• Journal of KIISE:Computing Practices and Letters
• /
• v.16 no.10
• /
• pp.1005-1009
• /
• 2010

Elliptical and anaphoric utterances occur frequently during spoken dialogue. Because discourse analysis rests on the basic premise that linguistic items cannot be understood without reference to the context, ellipsis and anaphora resolution plays an important role in discourse analysis. In this paper, we present a spoken dialogue system improving the robustness at dialogue level based on discourse analysis, such as anaphora and ellipsis resolution. The applicability and effectiveness of the proposed method is evaluated in the TV domain.

• Annual Conference on Human and Language Technology
• /
• 2022.10a
• /
• pp.320-323
• /
• 2022

상호참조해결은 문서에서 명사, 대명사, 명사구 등의 멘션 후보를 식별하고 동일한 개체를 의미하는 멘션들을 찾아 그룹화하는 태스크이다. 딥러닝 기반의 한국어 상호참조해결 연구들에서는 BERT를 이용하여 단어의 문맥 표현을 얻은 후 멘션 탐지와 상호참조해결을 동시에 수행하는 End-to-End 모델이 주로 연구가 되었으며, 최근에는 스팬 표현을 사용하지 않고 시작과 끝 표현식을 통해 상호참조해결을 빠르게 수행하는 Start-to-End 방식의 한국어 상호참조해결 모델이 연구되었다. 최근에 한국어 상호참조해결을 위해 구축된 ETRI 데이터셋은 WIKI, QA, CONVERSATION 등 다양한 도메인으로 이루어져 있으며, 신규 도메인의 데이터가 추가될 경우 신규 데이터가 추가된 전체 학습데이터로 모델을 다시 학습해야 하며, 이때 많은 시간이 걸리는 문제가 있다. 본 논문에서는 이러한 상호참조해결 모델의 도메인 적응에 Continual learning을 적용해 각기 다른 도메인의 데이터로 모델을 학습 시킬 때 이전에 학습했던 정보를 망각하는 Catastrophic forgetting 현상을 억제할 수 있음을 보인다. 또한, Continual learning의 성능 향상을 위해 2가지 Transfer Techniques을 함께 적용한 실험을 진행한다. 실험 결과, 본 논문에서 제안한 모델이 베이스라인 모델보다 개발 셋에서 3.6%p, 테스트 셋에서 2.1%p의 성능 향상을 보였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.672-674
• /
• 2023

본 논문은 뉴로-심볼릭 구조를 바탕으로 일반 텍스트로부터 온톨로지 생성이 가능한 심층 신경망 기반 온톨로지 추출기를 제안한다. 온톨로지 추출 단계를 (i) 온톨로지 학습 및 (ii) 온톨로지 생성의 2 단계로 상정, (i) 일반 텍스트로부터 문장 구조 및 논리적 관계를 학습하는 트랜스포머 기반 심층 생성 신경망 출력을 이용하여 (ii) 계층적으로 결합한 심볼릭 추론기로 온톨로지를 생성하는 뉴로-심볼릭 구조 온톨로지 추출기를 구현하였다. 1800 개 훈련 집합으로 학습 후 200 개 테스트 집합으로 평가한 결과, 정확도 91.9%, Precision 100%, Recall 99.1%로 비교 모델 OpenIE 의 성능에 비해서 각각 83.8%, 1.8%, 3.5% 개선된 것을 확인하였다. 정성적 품질에 있어서, 복잡한 문장 (예: 관계대명사, 접속사, 중첩 구조)에서도 비교 모델에 비해 더 정밀한 온톨로지 생성 결과를 보였다.

• School Mathematics
• /
• v.9 no.2
• /
• pp.181-196
• /
• 2007

It would have a trouble to communicate mathematically without an appropriate use of mathematical language. Therefore it is necessary to form mathematics classroom culture to encourage students to use mathematical language precisely. A four-month teaching experiment in a 4th grade mathematics class was conducted focused the accurate use of mathematical language. In the course of the teaching experiment, children became more careful to use their language precisely. The use of demonstrative pronouns such as this or that as well as the use of inaccurate or wrong expressions was diminished. Children became to use much more mathematical symbols and terms instead of their imprecise expressions. The result of the experiment suggests that the culture that encourage students to use mathematical language precisely can be formed in elementary mathematics classroom.

• Journal of KIISE
• /
• v.41 no.11
• /
• pp.992-1005
• /
• 2014

Coreference resolution finds all expressions that refer to the same entity in a document. Coreference resolution is important for information extraction, document classification, document summary, and question answering system. In this paper, we adapt Stanford's Multi-pass sieve system, the one of the best model of rule based coreference resolution to Korean. In this paper, all noun phrases are considered to mentions. Also, unlike Stanford's Multi-pass sieve system, the dependency parse tree is used for mention extraction, a Korean acronym list is built 'dynamically'. In addition, we propose a method that calculates weights by applying transitive properties of centers of the centering theory when refer Korean pronoun. The experiments show that our system obtains MUC 59.0%, $B_3$ 59.5%, Ceafe 63.5%, and CoNLL(Mean) 60.7%.

• Korean Journal of Cognitive Science
• /
• v.18 no.1
• /
• pp.35-56
• /
• 2007

The English language has developed from a language with optional subjecthood Into a language with obligatory subjecthood due to a general reduction of inflections. Two types of subject omission, pro-drop and conjunction reduction, have been reported in the history of English. Old English with rich inflections had both referential pro-drop and conjunction reduction. Middle English with much lesser inflections still witnessed pro-drop and conjunction reduction, but in such a decreasing way that modern English with a loss of inflections developed from Middle English hardly has either pro-drop or conjunction reduction. This paper explores both the phenomena relating to optional subjecthood in Old, Middle, and Modern English in light of the cognitive processes of the universal, hierarchical constraints that are assumed to be inherent in English speakers' cognitive fatuity. It is found that optional subjecthood in Old, Middle, and Modern English is correctly raptured in terms of the distinct rankings of the proposed constraints, and that it is closely related to whether each of Old, Middle, and Modern English has rich inflections.

• The KIPS Transactions:PartB
• /
• v.9B no.6
• /
• pp.817-826
• /
• 2002

In Korean dialogue honorification phenomenon may occur, an honorific pronoun may be used, and a subject or an object may be completely omitted when it can be recovered based on context. This paper proposes that in order to process Korean dialogue in which such distinct linguistic phenomena occur and to construct its representation structure we mark and use the following information explicitly, not implicitly : information about dialogue participants, information about the speech act of an utterance, information about the relative order of social status for the people involved in dialogue, and information flow among utterances of dialogue. In addition, this paper presents a method of marking and using such situational information and an appropriate representation structure of Korean dialogue. In this paper we set up Korean dialogue representation structure by modifying and extending DRT (Discourse Representation Theory) and SDRT (Segmented Discourse Representation Theory). Futhermore, this paper shows how to process Korean dialogue computationally and construct its representation structure by using Prolog programming language, and then applies such representation structure to spontaneous Korean dialogue to know its validity.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.21 no.5
• /
• pp.426-432
• /
• 2020

Music with the development of the media since the 20th century marks a change into an era where the live performance and music of the masses, which were hard to imagine in previous times, are shared. Based on this cultural trend, teenage music, which has been completely alienated from existing culture, is in line with the birth of Rock 'n' Roll, an event that has entered the mainstream, and the emergence of the guitar, especially the solid body electric guitar. The Fender Stratocaster, which is referred to as the epitome of this electric guitar, has joined the history of popular music for rock 'n' roll. To this day, the immense influence, which still encompasses many followers and generations, has served as a bridge that continues to be reproduced, even in the historical trend of popular music and the status of electric guitars. In addition, even in the rapid development of popular music and media, we will always be able to give true meaning and value to the vitality with the times. This paper examines the features and marks of these Fender Stratocasters.

• The Journal of the Korea Contents Association
• /
• v.17 no.12
• /
• pp.378-385
• /
• 2017

The aim of this study was to analyze cohesion and word information among different types of questions in the English reading section of the College Scholastic Ability Tests (CSAT). The types of questions were divided into three categories: macro reading, micro reading, and indirect writing. Reading texts from 1994 to 2017 CSAT were analyzed by Coh-Metrix, an automated evaluation program of text and discourse. The findings of this study indicated that there were statistical differences among the three categories of questions for noun overlap, stem overlap, adversative and contrastive connective, additive connective, pronoun incidence, age of acquisition, concreteness for content word, imagability, and meaningfulness. The information of the findings bore pedagogic implications for developing textbooks, questions for CSAT, and reading strategies by students.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.13 no.6
• /
• pp.155-159
• /
• 2013

This article employs Naive Bayesian Classifier to realize a system that can distinguish the sex of a name. Unlike foreign names, in Korean names, the pronoun referring to a person shows discordance with sex. With the characteristics of Korean names, however, the study distinguishes names frequently used for men and for women. And as it also includes names of which sex is rather ambiguous such as proper nouns, the accuracy of it is somewhat low. The result of the experiment conducted in this article indicates 84% accuracy for Korean men and 88% for Korean women; thus, the total accuracy equals 86%. Meanwhile, about foreign names, men show 80% accuracy, and women 84%, so the total accuracy equals 83%.