• Annual Conference on Human and Language Technology
• /
• 2019.10a
• /
• pp.3-7
• /
• 2019

자연어 표상은 자연어가 가진 정보를 컴퓨터에게 전달하기 위해 표현하는 방법이다. 현재 자연어 표상은 학습을 통해 고정된 벡터로 표현하는 것이 아닌 문맥적 정보에 의해 벡터가 변화한다. 그 중 BERT의 경우 Transformer 모델의 encoder를 사용하여 자연어를 표상하는 기술이다. 하지만 BERT의 경우 학습시간이 많이 걸리며, 대용량의 데이터를 필요로 한다. 본 논문에서는 빠른 자연어 표상 학습을 위해 의미 정보와 BERT를 결합한 개념 언어 모델을 제안한다. 의미 정보로 단어의 품사 정보와, 명사의 의미 계층 정보를 추상적으로 표현했다. 실험을 위해 ETRI에서 공개한 한국어 BERT 모델을 비교 대상으로 하며, 개체명 인식을 학습하여 비교했다. 두 모델의 개체명 인식 결과가 비슷하게 나타났다. 의미 정보가 자연어 표상을 하는데 중요한 정보가 될 수 있음을 확인했다.

• Proceedings of the Korean Society for Cognitive Science Conference
• /
• 2002.05a
• /
• pp.59-64
• /
• 2002

인지심리학 및 인지언어학 분야에서 시도한 어휘 표상, 특히 움직임과 관련된 동사의 인지도식에 관한 연구들을 비교해보고자 한다. 인간의 언어학적인 지식을 도식적으로 표상 하고자 하는 노력은 언어의 통사적인 외형에만 치중하는 연구에서는 언어의 의미구조를 파악하기 힘들다고 판단하고 의미적인 범주화를 중요시하게 되었다. 본 연구에서는 시각적 이미지 도식을 중점적으로 살펴보기로 한다. 이미지 도식은 공간적 위치 관계, 이동, 형상 등에 관한 지각과 결부되어 있다. 이미지로 나타낸 표상은 근본적으로 세상의 인식과 세상에 대한 행동방법을 사용하게 하는 유추적이고 은유적인 원칙에 기초하고 있다. 이러한 점에 있어서, 언술을 발화한 화자는 어느 정도 주관적인 행동의 능력과 그가 인식한 개념화에서부터 문자화시킨 표상을 구성한다. 인지 원칙에 입각한 의미 표상에 중점을 둔 도식으로는, Langacker, Lakoff, Talmy의 도식이 있다. 프랑스에서 톰 R. Thom과 같은 수학자들은 질적인 현상에 관심을 가져 형역학(morphodynamique)이론을 확립하였는데, 이 이론은 요즘의 인지 연구에 수학적 기초를 제공하였다. R. Thom, J. Petitot-Cocorda의 도식 및 구조 의미론의 창시자라고 불리는 B.Pottier의 도식이 여기에 속한다 J.-P. Descles가 제시한 인지연산문법(Grammaire Applicative et Cognitive)은 다른 인지문법과는 달리 정보 자동처리과정에서 사용할 수 있는 연산자와 피연산자의 관계에 기초한 수학적 연산작용을 발전시켰다. 동사의 의미는 의미-인지 도식으로 설명되는데, 이것은 서로 다른 연산자와 피연산자로 구성된 형식화된 표현이다. 인간의 인지 기능은 언어로 표현되며, 언어는 인간의 의사소통, 사고 행위 및 인지학습의 핵심적 기능을 담당한다. 인간의 언어정보처리 메카니즘은 매우 복잡한 과정이기 때문에 언어정보처리와 관련된 언어심리학, 인지언어학, 형식언어학, 신경해부학 및 인공지능학 등의 관련된 분야의 학제적 연구가 필요하다.

• Journal of the Korean School Mathematics Society
• /
• v.17 no.4
• /
• pp.507-539
• /
• 2014

The algebra is an important tool influencing on a mathematics in general. To make good use of the algebra, it is necessary to transfer from a given situation to a proper algebraic representation. But some research in related to algebraic word problems have reported the difficulty changing to a proper algebraic representation. Our study have focused on transformation and elaboration of algebraic representation. We investigated in detail the responses and perceptions of 29 Grade 7 students while transforming to algebraic representation, only concentrating on the literature expression form the problematic situations given. Most of students showed difficulties in transforming both descriptive and geometric problems to algebraic representation. 10% of them responded wrong answers except only a problem. Four of them were interviewed individually to show their thinking and find the factor influencing on a positive elaboration. As results, we could find some characteristics of their thinking including the misconception that regard the problem finding a functional formula because there are the variables x and y in the problematic situation. In addition, we could find the their fixation which student have to set up the equation. Furthermore we could check that making student explain own algebraic representation was able to become the factor influencing on a positive elaboration. From these, we also discussed about several didactical implications.

• Annual Conference on Human and Language Technology
• /
• 2019.10a
• /
• pp.272-275
• /
• 2019

단어 중의성 해소 방법은 지식 정보를 활용하여 문제를 해결하는 지식 기반 방법과 각종 기계학습 모델을 이용하여 문제를 해결하는 지도학습 방법이 있다. 지도학습 방법은 높은 성능을 보이지만 대량의 정제된 학습 데이터가 필요하다. 반대로 지식 기반 방법은 대량의 정제된 학습데이터는 필요없지만 높은 성능을 기대할수 없다. 최근에는 이러한 문제를 보완하기 위해 지식내에 있는 정보와 정제된 학습데이터를 기계학습 모델에 학습하여 단어 중의성 해소 방법을 해결하고 있다. 가장 많이 활용하고 있는 지식 정보는 상위어(Hypernym)와 하위어(Hyponym), 동의어(Synonym)가 가지는 의미설명(Gloss)정보이다. 이 정보의 표상을 기존의 문장의 표상과 같이 활용하여 중의성 단어가 가지는 의미를 파악한다. 하지만 정확한 문장의 표상을 얻기 위해서는 단어의 표상을 잘 만들어줘야 하는데 기존의 방법론들은 모두 문장내의 문맥정보만을 파악하여 표현하였기 때문에 정확한 의미를 반영하는데 한계가 있었다. 본 논문에서는 의미정보와 문맥정보를 담은 단어의 표상정보를 만들기 위해 구문정보, 의미관계 그래프정보를 GCN(Graph Convolutional Network)를 활용하여 임베딩을 표현하였고, 기존의 모델에 반영하여 문맥정보만을 활용한 단어 표상보다 높은 성능을 보였다.

• Journal of the Korean Chemical Society
• /
• v.56 no.6
• /
• pp.759-767
• /
• 2012

The purpose of this study was to examine the impact of Science Writing Heuristic (SWH) on multiple representations in students' writing and to survey experimental group students' recognition about the use of multiple representations. For this study, Participants of this study were 158 students in 7th grade. 94 students were assigned to the experimental group and 64 students were assigned to the comparative group. The experimental group showed significantly higher mean score than comparative group at utilizing multiple representation in summary writing. Interview analysis indicated that all students who participated in interviews, regardless of solid multi-modal competency, recognized that use of multiple representations with appropriate explanations enable to communicate science information persuasively.

• Journal of the Korean Chemical Society
• /
• v.64 no.1
• /
• pp.19-29
• /
• 2020

We analyzed the representations of acid-base models in 4 kinds of Chemistry I and 4 kinds of Chemistry II textbooks of the 2009 revised curriculum, and 9 kinds of Chemistry I textbooks and 6 kinds of chemistry II textbooks of the 2015 revised curriculum in this study. The problems of the textbook were divided into the problems of definitions and the representations of the logical thinking. As a result of the study, the lack of the concept of chemical equilibrium had a problem with the representation of reversible reactions in the definition of the Brønsted-Lowry model in the Chemistry I textbooks of 2009 revised curriculum, it also appeared to persist in Chemistry I textbooks of 2015 revised curriculum which contains the concept of chemical equilibrium. The representations of logical thinking were related to particle kinds of conservation logic, combinational logic, particle number conservation logic, and proportion logic. There were few problems related to representation of logical thinking in Chemistry I textbook in 2009 revision curriculum, but more problems of representations related to logics are presented in Chemistry I textbooks in 2015 revision curriculum. Therefore, as the curriculum is revised, the representations of chemistry textbooks related to acid and base models need to be changed in a way that can help students' understanding.

• Journal of the Korean Chemical Society
• /
• v.51 no.5
• /
• pp.423-432
• /
• 2007

In this study we investigated the types of inscriptions representing the concept of atom and molecule in the 9th grade science textbooks developed under the 7th National Curriculum, and analyzed the processes of interpreting those inscriptions on the view of semiotics. The concept of atom and molecule was represented in various ways such as circular diagram, matters of everyday life, ball-and-stick, and STM graphic, etc. We identified two kinds of inscriptions on the basis of the possibilities of handling models: the inscriptions that represent concept of atom and molecule by image; the inscriptions that represent concept of atom and molecule by using specific matters. We analyzed the processes of interpreting inscriptions systematically with a semiotic model that includes ‘structuring processes' and ‘translating processes'. The interpretation of inscriptions representing concept of atom and molecule was a complex process requiring many steps of interpreting works. Educational implications of these findings are discussed.

• Communications of Mathematical Education
• /
• v.24 no.1
• /
• pp.49-65
• /
• 2010

The purpose of this study is to encourage the role of domain to consider the teaching of the concept of functions in modeling real situations. To do this, it is analyzed that how to introduce the concept of functions and linear functions in textbooks treated in the 1st grade and the 2nd grade of middle school. This study also reviewed the role of domain in representational activities for modeling real situations using linear functions. In these reviews, it found that many textbooks do not consider the domain in the equations of functions and these graphs and several text books used linear functions for modeling real situations which are not represented by linear functions contextually. It is concluded that the domain of function is an important concept that will be considered any representational activities for functions.

• Korean Journal of Culture and Social Issue
• /
• v.19 no.3
• /
• pp.507-528
• /
• 2013

The study examined the mental representation of children in foster care through the MacArthur Story Stem Battery (MSSB) and reviewed difference between children in foster care and typical children in terms of mental representation. Subjects of the study included a total of 50 children: 25 children (11 boys and 14 girls), aged from 5 to 8, in foster care in Gyeonggi Province, and 25 typical children (11 boys and 14 girls), aged from 5 to 8, who lived in Seoul and Gyeonggi Province. The research findings are summarized as follows: First, in terms of mental representation, children in foster care showed less subjects of empathic relations and more subjects of disorganized aggression than typical children. Second, children in foster care showed lower coherence mental representation than typical children. Finally, children in foster care were more likely to fail in emotional regulation than typical children in mental representation.

• Science of Emotion and Sensibility
• /
• v.26 no.1
• /
• pp.33-42
• /
• 2023

Previous emotion studies employing facial expressions have focused on the differences between age groups for each of the emotion categories. Instead, Kim (2021) has compared representations of facial expressions in the lower-dimensional emotion space. However, he reported descriptive comparisons without statistical significance testing. This research used representational similarity analysis (Kriegeskorte et al., 2008) to directly compare empirical datasets from young, middle-aged, and old groups and conceptual models. In addition, individual differences multidimensional scaling (Carroll & Chang, 1970) was conducted to explore individual weights on the emotional dimensions for each age group. The results revealed that the old group was the least similar to the other age groups in the empirical datasets and the valence model. In addition, the arousal dimension was the least weighted for the old group compared to the other groups. This study directly tested the differences between the three age groups in terms of empirical datasets, conceptual models, and weights on the emotion dimensions.