• 한국HCI학회:학술대회논문집
• /
• 2006.02a
• /
• pp.1379-1386
• /
• 2006

전자우편이나 문자 메세지를 이용할 때 겪는 불편함 중 하나는 상대방이나 기계에 정서 정보를 전달하기 어렵다는 점이다. 정서 정보를 메시지에 싣기 위해서는 컴퓨터나 디지털 기기가 정서를 인식하거나 사용자가 정서를 입력해야 한다. 기존의 정서 인식 방법은 생리적, 신체적 측정치를 이용하는 것인데, 이 경우 측정을 위한 별도의 장비가 필요하고 현재 자신의 정서 상태와 다른 정서를 표현할 수 없다는 단점이 있다. 특히 소형 모바일 기기를 이용할 때 다른 측정 장치를 사용하는 것은 더욱 어렵다. 이런 문제를 해결하기 위해 모바일 기기를 사용하는 환경에서 사용자가 원하는 정서를 기계에 입력하기 위해 동작을 이용하려는 연구가 진행되었다(Fargerberg, Stahl, & Hook, 2003). 본 연구에서는 Laban Movement Analysis에서 동작을 구성하는 다섯 요소 중 노력(effort)과 모양(shape) 요소를 재구성하여, 방향성 차원, 무게감 차원, 시간감 차원으로 동작을 구분하고 총 20개의 동작을 선정하였다. 또한 한덕웅과 강혜자(2000)가 수집한 834개 정서 어휘를 평정하여 동작을 통해 표현하고 전달되기 쉬운 정서 어휘 50개를 선택하였다. 최종 실험에서 참가자들은 20개의 동작에 대해 50개의 정서 어휘를 평정하고 데이터는 범주형 주성분분석을 이용하여 분석하였다. 분석 결과 Russell(1980)의 이차원 정서 구조 모형에서 각성 수준 차원은 동작의 무게감과 시간감 차원과 관련이 있는 것으로 나타났다. 강하고 빠른 동작일수록 각성 수준이 높은 정서가 나타났다. 또한 동작의 방향성 차원은 정서의 종류와 관련이 있는 것으로 드러났다. 직선 움직임은 높은 각성 수준의 부정적 정서와, 흔듦 움직임은 불안 및 초조와, 원형 움직임은 즐거운 정서와 관련이 있는 것으로 나타났다. 이는 동작을 통하여 정서 정보를 효과적으로 전달할 수 있음을 보여주었고, 동작과 정서를 연관 짓기 위해 방향성 차원과 무게감 차원 그리고 시간감 차원을 고려할 필요가 있음을 시사한다.

• The Journal of the Korea Contents Association
• /
• v.15 no.4
• /
• pp.545-555
• /
• 2015

We investigated the influence of negative emotion such as state-trait anxiety and depression and working memory (WM) on Foreign Language Vocabulary Learning (FLVL) of South Korean university students. Also, its implications for developing contents for FLVL were discerned. To do so, state-trait anxiety and depression inventories as well as four kinds of WM test were performed for 132 undergraduate students. Participants also had two semantic learning sessions for Swahili words. The mean scores of negative emotions were normal level. The results of structural equation modeling (SEM) showed that there was no effect of negative emotion on FLVL, while direct effects of the negative emotion on WM and the WM on FLVL were significant. Such results suggested that FLVL would be weakened, with the result that WM had been impaired by negative emotions. These outcomes suggested that when developing FLVL content for university students, it is necessary to consider the negative emotions of foreign language learners and to develop the contents for FLVL in the light of WM load.

• The Journal of the Acoustical Society of Korea
• /
• v.22 no.3
• /
• pp.194-201
• /
• 2003

In speech recognition, confidence measuring is to decide whether it can be accepted as the recognized results or not. The confidence is measured by integrating frames into phone and word level. In case of word recognition, the confidence measuring verifies the results of recognition and Out-Of-Vocabulary (OOV). Therefore, the post-processing could improve the performance of recognizer without accepting it as a recognition error. In this paper, we measure the confidence modifying log likelihood ratio (LLR) which was the previous confidence measuring. It accumulates only those which the log likelihood ratio is negative when integrating the confidence to phone level from frame level. When comparing the verification performance for the results of word recognizer with the previous method, the FAR (False Acceptance Ratio) is decreased about 3.49% for the OOV and 15.25% for the recognition error when CAR (Correct Acceptance Ratio) is about 90%.

• Phonetics and Speech Sciences
• /
• v.14 no.3
• /
• pp.41-49
• /
• 2022

The average acoustic distribution of American English speakers was statistically compared with the English-speaking patterns of gifted students in a Science Academy in Korea. By analyzing speech recordings, the duration time of which is much longer than in previous studies, this research identified the degree of acoustic proximity between the two parties and the predictability of English academic achievement of gifted high school students. Long-term spectral acoustic power distribution vectors were obtained for 2,048 center frequencies in the range of 20 Hz to 20,000 Hz by applying an long-term average speech spectrum (LTASS) MATLAB code. Three more variables were statistically compared to discover additional indices that can predict future English academic achievement: the receptive vocabulary size test, the cumulative vocabulary scores of English formative assessment, and the English Speaking Proficiency Test scores. Linear regression and correlational analyses between the four variables showed that the receptive vocabulary size test and the low-frequency vocabulary formative assessments which require both lexical and domain-specific science background knowledge are relatively more significant variables than a basic suprasegmental level English fluency in the predictability of gifted students' academic achievement.

• Journal of KIISE:Software and Applications
• /
• v.36 no.8
• /
• pp.642-652
• /
• 2009

In order to improve the performance of the existing relation extraction approaches, we propose a method for combining two pivotal concepts which play an important role in classifying semantic relationships between entities in text. Having built a composite kernel-based relation extraction system, which incorporates both entity features and syntactic structured information of relation instances, we define nine classes of lexical features and synthetically apply them to the system. Evaluation on the ACE RDC corpus shows that our approach boosts the effectiveness of the existing composite kernels in relation extraction. It also confirms that by integrating the three important features (entity features, syntactic structures and contextual lexical features), we can improve the performance of a relation extraction process.

• The Journal of the Korea Contents Association
• /
• v.14 no.1
• /
• pp.150-157
• /
• 2014

This thesis intends to grasp the degree of utilization of visual images by understanding the relational properties between picture and letter and considering the children as users, through the analysis of currently published books of English vocabulary for children. Accordingly, the types of picture used in the books of English vocabulary for children, the degree of utilization of picture, combination types of picture and letter, and semantic consistency of picture and letter are reviewed. As a result of analysis, the degree of utilization of picture is high in general, in order of illustration, cartoon, and the mix of illustration and cartoon. In the combination form of picture and letter, the degree of utilization appears in order of picture plus vocabulary, letters without illustration, and pictorial symbol. In particular, the higher semantic consistency of picture and letter, it is effective in learning, however, semantic consistency is low, generally. Pictorial symbol type shows the frequency of the highest combination type in the five groups of higher semantic consistency. In conclusion, the presented types of picture and letter, shown in the currently published books of English vocabulary for children, are similar types by the publishing companies, thus, effective design research should be required based on diverse levels of children.

• Annual Conference on Human and Language Technology
• /
• 2020.10a
• /
• pp.227-232
• /
• 2020

본 논문은 의미역 주석(Semantic Role Labeling) 자원인 FrameNet을 준구어 말뭉치인 드라마 대본에 주석하는 과정과 주석 결과에 대해 서술한다. 본 논문에서는 프레임 - 프레임 논항 구조의 주석 범위를 한 문장에서 여러 발화로 이루어진 장면 (Scene) 단위의 대본으로 확장하여 문장 내에서 생략된 프레임 논항(Null-Instantiated Frame Elements)을 장면 단위 대본 내의 다른 발화에서 복원하였다. 본 논문은 프레임 자동 분석기를 통해 동일한 드라마의 한국어, 영어 대본에 FrameNet 주석을 한 드라마 대본을 선발된 주석자에 의해 대상 어휘 적합성 평가, 프레임 적합성 평가, 생략된 프레임 논항 복원을 실시하고, 자동 주석된 대본과 주석자 작업 후의 대본 결과를 비교한 결과와 예시를 제시한다. 주석자가 자동 주석된 대본 중 총 2,641개 주석 (한국어 1,200개, 영어 1,461개)에 대하여 대상 어휘 적합성 평가를 실시하여 한국어 190개 (15.83%), 영어 226개 (15.47%)의 부적합 대상 어휘를 삭제하였다. 프레임 적합성 평가에서는 대상 어휘에 자동 주석된 프레임의 적합성을 평가하여 한국어 622개 (61.68%), 영어 473개 (38.22%)의 어휘에 대하여 새로운 프레임을 부여하였다. 생략된 프레임 논항을 복원한 결과 작업된 평균 프레임 논항 개수가 한국어 0.780개에서 2.519개, 영어 1.290개에서 2.253개로 증가하였다.

• Journal of Korea Entertainment Industry Association
• /
• v.14 no.5
• /
• pp.181-191
• /
• 2020

The purpose of this study is to suggest a direction for efficient teaching and learning idiomatic expressions in Korean to improve the vocabulary of Korean language learners. In order to make learning more interesting and enhance learning effectiveness for Korean language learners, the drama, , which was popular in Korea, was used as educational material. Since idomatic language is formed and used based on Korean history, culture, and social background, dramas containing Korean culture and sentiments can be said to be suitable materials for the teaching and learning of Korean idiomatic expressions. By analyzing the drama , 277 significant vocabularies were extracted from the drama based on vocabulary actually used. Among these, 124 idiomatic expressions were extracted after excluding overlapping expressions. Idiomatic expressions extracted in this way were classified based on vocabulary used more than 2 times. In addition, in order to select idiomatic expressions suitable for the level of the learners, 46 final expressions for Korean language education were selected considering the difficulty of vocabulary. Lastly, when the materials selected in the drama were used for education, the precautions for teaching and learning, and the direction of education on idiomatic language were classified into elementary, intermediate, and advanced grades and presented.

• The KIPS Transactions:PartB
• /
• v.19B no.2
• /
• pp.149-154
• /
• 2012

The aim of this paper is to collect relevant keywords from clicklog data including user's keywords and URLs accessed using them. Our main hyphothesis is that two or more different keywords may be relevant if users access same URLs using them. Also, they should have higher relationship when the more same URLs are accessed using them. To validate our idea, we collect relevant keywords from clicklog data which is offered by a portal site. As a result, our experiment shows 89.32% precision when we define answer set to only semantically same words, and 99.03% when we define answer set to broader sense. Our approach has merits that it is independent on language and collects relevant words from real world data.

• 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.