• Proceedings of the Korea Information Processing Society Conference
• /
• 2017.04a
• /
• pp.941-944
• /
• 2017

심층 학습에 기반을 둔 통계적 언어모형에서 가장 중요한 작업은 단어의 분산 표현(Distributed Representation)이다. 단어의 분산 표현은 단어 자체가 가지는 의미를 다차원 공간에서 벡터로 표현하는 것으로서, 워드 임베딩(word embedding)이라고도 한다. 워드 임베딩을 이용한 심층 학습 기반 통계적 언어모형은 전통적인 통계적 언어모형과 비교하여 성능이 우수한 것으로 알려져 있다. 그러나 워드 임베딩 역시 자료 부족분제에서 벗어날 수 없다. 특히 학습데이터에 나타나지 않은 단어(unknown word)를 처리하는 것이 중요하다. 본 논문에서는 고품질 한국어 워드 임베딩을 위하여 단어의 의미적 계층정보를 이용한 워드 임베딩 방법을 제안한다. 기존연구에서 제안한 워드 임베딩 방법을 그대로 활용하되, 학습 단계에서 목적함수가 입력 단어의 하위어, 동의어를 반영하여 계산될 수 있도록 수정함으로써 단어의 의미적 계층청보를 반영할 수 있다. 본 논문에서 제안한 워드 임베딩 방법을 통해 생성된 단어 벡터의 유추검사(analog reasoning) 결과, 기존 방법보다 5%가 증가한 47.90%를 달성할 수 있었다.

• Journal of the Korean Society of Clothing and Textiles
• /
• v.44 no.3
• /
• pp.413-426
• /
• 2020

This study researches the association between associative vocabulary and fashion image language in order to extract ideas that can be used as basic data for design ideas. Classic - avant-garde imaged language were chosen as theme words and each 70 questionnaires per a final image word were used for analysis. We obtained the following results by researching keywords that explained classic image words through a word cloud technique. It was found to have high central representation in the order of suit, classical, basic, music, Chanel, black and traditional. The core key words explaining avant-garde image language were found to have a central representation in the order of : peculiar, huge, Comme des Garçons, artistic, creative, deconstruction and individuality. We extracted the necessary idea dimensions needed for design ideas through associative network graph analysis. In the case of classical image language, it was named as the Mannish Item, Music, Modern Color, and the Traditional Classicality dimensions. In the case of avant-garde image language, it was named as the Key Image, Artistic Aura, Key Design and Designers dimensions.

• Education of Primary School Mathematics
• /
• v.25 no.4
• /
• pp.395-410
• /
• 2022

Nominalization is one of the grammatical metaphors, and it is the representation of verbal meaning through noun equivalent phrases. In mathematical word problems, texts using nominalization have both the advantage of clarifying the object to be noted in the mathematization stage, and the disadvantage of complicating sentence structure, making it difficult to understand the sentences and hindering the experience of the full steps in mathematical modelling. The purpose of this study is to analyze word problems in the textbooks from the perspective of nominalization, a linguistic element, and to derive implications in relation to students' difficulties during solving the word problems. To this end, the types of nominalization of 341 word problems from the content domain of 'Numbers and Operations' of elementary math textbooks according to the 2015 revised national curriculum were analyzed in four aspects: grade-band group, main class and unit assessment, specialized class, and mathematical expression required word problems. Based on the analysis results, didactical implications related to the linguistic expression of the mathematical word problems were derived.

• ETRI Journal
• /
• v.44 no.5
• /
• pp.794-804
• /
• 2022

Transformer models have had a great impact on natural language processing (NLP) in recent years by realizing outstanding and efficient contextualized language models. Recent studies have used transformer-based language models for various NLP tasks, including Persian named entity recognition (NER). However, in complex tasks, for example, NER, it is difficult to determine which contextualized embedding will produce the best representation for the tasks. Considering the lack of comparative studies to investigate the use of different contextualized pretrained models with sequence modeling classifiers, we conducted a comparative study about using different classifiers and embedding models. In this paper, we use different transformer-based language models tuned with different classifiers, and we evaluate these models on the Persian NER task. We perform a comparative analysis to assess the impact of text representation and text classification methods on Persian NER performance. We train and evaluate the models on three different Persian NER datasets, that is, MoNa, Peyma, and Arman. Experimental results demonstrate that XLM-R with a linear layer and conditional random field (CRF) layer exhibited the best performance. This model achieved phrase-based F-measures of 70.04, 86.37, and 79.25 and word-based F scores of 78, 84.02, and 89.73 on the MoNa, Peyma, and Arman datasets, respectively. These results represent state-of-the-art performance on the Persian NER task.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.15 no.7
• /
• pp.1517-1524
• /
• 2011

In this paper, a new document summarization method, which uses the semantic features and the pseudo relevance feedback (PRF) by using WordNet, is introduced to extract meaningful sentences relevant to a user query. The proposed method can improve the quality of document summaries because the inherent semantic of the documents are well reflected by the semantic feature from NMF. In addition, it uses the PRF by the semantic features and WordNet to reduce the semantic gap between the high level user's requirement and the low level vector representation. The experimental results demonstrate that the proposed method achieves better performance that the other methods.

• Journal of the Korea Society of Computer and Information
• /
• v.11 no.5 s.43
• /
• pp.115-125
• /
• 2006

A concept hierarchy represents the knowledge with multi-level form, which is very useful to categorize, store and retrieve the data. Traditionally, a concept hierarchy has been built manually by domain experts. However, the manual construction of a concept hierarchy has caused many problems such as enormous development and maintenance costs and human errors such as inconsistency. This paper proposes the automatic creation of concept hierarchies using the predefined hypernym relation. To create the hierarchy automatically, we first eliminate the ambiguity of the senses of data values, and construct the hierarchy by grouping and leveling of the remaining senses. We use the WordNet explanations for multi-meaning word to eliminate the ambiguity and use the WordNet hypernym relations to create multi-level hierarchy structure.

• The Journal of Society for e-Business Studies
• /
• v.26 no.3
• /
• pp.119-132
• /
• 2021

Identification of associated knowledge based on content relevance is a fundamental functionality in managing service and security of core knowledge. This study compares the performance of methods to identify associated knowledge based on content relevance, i.e., the associated document network composition performance of keyword-based and word-embedding approach, to examine which method exhibits superior performance in terms of quantitative and qualitative perspectives. As a result, the keyword-based approach showed superior performance in core document identification and semantic information representation, while the word embedding approach showed superior performance in F1-Score and Accuracy, association intensity representation, and large-volume document processing. This study can be utilized for more realistic associated knowledge service management, reflecting the needs of companies and users.

• School Mathematics
• /
• v.11 no.3
• /
• pp.497-511
• /
• 2009

This study compared two problem solving instructional approaches, schema based sequence instruction and schema based parallel instruction on word problem solving performance of elementary school students who were in general students group. The subjects totaled 48 third grade students who were exposed to a test that consisted of 9 word problem items of three types for 4 sessions. First of all, the baseline of word problem performance level was measured without any training. During session 1, 2 and 3 participants were put into strategic training groups. The experiment was designed by two between factor(two intervention group and two within factors(two problem types, three sessions). The results of experiment were as follows. Schema based sequence instruction group performed significantly better than students in another group on word problem solving performance. The effect of strategic schema based Instruction revealed that solving word problems relied upon problem types, sessions and input orders which were of great value.

• 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.

• Education of Primary School Mathematics
• /
• v.16 no.3
• /
• pp.285-301
• /
• 2013

The purpose of the study is to analyze the 4th graders' visual representation in mathematics problem solving process and to find out how to teach the visual representation in mathematics problem solving process. on the basis of the results, this study gives several pedagogical implication related to the mathematics problem solving. The following were the conclusions drawn from the results obtained in this study. First, The achievement level of students and using visual representation in the mathematics problem solving are closely connected. High achieving students used visual representation in the mathematics problem solving process more frequently. Second, high achieving students realize the usefulness of visual representation in the mathematics problem solving process and use visual representation to solve mathematical problem. But low achieving students have no conception that visual representation is one of the method to solve mathematical problem. Third, students tend to especially focus on 'setting up an equation' when they solve a mathematical problem. Because they mostly experienced mathematical problems presented by the type of 'word problem-equation-answer'. Fourth even through students tried visual representation to solve a mathematical problem, they could not solve the problem successfully in numerous instances. Because students who face a difficulty in solving a problem try to construct perfect drawing immediately. But generating visual representation 2)to represent mathematical problem cannot be constructed at one swoop.