• Journal of KIISE:Software and Applications
• /
• v.31 no.12
• /
• pp.1649-1659
• /
• 2004

Text Generation is a process of generating comprehensible texts in human languages from some underlying non-linguistic representation of information. Among several sub-processes for text generation to generate coherent texts, this paper concerns referring expression generation which produces different types of expressions to refer to previously-mentioned things in a discourse. Specifically, we focus on pronominalization by zero pronouns which frequently occur in Korean. To build a generation model of referring expressions for Korean, several features are identified based on grammatical information and cost-based centering model, which are applied to various machine learning techniques. We demonstrate that our proposed features are well defined to explain pronominalization, especially pronominalization by zero pronouns in Korean, through 95 texts from three genres - Descriptive texts, News, and Short Aesop's Fables. We also show that our model significantly outperforms previous ones with a 99.9% confidence level by a T-test.

• Korean Institute of Interior Design Journal
• /
• no.18
• /
• pp.81-86
• /
• 1999

The purpose of this study is to find out to meaning and practical use of 'Anbang' of which unique roles and function has not been changed till now since Chosun dynasty. The Study has been carried out by selecting 27 novels in which the word 'Anbang' mentioned frequently and analyzing the sentences with 'Anbang' . The major finding were summarized as follows: 1. Since Chosun dynasty till now, 'Anbang' in most important space for Korean in residential area its meaning and practical use had not been changed. 2. The meaning of Anbang could be categorized into two parts; One was general symbolic meaning which was perceived mentally in general and the other was specific symbolic meaning which was varied by uses. In case of general symbolic meaning 'Anbang' implied a space with light turned on late at night, a very special space from the point of interior decoration and size, a space to keep very important stuffs and a space to deal with very important business. In case of specific symbolic meaning during Chosun dynasty, 'Anbang' was used as a pronoun to imply a mistress as major occupant was a mistress. Since then, gradually, 'Anbang' was used as a pronoun to imply married couple as married couple becomes major occupant. 3. The practical use of 'Anbang' cold be categorized into two; One was routine usage and the other was exceptional usage. The typical routine use were sleeping, dining family gathering, visitor greeting, clothes changing and sewing. The exceptional use were a Sebae(new year's bow), Pyebaek(make a deep bow and offer her gifts to her parents-in-law) and patient curing.

• Language and Information
• /
• v.7 no.1
• /
• pp.55-68
• /
• 2003

Most of the previous approaches to English agreement phenomena have relied upon only one component of the grammar (e.g., either syntax, or semantics, or pragmatics). This paper argues that interrelationships among different grammatical components play crucial roles in such phenomenon too (cf. Kathol 1999 and Hudson 1999). The paper proposes that, contrary to traditional wisdom, English determiner-noun agreement is morpho-syntactic whereas subject-verb and pronoun-antecedent agreement are reflections of index agreement (cf. Pollard and Sag 1994). The present hybrid analysis of English agreement shows the importance of the interaction of different components of the grammar in accounting for English agreement phenomena. In particular, once we allow morphology to tightly interact with the system of syntax, semantics, or even pragmatics, we could provide a solution to some puzzling English agreement phenomena. This allows a more principled theory of English agreement.

• Proceedings of the Korean Society for Language and Information Conference
• /
• 2000.06a
• /
• pp.36-48
• /
• 2000

이 논문에서는 국지적인 대명사의 해결을 위한 이론적인 틀로 널리 알려진 중심화이론의 여러 이론적인 장치들을 정보포장이론에 도입함으로써 정보포장이론을 통제 정보포장이론이라는 새로운 담화의미론으로 발전시킬 수 있음을 보이고자 한다. 이로써 한국어 담화상에 나타나는 대명사의 선행사 탐색이 상당히 설득력있게 이루어질 수 있음을 보인다. 이를 위해 먼저 제한된 영역대화라 할 수 있는 호텔예약대화에 나타나는 영대명사의 특성에 대해 논의한다. 이어서, 대명사 해결과 관련하여 정보구조가 중심화이론의 주요 구성요소인 전향적 중심리스트 서열 구성에 직접 반영되어야 함을 보인다. 마지막으로 대명사의 선행사는 대명사를 담화통어할 수 있는 위치에 있어야 한다는 담화통어제약을 제안하여 광역대화에 나타나는 명시적인 대명사의 해결을 위한 하나의 대안을 제시한다.

• Proceedings of the Korean Society for Language and Information Conference
• /
• 2007.11a
• /
• pp.532-541
• /
• 2007

This paper investigates children's comprehension and production of demonstrative pronouns (DPs), 'zhege' (this) and 'nage' (that), in Mandarin Chinese. Subjects are children of ages three, four, five and six. Based on the results of the present experiment, children's developmental stages and the corresponding age grading are provided. Also, the present study incorporates a physical clue into the experiment. The result suggests that in the acquisition of deixis children rely highly on physical context to work out the meaning distinction. In addition, Piaget's egocentrism hypothesis and H. Clark's marking hypothesis are examined in the study. The result seems to support the egocentrism hypothesis. Subjects under the age of six do fail to shift the deictic center when they and the experimenter have a different perspective. As for the marking hypothesis, the study seems to challenge the hypothesis. The result shows that children actually performed better on the marked term 'zhege' than the unmarked member 'nage'.

• Language and Information
• /
• v.5 no.1
• /
• pp.17-34
• /
• 2001

Kamp & Reyle's (1993) proposal to represent split antecedents to a plural pronoun in terms of sum- mation and abstraction is critically reviewed in this paper, In point our some weak- nesses of their analysis as well as wrong predictions they make. In propose to treat the partitive reading found in bare numeral constructions by separating the conven-tional DRS construction rules from the cognitively motivated DRS-operations at a different level. A preference rule is also proposed that would constrain the sortal structure of discourse referents when such operations as summation and abstraction are enforced at the DRS's of relevant levels. Evidence for the separate treatment of linguistically motivated processes apart from cognitively motivated ones comes from both English and Korean constructions involving definite plural pronouns and numeral classifiers.

• Annual Conference on Human and Language Technology
• /
• 2009.10a
• /
• pp.197-201
• /
• 2009

이 논문은 영어 문장에서 대명사의 참조해소 시스템을 구현한다. 대명사는 문장에서 반복되는 말 대신에 사용하는 단어이다. 반복되는 말을 선행어라고 하며 대명사는 선행어보다 간결한 형식으로 사용된다. 정보검색이나 정보추출에서 대명사를 그대로 색인하여 검색하면 정확한 정보를 추출할 수 없다. 따라서 대용어가 가리키는 개체를 정확히 파악해서 이 정보를 색인하고 검색하면 정보검색, 정보추출, 질의응답의 성능을 크게 개선할 수 있다. 이 논문에서는 CRF모델을 이용해서 이용하여 영어 문서에서 대명사 참조해결 방법을 제안하고 이를 구현한다.

• Annual Conference on Human and Language Technology
• /
• 2011.10a
• /
• pp.98-101
• /
• 2011

이 논문은 한-영 대화체 번역 시스템에서 영형 대명사 해소를 위한 새로운 방법론을 제시하였다. 영형 대명사는 문맥, 상황, 세상 지식으로부터 추론될 수 있는 문장에서 생략된 요소이다. 이 논문은 특히 주어-대명사 생략 현상에 대해 다루고 있는데, 그 이유는 드라마 대본이나 인스턴트 메신저 채팅과 같은 한국어 대화체에서는 매우 일반적인 현상이기 때문이다. 이 논문에서 우리는 많은 양의 지식을 요구하지 않는 간단한 방법론을 제시하였다. 평가결과 우리의 방법은 0.79의 F-measure 스코어를 달성하였고, 전체번역률의 측면에서는 약 4.1% 정도의 향상효과가 있었다.

• Language and Information
• /
• v.6 no.1
• /
• pp.91-103
• /
• 2002

The temporal structure of events on the discourse level has long been of great interest in both theoretical and computational linguistics. In this paper, we offer a unified approach to the temporal relationships related to a hierarchical discourse structure. We apply the method of pronoun resolution to the interpretation of tense. It is based on an analysis within the framework of the controlled information packaging theory. A unique aspect of our account is that temporal interpretation across discourse segments in global discourse is subject to the same principles as the interpretation of global anaphora, and that there is thus no need to postulate independent principles to account for the discourse behaviour of tense. In this way, we can neatly explain the general view that tense parallels the anaphoric nature of pronouns.

• Korean Journal of Cognitive Science
• /
• v.12 no.3
• /
• pp.29-46
• /
• 2001

This study attempted to demonstrate that the integration of the two premises in linear syllogism is mediated by the middle term the term that is repeated in the two premises. In Experiment 1. we examined whether representing the middle term is more important than representing the end terms. We asked a question to each premise. Depending on the order of the questions either the two end terms or the middle term became the answer in both premises. Participants solved the problems better when the middle term became the answer suggesting that it is more important to represent the middle term than the end terms. In Experiment 2 we examined whether additional processing is needed for the integration beyond establishing co-referential link through the middle term. We pronominalized the middle term in the second premise and provided two kinds of information to disambiguate the pronoun. In the direct information condition we provided information about who the pronoun is whereas in the indirect information condition we provided information about the relative location of the pronoun. Participants solved the problems more quickly in the indirect information condition than in the direct information condition indicating that mere co-referential link was not enough and that the relative location of the middle term needs to be computed for the integration of the two premises.