• Annual Conference on Human and Language Technology
• /
• 2019.10a
• /
• pp.460-464
• /
• 2019

번역 품질 예측은 기계번역 시스템이 생성한 번역문의 품질을 정답 번역문을 참고하지 않고 예측하는 과정으로, 번역문의 사후 교정을 위한 번역 오류 검출의 역할을 담당하는 중요한 연구이다. 본 논문은 문장 수준의 번역 품질 예측 문제를 HTER 구간의 분류 문제로 간주하여, 번역 품질 말뭉치의 HTER 분포 불균형으로 인한 성능 제약을 완화하기 위해 인조 사후 교정 말뭉치를 이용하는 방법을 제안하였다. 결과적으로 HTER 분포를 균등하게 조정한 학습 말뭉치가 그렇지 않은 쪽에 비해 번역 품질 예측에 더 효과적인 것을 보였다.

• KIISE Transactions on Computing Practices
• /
• v.20 no.12
• /
• pp.700-705
• /
• 2014

An anagram is a form of word play to find a new word from a set of given alphabet letters. Good human anagram solvers use the strategy of bigrams. They explore a constraint satisfaction network in parallel and answers consequently pop out quickly. In this paper, we propose a molecular computational algorithm using the same process as this. We encoded letters into DNA sequences and made bigrams and then words by connecting the letter sequences. From letters and bigrams, we performed DNA hybridization, ligation, gel electrophoresis and finally, extraction and separation to extract bigrams. From the matched bigrams and words, we performed the four molecular operations again to distinguish between right and wrong results. Experimental results show that our molecular computer can identify cor rect answers and incorrect answers. Our work shows a new possibility for modeling the cognitive and parallel thinking process of a human.

• Phonetics and Speech Sciences
• /
• v.13 no.2
• /
• pp.27-35
• /
• 2021

This study investigated the status of the wh-island constraint in Seoul Korean. The syntactic movement of a wh-phrase out of an embedded sentence so as to have wide scope at LF is known to be invalid as it violates the wh-island constraint, but there have been studies in which such a movement was possible when the sentence is read with a wh-intonation. We conducted perceptual tests in which subjects were asked to select an answer after listening to each of the four types of interrogative sentences. Three of them were with 'Nugu-leul', which is an accusative form of the wh-phrase 'who' as well as an indefinite form. The fourth sentence contained the name of a person. 'Nugu-leul' and the noun were positioned in the same embedded sentence to see whether the subjects accepted a matrix scope interpretation of the wh-phrases. Response time was transformed to normalized log response time and checked to find any differences in the time taken to select the answers depending on different types of interrogative sentences. The results showed the subjects had a definite preference for the matrix scope interpretation for the sentences with a wh-intonation. The response time required to select the matrix scope interpretation was longer than for any other type of interrogative sentence. We concluded that the wh-island constraint in Seoul Korean is weak.

• Journal of the Korea Society of Computer and Information
• /
• v.19 no.1
• /
• pp.191-201
• /
• 2014

The purpose of this study is to propose strategies of puzzle-based learning for Informatics gifted education through analyzing Informatics gifted elementary students' computational problem solving approaches in puzzle-based learning contexts. Six types of educational puzzles, which are constraints, optimization, probability, statistically speaking, pattern recognition, and strategy, were used in teaching 14 Informatics gifted students for 8 sessions. The results of pre and post test and each students' answers were analyzed to identify why students were not able to solve the puzzles. We also analysed what essential computational strategies are needed to solve each type of puzzles, and what students did not know in solving puzzle problems. We identified some problems caused by puzzle representation methods, and various students' intuitions that disturb puzzle solving. Also, we identified essential computational strategies to solve puzzles: backtracking, dynamic programming, abstraction, modeling, and reduction of big problem. However, students had difficulties in applying these strategies to solve their puzzle problems. We proposed the revised puzzle-based learning strategies, which is based on the improved problem representation, just-in-time cognitive feedbacks, and web-based learning system.

• Journal of the Korea Society of Computer and Information
• /
• v.13 no.6
• /
• pp.69-76
• /
• 2008

The biggest strength of the Internet is to enable to access information without any limitation of time and space. As the Internet and IT technologies have been developed, various kinds of teaching ways in education field such as remote lectures, video lectures, and CAI(Computer Adapted Instruction) have emerged. In terms of education, evaluation can be a basic foundation to help teach students in the next learning stage according to each student's level. In addition, it is able to give the information of students'abilities and provides proper learning programs to teach students on a case-by-case basis. The purpose of the paper is to establish evaluation system on the WWW(World Wide Web) that can reflect learning activities part of students in their evaluation scores based on the two important learning theories, Behaviorism and constructivism, which are mainly used in evaluation procedures to judge learning ability of students. This system will give information about learners, and analyze the learning interest of learners. The proposed system enables teachers to evaluate learning ability of students through various kinds of information of learners, and to execute level-based education.

• Journal of Intelligence and Information Systems
• /
• v.20 no.1
• /
• pp.15-34
• /
• 2014

Overall, accuracy as a performance measure does not fully consider modular accuracy: the accuracy of classifying 1 (or true) as 1 is not same as classifying 0 (or false) as 0. A smarter classification algorithm would optimize the classification rules to match the modular accuracies' goals according to the nature of problem. Correspondingly, smarter algorithms must be both more generalized with respect to the nature of problems, and free from decretization, which may cause distortion of the real performance. Hence, in this paper, we propose a novel vertical boosting algorithm that improves modular accuracies. Rather than decretizing items, we use simple classifiers such as a regression model that accepts continuous data types. To improve the generalization, and to select a classification model that is well-suited to the nature of the problem domain, we developed a model selection algorithm with smartness. To show the soundness of the proposed method, we performed an experiment with a real-world application: predicting the intellectual properties of e-transaction technology, which had a 47,000+ record data set.

• 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 KIISE:Software and Applications
• /
• v.33 no.12
• /
• pp.1073-1089
• /
• 2006

The goal of Natural Language Processing(NLP) is to make a computer understand a natural language and to deliver the meanings of natural language to humans. Word sense Disambiguation(WSD is a very important technology to achieve the goal of NLP. In this paper, we describe a technology for automatic homonyms disambiguation using both Mutual Information(MI) and a Sense-Tagged Compound Noun Dictionary. Previous research work using word definitions in dictionary suffered from the problem of data sparseness because of the use of exact word matching. Our work overcomes this problem by using MI which is an association measure between words. To reflect language features, the rate of word-pairs with MI values, sense frequency and site of word definitions are used as weights in our system. We constructed a Sense-Tagged Compound Noun Dictionary for high frequency compound nouns and used it to resolve homonym sense disambiguation. Experimental data for testing and evaluating our system is constructed from QA(Question Answering) test data which consisted of about 200 query sentences and answer paragraphs. We performed 4 types of experiments. In case of being used only MI, the result of experiment showed a precision of 65.06%. When we used the weighted values, we achieved a precision of 85.35% and when we used the Sense-Tagged Compound Noun Dictionary, we achieved a precision of 88.82%, respectively.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2015.05a
• /
• pp.277-280
• /
• 2015

Smishing, Phishing personal privacy caused by Incident accidents such as Phishing information security has become a hot topic. Such incidents have privacy in personal information management occurs due to a lack of user awareness. This paper is based on the existing structure of the XML Tag question bank used a different Key-Value Structure-based JSON. JSON is an advantage that does not depend on the language in the text-based interchange format. The proposed system is divided into information security sector High, Middle and Low grade. and Provides service to the user through the free space and the smart device and the PC to the constraints of time. The use of open source Apache Load Balancing technology for reliable service. It also handles the user's web page without any training sessions Require server verification result of the training(training server). The result is sent to the training server using jQuery Ajax. and The resulting data are stored in the database based on the user ID. Also to be used as a training statistical indicators. In this paper, we design a level training system to enhance the user's information security awareness.

• Journal of The Korean Association For Science Education
• /
• v.38 no.4
• /
• pp.541-553
• /
• 2018

The purpose of this study is to explore the relationship between epistemic goals, epistemic considerations and complexity of reasoning of science high school students in an open inquiry and to explore the context on how open inquiry compares with the characteristics of an authentic scientific inquiry. Two teams were selected as focus groups and a case study was conducted. The findings are as follows: First, the contexts, such as 'sharing the value for the phenomenon understanding, reflection on the value of the research, task characteristics that require collaboration and consensus, and sufficient communication opportunities,' promote epistemic goals and considerations. On the other hand, contexts such as 'lack of opportunity for critical review of related literature and environmental constraints' lowered epistemic sides. Second, epistemic goals and considerations influenced the reasoning complexity. The goal of 'scientific sense making' led to reasoning that pose testable hypotheses based on students' own questions. The high justification considerations led to purposely focusing attention to the control designs and developing creative experimental know-how. The high audience considerations led to defending their findings through argumentation and suggesting future research. On the other hand, the goal of 'doing the lesson' and the low justification considerations led to reasoning that did not interpret the meaning of the data and did not control the limit of experiment. The low audience considerations led to reasoning that did not actively defend their findings and not suggest future research. The results of this study suggest that guidance should provide communication and critical review opportunities.