트 분류 태스크에서의 성능을 비교 평가하였다. 의미/구조적 분석 실험을 위해 KLUE의 의미유사도와 토픽분류 데이터셋을 이용하였고, 의미적 추론 실험을 위해서 KLUE의 자연어추론 데이터셋을 이용하였다. 실험을 통해, MLM모델의 특성을 반영한 프롬프트기반 텍스트 분류에서는 의미유사도와 토픽분류 태스크에서 우수한 성능을 보였고, NSP모델의 특성을 반영한 헤드기반 텍스트 분류에서는 자연어추론 태스크에서 우수한 성능을 보였다.

• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.113-118
• /
• 2023

딥러닝을 이용한 언어 모델은 다양한 분야에서 사용되고 있는데, 그 중 교육 분야에선 꾸준히 시험 문항을 자동으로 생성하려는 요구가 존재해 왔다. 그러나 빈칸 추론 문항, 그 중에서도 어구 단위 빈칸 추론 문항은 학습 및 평가 목적으로 널리 쓰이고 있지만, 이를 자동 생성하려는 연구는 상대적으로 드물다. 이에 본 연구에선 masked language modeling (MLM)을 이용한 난이도 조절이 가능한 어구 단위 빈칸 추론 문항 생성 시스템을 제안한다. 본 시스템은 정답 생성 모델의 attention 정보에 따라 지문 내 중요한 어구를 삭제해 오답을 생성하고, 동시에 어구의 삭제 비율을 조절함으로써 더 쉽거나 더 어려운 오답을 만들어낼 수 있다. 평가 결과, 제안한 시스템은 기존 접근법보다 정답과의 유사도가 최고 28.3% 낮았고, 또한 난이도 설정에 따라 쉬운 오답이 어려운 오답에 비해 유사도가 15.1% 낮아, 더 정답과 먼 뜻의 오답을 생성해내었다.

• Proceedings of the Materials Research Society of Korea Conference
• /
• 1998.11a
• /
• pp.14-14
• /
• 1998