• KIPS Transactions on Software and Data Engineering
• /
• v.7 no.11
• /
• pp.411-418
• /
• 2018

With the proliferation of artificial intelligence technology, it is becoming important to obtain, store, and utilize scientific data in research and science sectors. A number of methods for extracting meaningful objects such as graphs and tables from research articles have been proposed to eventually obtain scientific data. Existing extraction methods using heuristic approaches are hardly applicable to electronic documents having heterogeneous manuscript formats because they are designed to work properly for some targeted manuscripts. This paper proposes a prototype of an object extraction system which exploits a recent deep-learning technology so as to overcome the inflexibility of the heuristic approaches. We implemented our trained model, based on the Faster R-CNN algorithm, using the Google TensorFlow Object Detection API and also composed an annotated data set from 100 research articles for training and evaluation. Finally, a performance evaluation shows that the proposed system outperforms a comparator adopting heuristic approaches by 5.2%.

• Annual Conference on Human and Language Technology
• /
• 2022.10a
• /
• pp.125-130
• /
• 2022

최근 자연어처리 연구로 지식 기반 대화에서 대화 내용에 자유로운 주제와 다양한 지식을 포함하는 연구가 활발히 이루어지고 있다. 지식 기반 대화는 대화 내용이 주어질 때 특정 지식 정보를 포함하여 이어질 응답을 생성한다. 이때 대화에 필요한 지식이 검색 가능하여 선택에 제약이 없는 오픈 도메인(Open-domain) 지식 기반 대화가 가능하도록 한다. 오픈 도메인 지식 기반 대화의 성능 향상을 위해서는 대화에 이어지는 자연스러운 답변을 연속적으로 생성하는 응답 생성 모델의 성능 뿐만 아니라, 내용에 어울리는 응답이 생성될 수 있도록 적합한 지식을 선택하는 지식 검색 모델의 성능 향상도 매우 중요하다. 본 논문에서는 오픈 도메인 지식 기반 한국어 대화에서 지식 검색 성능을 높이기 위해 밀집 벡터 기반 검색 방식과 주제어(Keyword) 기반의 검색 방식을 함께 사용하는 것을 제안하였다. 먼저 밀집 벡터 기반의 검색 모델을 학습하고 학습된 모델로부터 고난도 부정(Hard negative) 지식 후보를 생성하고 주제어 기반 검색 방식으로 고난도 부정 지식 후보를 생성하여 각각 밀집 벡터 기반의 검색 모델을 학습하였다. 성능을 측정하기 위해 전체 지식 중에서 하나의 지식을 검색했을 때 정답 지식인 경우를 계산하였고 고난도 부정 지식 후보로 학습한 주제어 기반 검색 모델의 성능이 6.175%로 가장 높은 것을 확인하였다.

• Journal of Practical Engineering Education
• /
• v.11 no.1
• /
• pp.25-31
• /
• 2019

The purpose of this study is to investigate the difference between motivation and learning strategies of students who have received academic probation and who have not while attending engineering college. The subjects of this study were engineering college students in Seoul and they participated in the learning strategy test at the teaching and learning center. The subjects of this study were 553 students, 22 of whom received academic probation and 531 didn't. In order to achieve the purpose of this study, we used MLST (Multi-dimensional Learning Strategy test) learning strategy checklists of Korea Guidance, which is a standardized test. A t-test was conducted to compare motivational and learning strategies between students with and without academic probation. As a result, the motivation score of the students with the academic probation was lower than that of those without the academic probation, and the score of the time management and note taking factors of the students with the academic probation were lower than those of the students without the academic probation.

• Journal of the Korean Society for information Management
• /
• v.35 no.4
• /
• pp.141-164
• /
• 2018

Recently, as the amount of academic literature has increased rapidly and complex researches have been actively conducted, researchers have difficulty in analyzing trends in previous research. In order to solve this problem, it is necessary to classify information in units of academic papers. However, in Korea, there is no academic database in which such information is provided. In this paper, we propose an automatic classification system that can classify domestic academic literature into multiple classes. To this end, first, academic documents in the technical science field described in Korean were collected and mapped according to class 600 of the DDC by using K-Means clustering technique to construct a learning set capable of multiple classification. As a result of the construction of the training set, 63,915 documents in the Korean technical science field were established except for the values in which metadata does not exist. Using this training set, we implemented and learned the automatic classification engine of academic documents based on deep learning. Experimental results obtained by hand-built experimental set-up showed 78.32% accuracy and 72.45% F1 performance for multiple classification.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2012.11a
• /
• pp.1194-1197
• /
• 2012

본 논문에서는 가상세계를 이용한 직업체험 콘텐츠를 통해 사용자가 직업에 대한 이해와 관련 업무의 지식 및 프로세스의 습득, 작성 및 흥미도 검사 등 직업에 대한 전반적인 정보 제공 및 시뮬레이션을 통한 학습의 몰입감 실제감 있는 교육 환경으로 기능적 목적 충족은 물론 재미적 요소를 가미하여 더욱 흥미있고 효과적인 직업체험을 할 수 있도록 콘텐츠를 구축하였다. 본 논문에서는 영어를 주로 사용하는 직업을 선정하여 직업 수행을 주제로 역할놀이(Role play)를 통해 직업 체험을 이용할 수 있도록 영어 강사 직업 체험 콘텐츠, 공항직원/승무원 직업 체험콘텐츠, 레스토랑 종업원 직업 체험 콘텐츠, 호텔 종업원 직업 체험 콘텐츠로 총 4개의 콘텐츠 서비스를 구성하였다.

• 대한공업교육학회지
• /
• v.40 no.1
• /
• pp.87-104
• /
• 2015

This study was carried out to evaluate on the job implementation of ITEC(Invention Teacher Education Center)'s in-service teacher training. To accomplish this purpose, teacher's perception on retention, usefulness, utilization, and enhancement were investigated after one year completing in-service teacher training. The populations of this study were 60 teachers completing in-service teacher training of the southeast ITEC. The methods of this study were survey research and qualitative content analysis. The subjects of survey research were 30 teachers and the case analysis objects were 12 epilogues of in-service teacher training. The mean of job implementation of southeast ITEC's in-service teacher training was 4.07 on a five-point scale. Teacher's perception on job implementation sub factors after one year completing in-service teacher training were as follows. The mean of enhancement factor was 4.39, usefulness factor was 4.11, utilization factor was 3.94, and retention factor was 3.82 on a five-point scale. The mean of enhancement factor and usefulness factor were relatively high level as compared with the mean of job implementation and the mean of utilization factor and retention factor were relatively low level as compared with the mean of job implementation. The results of qualitative content analysis about 12 epilogues of in-service teacher training, some significant instances were deducted that representable retention, usefulness, utilization, and enhancement factors directly and indirectly.

• Annual Conference on Human and Language Technology
• /
• 2022.10a
• /
• pp.473-476
• /
• 2022

문장 임베딩은 문장의 의미를 고려하여 모델이 적절하게 의미적인 벡터 공간에 표상하는 것이다. 문장 임베딩을 위해 다양한 방법들이 제안되었지만, 최근 가장 높은 성능을 보이는 방법은 대조적 학습 방법이다. 대조적 학습을 이용한 문장 임베딩은 문장의 의미가 의미적으로 유사하면 가까운 공간에 배치하고, 그렇지 않으면 멀게 배치하도록 학습하는 방법이다. 이러한 대조적 학습은 비지도와 지도 학습 방법이 존재하는데, 본 논문에서는 효과적인 비지도 학습방법을 제안한다. 기존의 비지도 학습 방법은 문장 표현을 학습하는 언어모델이 자체적인 정보를 활용하여 문장의 의미를 구별한다. 그러나, 하나의 모델이 판단하는 정보로만 문장 표현을 학습하는 것은 편향적으로 학습될 수 있기 때문에 한계가 존재한다. 따라서 본 논문에서는 Cross-Encoder의 Re-Ranker를 통한 의미 검색으로부터 추천된 문장 쌍을 학습하여 기존 모델의 성능을 개선한다. 결과적으로, STS 테스크에서 베이스라인보다 2% 정도 더 높은 성능을 보여준다.

• Journal of Korea Entertainment Industry Association
• /
• v.15 no.8
• /
• pp.269-278
• /
• 2021

This study was conducted to develop a learning competency scales customized for learning programs conducted by Center for Teaching & Learning at A University. To achieve this purpose, a preliminary study was set up, which consists of three competency groups (basic competency, intensity competency, application competency) and 13 learning competency factors through a review of previous studies. In order to verify the reliability and validity of the provisional learning competency scales, an online survey was conducted on A university students in September 2020, The collected questionnaire data were organized and exploratory factor analysis and confirmatory factor analysis were conducted. As a result of exploratory factor analysis, 13 learning competency was reduced to 10 as the three competency groups were maintained. As a result of the confirmatory factor analysis, the model was found to be good, Also, as a result of analyzing the reliability of the confirmed learning competency factors, all 10 factors showed a good level of .7 or more. The learning competency scales developed through this study can be used as basic data for performance evaluation and development of new programs of CTL learning program.

• Journal of the Korean Society of Earth Science Education
• /
• v.12 no.1
• /
• pp.82-93
• /
• 2019

The purpose of this study is to analyze the results of satisfaction for learning eco-system on the teacher training program conducted at the IS-Geo (International School for Geoscience Resources) which is KIGAM (Korea Institute of Geoscience and Mineral Resources), and to determine the satisfaction and educational effects of the teacher training programs on elementary and secondary teachers. And then, to suggest improvement points in the future operation of the teacher training program at the IS-Geo. Therefore, we conducted questionnaire of satisfaction for learning eco-system based on the data collected by a survey of 98 elementary and secondary teachers who participated in the teacher training program at the IS-Geo, from July 2017 to August 2018. The research results are as follows. First, the results of satisfaction for learning eco-system showed high values of 4.58 or higher in both the elementary and secondary programs, and the teacher training program conducted by the IS-Geo had a positive effect on the training participants. Second, internal factors indicating learning motivation and learning development were elementary teacher training 4.70 and secondary teacher training 4.64, and it is necessary to develop training contents and programs by classifying them into majors other than the earth science department. Third, intermediate factors indicating contents of education and learning curriculum were 4.67 for an elementary teacher training program and 4.72 for secondary teacher training program. In addition, in order to operate the teacher training program according to the purpose of science and technology culture, it is necessary to develop a teaching-learning model and to improve the quality of teaching. Fourth, external factors indicating learner support and quality of instructors were 4.83 for an elementary teacher training program and 4.72 for a secondary teacher training program. In particular, it is necessary to develop teaching materials that can be used immediately in school classes and can generate interest.

• Journal of Convergence for Information Technology
• /
• v.8 no.6
• /
• pp.281-285
• /
• 2018

The purpose of this research is to make a Flipped-MOOC model which can be applied and practiced in the college course after analyzing the characteristics and cases of MOOC and Flipped learning. For this, this study implemented the following tasks. First, this study analyzed the management and class types of MOOC and flipped learning through literature research. Secondly, flipped learning was applied in the course for a semester and the strong point and weak point of the course was analyzed and the alternative was suggested. Thirdly, the core ideas and strategies of Flipped-MOOC model was deducted for enhancing the participation and interaction of the students in the course which uses the MOOC content and applies flipped learning, and the instructional strategies and direction for the effective management in the real educational field was suggested. As a result, Flipped-MOOC model is expected to contribute for the educational revolution, change and quality improvement, and it is expected that Flipped-MOOC model might contribute to the lifelong education and educational competitiveness.