• The Journal of Engineering Geology
• /
• v.32 no.3
• /
• pp.377-388
• /
• 2022

Recently, several studies have utilized machine learning to efficiently and accurately analyze seismic data that are exponentially increasing. In this study, we expand earthquake information such as occurrence time, hypocentral location, and magnitude to produce a dataset for applying to machine learning, reducing the dimension of the expended data into dominant features through principal component analysis. The dimensional extended data comprises statistics of the earthquake information from the Global Centroid Moment Tensor catalog containing 36,699 seismic events. We perform data preprocessing using standard and max-min scaling and extract dominant features with principal components analysis from the scaled dataset. The scaling methods significantly reduced the deviation of feature values caused by different units. Among them, the standard scaling method transforms the median of each feature with a smaller deviation than other scaling methods. The six principal components extracted from the non-scaled dataset explain 99% of the original data. The sixteen principal components from the datasets, which are applied with standardization or max-min scaling, reconstruct 98% of the original datasets. These results indicate that more principal components are needed to preserve original data information with even distributed feature values. We propose a data processing method for efficient and accurate machine learning model to analyze the relationship between seismic data and seismic behavior.

• Journal of the Korean association of regional geographers
• /
• v.9 no.2
• /
• pp.153-168
• /
• 2003

Since 1945, geography textbooks of South Korea and North Korea have been developed differently. Taking both geography textbooks of South Korea(printed in 1996 and 1997) and them of North Korea(printed in 1995), this study compares and analyse geography textbooks of South and North Korea systematically, then to show their likeness and difference. Results of this study are as follows: First, Most of contents in South Korean textbooks consist of explations and many tables, graphs and photographs leading to inquiry activities, whereas those in North Korean textbooks consist of mostly sentences that explain topics, diagrams and sketches. Second, Geographical education in South Korea puts strong emphasis on human geography and regional geography, whereas that in North Korea does strong emphasis on physical geography, economic geography and education of cartography. Third, The geographical words of South Korean textbooks contain words written in Chinese character and words of English origin, whereas Those of North Korean textbooks do many pure Korean words.

• The Journal of Korean Association of Computer Education
• /
• v.4 no.2
• /
• pp.125-133
• /
• 2001

With development of computer engineering and the examination theories, new techniques to use computers in making or administrating examinations had been studied and developed. Recently, studies on web-based testing are in progress with development of Internet and information and communication technology. However, as compared with development of these technical aspects, studies on the user legibility to provide learners with comfortable environment for the evaluation have been insufficient in fact. In the web-based testing, user legibility decreases the anxiety of examinees in the examination and plays an important role in measuring their competence exactly. In this study, we tried to provide the optimum web-based evaluation interface by obtaining information through a questionnaire for primary-school students. We think that principles suggested in this study can give a great help in making the web-based evaluation screen in future. Besides, studies on more various fields including making individual questions, instructing sentences and layout of multimedia factors as well as the principles mentioned in this study should be performed.

• Journal of Practical Engineering Education
• /
• v.7 no.2
• /
• pp.113-117
• /
• 2015

In this academic paper, we introduce a bread board simulation which is developed based on HTML language. this program is developed because actual electronic practices have some problems. Electronic circuit practice using bread board is widely practiced in colleges or high schools in South Korea, but actual electronic circuit practice needs lots of consumable stuff, and they are needed to be purchased and managed continuously. Besides, composing complicated electronic circuit restricts the efficiency of studying. Therefore, we propose breadboard simulation program which helps to resolve those problems.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2023.07a
• /
• pp.541-544
• /
• 2023

LLM(Largescale Language Model)의 성능 향상을 위한 비용 효율적인 방법으로 ChatGPT, GPT-4와 같은 초거대 모델의 output에 대해 SLM(Small Language Model)을 finetune하는 방법이 주목받고 있다. 그러나, 이러한 접근법은 주로 범용적인 지시사항 모델을 위한 학습 방법으로 사용되며, 제한된 특정 도메인에서는 추가적인 성능 개선의 여지가 있다. 본 연구는 특정 도메인(Writing Assistant)에서의 성능 향상을 위한 새로운 방법인 Self-Guided Approach를 제안한다. Self-Guided Approach는 (1) LLM을 활용해 시드 데이터에 대해 도메인 특화된 metric(유용성, 관련성, 정확성, 세부사항의 수준별) 점수를 매기고, (2) 점수가 매겨진 데이터와 점수가 매겨지지 않은 데이터를 모두 활용하여 supervised 방식으로 SLM을 미세 조정한다. Vicuna에서 제안된 평가 방법인, GPT-4를 활용한 자동평가 프레임워크를 사용하여 Self-Guided Approach로 학습된 SLM의 성능을 평가하였다. 평가 결과 Self-Guided Approach가 Self-instruct, alpaca와 같이, 생성된 instruction 데이터에 튜닝하는 기존의 훈련 방법에 비해 성능이 향상됨을 확인했다. 다양한 스케일의 한국어 오픈 소스 LLM(Polyglot1.3B, PolyGlot3.8B, PolyGlot5.8B)에 대해서 Self-Guided Approach를 활용한 성능 개선을 확인했다. 평가는 GPT-4를 활용한 자동 평가를 진행했으며, Korean Novel Generation 도메인의 경우, 테스트 셋에서 4.547점에서 6.286점의 성능 향상이 발생했으며, Korean scenario Genration 도메인의 경우, 테스트 셋에서 4.038점에서 5.795 점의 성능 향상이 발생했으며, 다른 유사 도메인들에서도 비슷한 점수 향상을 확인했다. Self-Guided Approach의 활용을 통해 특정 도메인(Writing Assistant)에서의 SLM의 성능 개선 가능성을 확인했으며 이는 LLM에 비용부담을 크게 줄이면서도 제한된 도메인에서 성능을 유지하며, LLM을 활용한 응용 서비스에 있어 실질적인 도움을 제공할 수 있을 것으로 기대된다.

• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.448-453
• /
• 2023

대규모 언어 모델 (Large Language Model, LLM)을 인간의 선호도 관점에서 평가하는 것은 기존의 벤치마크 평가와는 다른 도전적인 과제이다. 이를 위해, 기존 연구들은 강력한 LLM을 평가자로 사용하여 접근하였지만, 높은 비용 문제가 부각되었다. 또한, 평가자로서 LLM이 사용하는 주관적인 점수 기준은 모호하여 평가 결과의 신뢰성을 저해하며, 단일 모델에 의한 평가 결과는 편향될 가능성이 있다. 본 논문에서는 엄격한 기준을 활용하여 편향되지 않은 평가를 수행할 수 있는 평가 프레임워크 및 평가자 모델 'FubaoLM'을 제안한다. 우리의 평가 프레임워크는 심층적인 평가 기준을 통해 다수의 강력한 한국어 LLM을 활용하여 연쇄적 사고(Chain-of-Thought) 기반 평가를 수행한다. 이러한 평가 결과를 다수결로 통합하여 편향되지 않은 평가 결과를 도출하며, 지시 조정 (instruction tuning)을 통해 FubaoLM은 다수의 LLM으로 부터 평가 지식을 증류받는다. 더 나아가 본 논문에서는 전문가 기반 평가 데이터셋을 구축하여 FubaoLM 효과성을 입증한다. 우리의 실험에서 앙상블된 FubaoLM은 GPT-3.5 대비 16% 에서 23% 향상된 절대 평가 성능을 가지며, 이항 평가에서 인간과 유사한 선호도 평가 결과를 도출한다. 이를 통해 FubaoLM은 비교적 적은 비용으로도 높은 신뢰성을 유지하며, 편향되지 않은 평가를 수행할 수 있음을 보인다.

• Asia-Pacific Journal of Business Venturing and Entrepreneurship
• /
• v.17 no.3
• /
• pp.201-214
• /
• 2022

This study theoretically and empirically examined whether organizational communication mediates the effect of organizational learning culture perceived by members in the organization on task performance and contextual performance. Organizational learning culture is defined as a culture that is good at creating, acquiring, transferring, and modifying behavior to reflect new knowledge and insights. The hypothesis of this study is that the perceived organizational learning culture can increase performance through organizational communication between members. In particular, we measured communication within the organization into three types: upward, horizontal, and downward. These communications were set as mediating variables. In empirical studies, independent variables were perceived organizational learning culture, mediation variables were upward, horizontal and downward communication, and dependent variables were task performance and contextual performance. Hypothesis 1 is that the organizational learning culture will have a positive effect on employees' tasks and contextual performance. Hypothesis 2 is about the mediating effect of communication on the relationship between Hypothesis 1. In the empirical study, after verifying the validity and reliability of the research variables, correlation analysis and hypothesis verification were conducted. Hypothesis 1 was verified through regression analysis, and all detailed hypotheses were supported. To verify Hypothesis 2, we conducted a bootstrap test using process macro to separate the total, direct, and indirect effects and examine the significance of the indirect effects. As a result, Hypothesis 2 was partially supported. Downward communication mediated organizational learning culture and task and contextual performance, and horizontal communication mediated organizational learning culture and contextual performance. The mediating effect of upward communication was not significant. The results of this study contributed to the suggestion of implications, research limitations, and research directions. Organizational learning culture is the direction and intention of the organization to achieve its goals through the learning and growth of its members. By strengthening internal motivation, organizational members can take voluntary desirable actions that help groups and organizations as well as essential tasks given. since this relationship appears as a medium of downward communication, organizations can strengthen the relationship between organizational learning culture and performance through leadership education.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.14 no.2
• /
• pp.433-438
• /
• 2019

This paper proposes a method of improving deep learning based numbers and characters recognition performance on parts of drawing through image preprocessing. The proposed character recognition system consists of image preprocessing and 7 layer deep learning model. Mathematical morphological filtering is used as preprocessing to remove the lines and shapes which causes false recognition of numbers and characters on parts drawing. Further.. Further, the used deep learning model is a 7 layer deep learning model instead of VGG-16 model. As a result of the proposed OCR method, the recognition rate of characters is 92.57% and the precision is 92.82%.

• Annual Conference on Human and Language Technology
• /
• 2023.10a
• /
• pp.627-632
• /
• 2023

대규모 언어 모델은 높은 연산 능력과 방대한 양의 데이터를 기반으로 탁월한 성능을 보이며 자연어처리 분야의 주목을 받고있다. 이러한 모델들은 다양한 언어와 도메인의 텍스트를 처리하는 능력을 갖추게 되었지만, 전체 학습 데이터 중에서 한국어 데이터의 비중은 여전히 미미하다. 결과적으로 이는 대규모 언어 모델이 영어와 같은 주요 언어들에 비해 한국어에 대한 이해와 처리 능력이 상대적으로 부족함을 의미한다. 본 논문은 이러한 문제점을 중심으로, 대규모 언어 모델의 한국어 처리 능력을 향상시키는 방법을 제안한다. 특히, Cross-lingual transfer learning 기법을 활용하여 모델이 다양한 언어에 대한 지식을 한국어로 전이시켜 성능을 향상시키는 방안을 탐구하였다. 이를 통해 모델은 기존의 다양한 언어에 대한 손실을 최소화 하면서도 한국어에 대한 처리 능력을 상당히 향상시켰다. 실험 결과, 해당 기법을 적용한 모델은 기존 모델 대비 nsmc데이터에서 2배 이상의 성능 향상을 보이며, 특히 복잡한 한국어 구조와 문맥 이해에서 큰 발전을 보였다. 이러한 연구는 대규모 언어 모델을 활용한 한국어 적용 향상에 기여할 것으로 기대 된다.

• Journal of The Korean Association of Information Education
• /
• v.12 no.1
• /
• pp.33-40
• /
• 2008

In social studies the ability of a student to investigate after deciding the subject and individual planning to search for society and to put investigation reports in order has been taken as highly important. The social studies performance evaluation methods used most frequently in education is also one of the investigation report method. However, when most students get the task "Write the report", they might write the report without any distinct object or organized structure. The evaluation result given by the teacher at the end of term is thus one sided. The recent performance evaluation is to keep the existing evaluation and let only the result have importance, not the essential meaning. So this paper is to build and materialize a web-based performance evaluation system, to be able to evaluate a course of classes. This paper's purposes was to write the reports in setting the standard form that the students research and collect the various data as based in the system and to improve the quality of evaluation and to inspire the study motivation as acting the mutuality action among the students and between them and teachers as giving a feedback at once after evaluating among colleagues by suggested evaluation views.