• Proceedings of the Korea Information Processing Society Conference
• /
• 2002.11a
• /
• pp.573-576
• /
• 2002

본 논문에서는 언어학의 관점에서 시제와 상의 특성을 알아 보고, Reichenbach 의 시제와 상을 살펴 보며, 상의 기술에 있어서 언어학자의 여러 주장과 문제점을 고찰하며, 한국어의 시제를 영어의 12 시제와 비교한다. 그리고 한국어의 여러 시제 의미를 분석하고, 시제의 구조를 개념그래프 이론에 의거하여 기술한다. 실험에서는 시제가 포함된 문장을 입력 받아 개념그래프로 변환하는 프로그램을 구현하고 그 결과를 기술한다.

• Annual Conference on Human and Language Technology
• /
• 2019.10a
• /
• pp.161-164
• /
• 2019

본 논문은 BERT 모델을 이용하여 동형이의어의 단어 표현(Word Representation) 차이에 대한 실험을 한다. BERT 모델은 Transformer 모델의 인코더 부분을 사용하여 양방향을 고려한 단어 예측과 문장 수준의 이해를 얻을 수 있는 모델이다. 실험은 동형이의어에 해당되는 단어의 임베딩으로 군집화를 수행하고 이를 Purity와 NMI 점수로 계산하였다. 또한 각 단어 임베딩 사이를 코사인거리(Cosine Distance)로 계산하고 t-SNE를 통해 계층에 따른 변화를 시각화하였다. 군집된 결과는 모델의 중간 계층에서 점수가 가장 높았으며, 코사인거리는 8계층까지는 증가하고 11계층에서 급격히 값이 변하는 것을 확인할 수 있었다.

• Annual Conference on Human and Language Technology
• /
• 2019.10a
• /
• pp.375-379
• /
• 2019

언어학에서의 환유법은 표현을 위해 빌려온 대상이 다양한 의미로 해석 가능하기에 매우 어렵고 난해한 분야이다. 환유의 특성 상 주어진 엔티티의 환유 여부를 구분하기 위해서는 앞뒤 단어와의 연관성 뿐만 아니라 문장 전체의 문맥 정보에 대한 고려가 필수적이다. 최근 이러한 문맥 정보를 고려하여 학습된 다양한 모델들이 등장하면서 환유법에 대한 연구를 하기에 좋은 환경이 구축되고 있다. 본 논문에서는 언어학적 자질 정보를 최소화한 딥러닝을 이용한 환유 해소 모델을 제안한다. LSTM 기반의 feature-based 모델과 및 BERT, XLNet, RoBERTa와 같은 fine-tuning 모델들에 대한 실험을 진행하였다. 실험 결과, fine-tuning 모델들이 baseline과 비교하여 뛰어난 성능 향상을 가져왔으며, 특히 XLNet 모델은 두 개의 환유 해소 데이터 SemEval 2007와 ReLocaR에 대해 각각 90.1%과 95.8%의 정확도를 보여주었다.

• Journal of The Korean Association For Science Education
• /
• v.37 no.1
• /
• pp.215-223
• /
• 2017

This study aims to find an effective way to present essential science concepts in national science curriculum through international comparisons. Next Generation Science Standard (US), Ontario Science Curriculum (Canada), Australia Science Curriculum, and British/English Science Curriculum were selected for comparison. In science curriculum documents, these countries used terms such as 'Key ideas,' 'Big ideas,' 'Key concepts,' 'Disciplinary core ideas.' and 'Fundamental concepts' to present essential concepts of science. This study reviewed the characteristics of the meaning, the status, and the role of essential concepts country by country. The result shows essential concepts have been used with different meanings and statutes in each case. Furthermore, various roles were performed through essential concepts in order to organize their science curriculum. From these foreign nation's cases, this study proposes several ways to present essential science concepts based on results. First, interdisciplinary integrated concepts were needed to organize an integrated science curriculum. In science curriculum documents of the United States, Canada, Australia and England, two types of terms were used in order to structuralize an integrated science curriculum. Second, essential concepts should include concepts related with function and value as well as scientific knowledge. Third, essential concepts need to be presented in such a way as to show specific contexts. Therefore, selecting appropriate contents and structure are needed to be able to improve the way to present essential concepts in Korea's educational environment.

• Archives of design research
• /
• v.17 no.3
• /
• pp.335-342
• /
• 2004

Recently there have been various research activities regarding Korean traditional aesthetics. However, those researches were mainly conducted individually, partially, and periodically, which resulted in unsystematic and incomprehensive works. Therefore, it is required to orginze all the precedent research works with more systematic and objective framework. Generally speaking, all the human activities including aesthetic activity have ends, procedure and means. In other words, human being needs three key elements for realizing any thought and those three elements include contents, formal, and practical element. Element of contents is ultimate goal to accomplish as value, concept, and meaning of thought with their aims. Formal element includes methods, principles, norms, procedure, formality and style comprising of thought in order to accomplish the goal. Finally, practical element refers to specific means, tool, media, material and techniques to concretize the contents through form. Almost all of thoughts and meaning which human being tries to express consist of language. Major elements in sentence include 'subject (omissible)' , 'objects (aim)', 'predicate (formality)', 'complement (means)' and they are composed systematically and hierarchically with rules in sentence. The study compared human activity model with language structure and analyzed their implication with design (aesthetics), which made it possible to propose analytic frameworks for traditional aesthetics. In addition, the study also systematically organized the way to realize traditional aesthetic value in the present context based on the methods developed in this study.

• Cross-Cultural Studies
• /
• v.45
• /
• pp.55-83
• /
• 2016

To provide a set of data with which instructors may boost the sagging demand for Anglo-American poetry classes, this thesis classifies the kinds of difficulties the students face in reading English poems. Asses to the classification is an analysis on the causes of the difficulties at different levels of the reading process, from the linguistic to the cultural. Arnoldian insight argues that poetry is the best of all forms of writing. Without an ample exposure to poetry, average English majors would barely sharpen the skills that they use to deal with other kinds of writing. To help ease the continuing need for a workable teaching model in English poetry reading classes, this thesis suggests focusing on the kinds of wrong translations produced by the students. According to the theory of cultural translation, any translation, even the wrong kind, is already a product of a very complicated process of interpretation that involves many cultural factors. With the analysis of these factors discovered in Korean college English reading classes, this thesis tries to explain the mechanisms through which wrong translations are produced, since these inevitably lead to wrong interpretations of given poetic texts.

• The Journal of Bigdata
• /
• v.5 no.1
• /
• pp.197-208
• /
• 2020

Lexical ambiguity means that a word can be interpreted as two or more meanings, such as homonym and polysemy, and there are many cases of word sense ambiguation in words expressing emotions. In terms of projecting human psychology, these words convey specific and rich contexts, resulting in lexical ambiguity. In this study, we propose an emotional classification model that disambiguate word sense using bidirectional LSTM. It is based on the assumption that if the information of the surrounding context is fully reflected, the problem of lexical ambiguity can be solved and the emotions that the sentence wants to express can be expressed as one. Bidirectional LSTM is an algorithm that is frequently used in the field of natural language processing research requiring contextual information and is also intended to be used in this study to learn context. GloVe embedding is used as the embedding layer of this research model, and the performance of this model was verified compared to the model applied with LSTM and RNN algorithms. Such a framework could contribute to various fields, including marketing, which could connect the emotions of SNS users to their desire for consumption.

• Journal of KIISE:Software and Applications
• /
• v.27 no.4
• /
• pp.448-462
• /
• 2000

Coordinate constructions in natural language pose a number of difficulties to natural language processing units, due to the increased complexity of syntactic analysis, the syntactic ambiguity of the involved lexical items, and the apparent deletion of predicates in various places. In this paper, we address the syntactic characteristics of the coordinate constructions in Korean from the viewpoint of constructing a competence grammar, and present a version of combinatory categorial grammar for the analysis of coordinate constructions in Korean. We also show how to utilize a unified lexicon in the proposed grammar formalism in deriving the sentential semantics and associated information structures as well, in order to capture the discourse functions of coordinate constructions in Korean. The presented analysis conforms to the common wisdom that coordinate constructions are utilized in language not simply to reduce multiple sentences to a single sentence, but also to convey the information of contrast. Finally, we provide an analysis of sample corpora for the frequency of coordinate constructions in Korean and discuss some problematic cases.

• Journal of Digital Convergence
• /
• v.17 no.10
• /
• pp.321-327
• /
• 2019

The purpose of this study is to develop a smart device-based Language Rehabilitation Program (LRP) to improve communication ability for the patients with language disability. The content of the LRP includes a variety of semantic categories and grammatical elements and consists of 17 semantic categories, 29 tasks and 3780 items to improve comprehension/production ability at word level, semantic category level, sentence level and discourse level. We developed LRP as a Windows-base management program and an Android-base language rehabilitation application. LRP was developed into an application for smart devices, providing real-time delivery of training contents, measurement and database of training task results, and patient progress and monitoring. A follow-up study will be conducted on the verification of the language rehabilitation effect using LRP by patients with language disability.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.10a
• /
• pp.34-36
• /
• 2022

In the existing text sentiment analysis models, the entire text is usually directly modeled as a whole, and the hierarchical relationship between text contents is less considered. However, in the practice of sentiment analysis, many texts are mixed with multiple emotions. If the semantic modeling of the whole is directly performed, it may increase the difficulty of the sentiment analysis model to judge the sentiment, making the model difficult to apply to the classification of mixed-sentiment sentences. Therefore, this paper proposes a sentiment analysis model BHGCN that considers the text hierarchy. In this model, the output of hidden states of each layer of BERT is used as a node, and a directed connection is made between the upper and lower layers to construct a graph network with a semantic hierarchy. The model not only pays attention to layer-by-layer semantics, but also pays attention to hierarchical relationships. Suitable for handling mixed sentiment classification tasks. The comparative experimental results show that the BHGCN model exhibits obvious competitive advantages.