• The Transactions of the Korean Institute of Electrical Engineers B
• /
• v.54 no.12
• /
• pp.569-575
• /
• 2005

This paper proposes a gain scheduled control technique using time-delay for the nonlinear system with plant uncertainties and unexpected disturbances. The time delay-based gain scheduled control depends on a direct estimation of a function representing the effect of uncertainties. The information from the estimation is used to cancel the unknown dynamics and the unexpected disturbances simultaneously. The proposed estimation scheme with a finite convergence time is formulated in order to estimate the unknown scheduling variable variation. In other words, the time delay-based gain scheduled control uses the past observation of the system's response and the control input to directly modify the control actions rather than to adjust the controller gains or to identify system parameters. It has a simple structure so as to minimize the computational burden. The benefits of this proposed scheme are demonstrated in the simulation of an electromagnetic suspension system with plant uncertainties and external disturbances, and the proposed controller is compared with the conventional state feedback controller.

• Journal of Korean language education
• /
• v.29 no.4
• /
• pp.201-223
• /
• 2018

The purpose of this study is to observe the effects of text modification by comparing differences in Korean reading comprehension levels that arise from differences in vocabulary levels in texts. This study intends to use simplified texts with the vocabulary difficulty adjusted differently from the original text to measure reading comprehension levels of Korean learners and analyze the result. To measure reading comprehension, the researcher divided 55 Korean learners of intermediate to advanced level of fluency into two groups; the control group read the original text and the treatment group read a simplified text in which complex vocabulary were substituted with easier words of medium difficulty. Then the two groups were tested with the same questionnaire to measure comprehension levels of each group. The result showed that the groups that read simplified texts scored higher than the control group; this suggests that the reading comprehension level was increased in the treatment group. The experiment confirmed that unknown vocabulary density has direct impact on Korean reading comprehension. The result shows that the proportion of unknown vocabulary should be reduced for meaning-focused reading. It also demonstrates that comprehension of the learner was enhanced with lexical simplification rather than structural simplification i.e. simplification of grammar or sentences. Thus, diverse reading materials adjusted to the learners' level of fluency should be developed to enable reading for learning Korean. By reducing the burden of understanding the meaning of each vocabulary, learners will be able to achieve the initial goal of reading.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.30 no.4
• /
• pp.669-677
• /
• 2020

As the number of Internet users exploded, attacks on the web increased. In addition, the attack patterns have been diversified to bypass existing defense techniques. Traditional web firewalls are difficult to detect attacks of unknown patterns.Therefore, the method of detecting abnormal behavior by artificial intelligence has been studied as an alternative. Specifically, attempts have been made to apply natural language processing techniques because the type of script or query being exploited consists of text. However, because there are many unknown words in scripts and queries, natural language processing requires a different approach. In this paper, we propose a new classification model which uses byte pair encoding (BPE) technology to learn the embedding vector, that is often used for web attack payloads, and uses an attention mechanism-based Bi-GRU neural network to extract a set of tokens that learn their order and importance. For major web attacks such as SQL injection, cross-site scripting, and command injection attacks, the accuracy of the proposed classification method is about 0.9990 and its accuracy outperforms the model suggested in the previous study.

• Language and Information
• /
• v.7 no.1
• /
• pp.23-40
• /
• 2003

In this paper, we have proposed Auto-TUS, an automatic transcription module of three union symbols-hyphen, dash and tilde (‘­’, ‘―’, ‘∼’)-using their linguistic contexts. Few previous studies have discussed the problems of ambiguities in transcribing symbols into Korean alphabetic letters. We have classified six different reading formulae of the union symbols, analyzed the left and right contexts of the symbols, and investigated selection rules and distributions between the symbols and their contexts. Based on these linguistic features, 86 stereotyped patterns, 78 rules and 8 heuristics determining the types of reading formulae are suggested for Auto-TUS. This module works modularly in three steps. The pilot test was conducted with three test suites, which contains respectively 418, 987 and 1,014 clusters of words containing a union symbol. Encouraging results of 97.36%, 98.48%, 96.55% accuracy were obtained for three test suites. Our next phases are to develop a guessing routine for unknown contexts of the union symbols by using statistical information; to refine the proper nouns and terminology detecting module; and to apply Auto-TUS on a larger scale.

• Journal of Environmental Science International
• /
• v.5 no.3
• /
• pp.387-398
• /
• 1996

Granular sludge formation and it's activity change are the most important factors in achieving successful start-up and operation of UASB reactor. Nevertheless, the detailed mechanism is still unknown. On the basic of the experiments in laboratory-scale UASB reactor, the effect of hydrogen partial pressure on sludge granulation was evaluated. Size distribution method and specific metabolic activity of the sludge with the operation time were used as a means for estimating the degree of the sludge granulation. At the constant hydrogen loading, the granulation increased as starvation periods in hydrogen supply increased, resulting in high organic removal efficiency. It was evidient that hydrogen play very important role in granulation and sludge granulation was achieved through mutual symbiosis between hydrogen utilizing bacteria and hydrogen producing bacteria under the hydrogen dificient conditions. Key words : granular sludge, UASB reactor, hydrogen partial pressure.

• Proceedings of the Korean Society for Cognitive Science Conference
• /
• 2006.06a
• /
• pp.73-79
• /
• 2006

본 논문에서는 미등록어 처리가 강화된 복합명사 분석 기법을 제시한다. 기본적으로 모든 복합명사 내에 한국어나 외래어의 미등록어가 포함되어 있을 수 있다는 가정하에 분석을 시도한다. 따라서 등록어로 구성된 복합명사에 대해서도 미등록어가 포함된 분해 후보가 생성될 수도 있다. 이는 분해 후보의 수를 크게 증가시키는 문제를 일으킨다. 이 문제에 대처하기 위하여 미등록어의 분류에 따라 미등록어로서의 가능성 여부의 판별 및 제거, 분해 후보 상호간의 견제에 의한 제거 등을 이용하였다. 이러한 과정은 정답 후보 선택시에도 영향을 미쳐 정답이 아닌 분해 후보가 선택되는 것을 방지할 수 있으며, 처리 시간을 줄일 수 있는 이점이 있다. 실험 결과 제시된 기법들이 매우 효과적임을 확인할 수 있었다.

• Proceedings of the IEEK Conference
• /
• 2003.11b
• /
• pp.145-148
• /
• 2003

The agglutinative nature of Korean language makes the problem of automatic indexing of Korean much different from that of Indo-Eroupean languages. Especially, indexing with compound nouns in Korean is very problematic because of the exponential number of possible analysis and the existence of unknown words. To deal with this compound noun indexing problem, we propose a new indexing methods which combines the merits of the morpheme-based indexing methods and the n-gram based indexing methods. Through the experiments, we also find that the best performance of n-gram indexing methods can be achieved with 1.75-gram which is never considered in the previous researches.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1995.06a
• /
• pp.195-199
• /
• 1995

자연언어 처리 시스템의 전단부인 형태소 분석 모듈은 해결해야 할 두 가지 문제를 갖고 있다. 하나는 형태소 분석기가 여러 개의 분석 결과를 출력하여 생기는 품사 중의성이고, 다른 하나는 주어진 문장에 미등록어가 사용되어 형태소 분석이 실패되었을 때이다. 본 논문에서는 이 문제들을 해결하는 한국어 품사 태깅 시스템 KTS를 소개한다. KTS는 주어진 어절에 대해 모든 가능한 분석을 하는 형태소 분석기, 미등록어를 예측하는 미등록어 추정 모듈, 음절 정보와 단서 형태소를 이용하여 미등록어 후보의 수를 줄이는 미등록어 후보 여과기, 그리고 미등록어의 출현을 모델안에 포함한 품사 태깅 모듈로 구성되어 있다. KTS 의 품사태깅 모듈에는 두가지 태깅 방법인 경로 기반 태깅과 상태 기반 태깅의 유일 출력과 다중 출력 기능이 모두 구현되어 있으며, 실험에 의하면, 미등록어가 포함되지 않은 어절에 대해서 89.12%, 미등록어가 포함된 어절에 대해서 68.63%의 정확률을 각각 나타내었다.

• Annual Conference on Human and Language Technology
• /
• 2009.10a
• /
• pp.186-190
• /
• 2009

품사 부착은 각종 자연어처리의 기본적인 요소이며, 크게 규칙 기반 방법과, 통계 기반 방법으로 나눌 수 있다. 대부분은 통계 기반의 기계학습을 이용하고 있으며, 대개 95% 이상의 성능을 보여주고 있다. 그러나 미등록어에 대해서는 성능이 그다지 높지 않다. 이 논문에서는 단어의 접사 정보를 이용해서 미등록어에 대한 품사 부착의 성능을 높이는 방법을 제안한다. 제안된 시스템은 CRF(Conditional Random Fields)를 이용하며, 그 자질의 일부로 접사 정보를 이용한다. 그 결과 미등록어에 대해서 약 40%의 성능이 개선되었다. 앞으로 미등록어에 적합한 자질을 연구하고 개발할 필요가 있을 것으로 생각된다.

• 한국방재학회:학술대회논문집
• /
• 2008.02a
• /
• pp.123-126
• /
• 2008

This research develops a methodology for standard design of spatial Database utilizing the disaster ethnography. Especially, the disaster response operation is sensitive to the size of the disaster, location, damage situation, resource a variability, etc. Moreover, there are many unknown and unexpected factors that will affect the disaster response strategy. But, the future Crisis Management Systems is needed that past disaster teaching. In another words, from now on the response systems need to prepare several scenarios and spatial data and manual etc. before the disaster. Then, this research is the experimental research which examined the relationship between the disaster-ethnography and the GIS spatial data of disaster.