• Proceedings of the Korean Information Science Society Conference
• /
• 2012.06a
• /
• pp.242-244
• /
• 2012

과학기술 문헌의 기술개체 인식에 관한 연구는 정보추출, 텍스트마이닝, 질의응답 분야 등의 선행 연구로서 다양한 통계적 방법론을 사용하여 기술개체 인식 정확률을 향상시키기 위해 연구되어 왔다. 하지만 기존의 연구는 단일-코어 또는 단일 머신 상에서 수행되었기 때문에, 폭발적으로 증가하는 문헌들에 대한 실시간 분석 요구를 처리할 수 없는 상황에 직면하고 있다. 이에 본 논문에서는 기술개체를 인식하는 과정에서 병목현상이 발생하는 작업을 "후보개체 추출 과정"의 언어처리 부분과 "개체 가중치 할당 과정"에서 통계정보를 취합하는 부분으로 분류하고, 각 작업을 하둡의 맵 작업과 리듀스 작업을 이용하여 해결하는 분산 병렬 처리 기반의 기술개체 인식 방법에 대해 살펴보고자 한다.

• 한국어정보학회:학술대회논문집
• /
• 2016.10a
• /
• pp.337-339
• /
• 2016

본 논문에서는 한국어 개체명 인식의 성능 향상을 위하여 워드 임베딩을 활용할 수 있는 방법에 대하여 기술한다. 워드 임베딩이란 문장의 단어의 공기정보를 바탕으로 그 단어의 의미를 벡터로 표현하는 분산표현이다. 이러한 분산 표현은 단어 간의 유의미한 정도를 계산하는데 유용하다. 본 논문에서는 이러한 워드 임베딩을 통하여 단어 벡터들의 코사인 유사도를 통한 개체명 사전 자가 학습 및 매칭 방법을 적용하고, 그 실험 결과를 보고한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.05a
• /
• pp.434-436
• /
• 2023

본 연구는 관세 데이터를 BERT 기반 모델을 활용한 개체명 인식(NER)모델을 제안한다. 관세 분야 국내 첫 시도이며, 선행연구들과 달리 개체명 인식에 초점을 맞춘다. 관세 관련 텍스트에서 고유한 의미의 개체를 인식하는 것이 주요 목표이다. 이 연구는 관세 분야의 개체명 인식에 대한 이해도를 높이고 향후 HS 코드 검색 시스템 개발에 대한 기초 연구를 제공한다.

• Journal of the Korean Society for information Management
• /
• v.27 no.4
• /
• pp.219-237
• /
• 2010

This study proposes a 'Bibliographic Universe Relationship Vocabulary'(burv) using the RDF/OWL Object Property under the SPO predicate logic according to the relationship type among all entities of bibliographic universe and implemented a 'relationship ontology system' to establish a new cataloging business domain called 'Relationship Description Cataloging' based on the ontology.

• Journal of KIISE:Computing Practices and Letters
• /
• v.15 no.8
• /
• pp.591-595
• /
• 2009

Most existing routing protocols for object detection and tracking in wireless sensor networks concentrate on finding ways to detect and track one and more individual objects, e.g. people, animals, and vehicles, but they do not be interested in detecting and tracking of continuous objects, e.g., poison gas and biochemical. Such continuous objects have quite different properties from the individual objects since the continuous objects are continuously distributed across a region and usually occupy a large area, Thus, the continuous objects could be detected by a number of sensor nodes so that sensing data are redundant and highly correlated. Therefore, an efficient data collection and report scheme for collecting and locally aggregating sensing data is needed, In this paper, we propose the Continuous Object Tracking Mechanism based on Dynamic Rectangle Zone for detecting, tracking, and monitoring the continuous objects taking into account their properties.

• The Journal of the Korea Contents Association
• /
• v.10 no.5
• /
• pp.29-36
• /
• 2010

Past clustering researches are focused on extraction of keyword for word similarity grouping. However, too many candidates to compare and compute bring high complexity, low speed and low accuracy. To overcome these weaknesses, this paper proposed a topical web document clustering model using not only keyword but also named entities such as person name, organization, location, and so on. By several experiments, we prove effects of our model compared with traditional model based on only keyword and analyze how different effects show according to characteristics of document collection.

• Journal of KIISE:Software and Applications
• /
• v.37 no.5
• /
• pp.363-369
• /
• 2010

Demonstration-based learning has the advantage that a user can easily teach his/her robot new task knowledge just by demonstrating directly how to perform the task. However, many previous demonstration-based learning techniques used a kind of attribute-value vector model to represent their state spaces and policies. Due to the limitation of this model, they suffered from both low efficiency of the learning process and low reusability of the learned policy. In this paper, we present a new demonstration-based learning method, in which the relational model is adopted in place of the attribute-value model. Applying the relational instance-based learning to the training examples extracted from the records of the user demonstrations, the method derives a relational instance-based policy which can be easily utilized for other similar tasks in the same domain. A relational policy maps a context, represented as a pair of (state, goal), to a corresponding action to be executed. In this paper, we give a detail explanation of our demonstration-based relational policy learning method, and then analyze the effectiveness of our learning method through some experiments using a robot simulator.

• Journal of KIISE
• /
• v.44 no.3
• /
• pp.306-313
• /
• 2017

Named entity recognition (NER) seeks to locate and classify named entities in text into pre-defined categories such as names of persons, organizations, locations, expressions of times, etc. Recently, many state-of-the-art NER systems have been implemented with bidirectional LSTM CRFs. Deep learning models based on long short-term memory (LSTM) generally depend on word representations as input. In this paper, we propose an approach to expand word representation by using pre-trained word embedding, part of speech (POS) tag embedding, syllable embedding and named entity dictionary feature vectors. Our experiments show that the proposed approach creates useful word representations as an input of bidirectional LSTM CRFs. Our final presentation shows its efficacy to be 8.05%p higher than baseline NERs with only the pre-trained word embedding vector.

• Journal of the Korea Society for Simulation
• /
• v.21 no.4
• /
• pp.81-90
• /
• 2012

Presented in this paper is a study for construction of a simulation system for the engagement of ground weapon systems. This paper proposes architecture for the simulation system based on agent simulation design methodology. Every entity of the proposed architecture is developed by assembling modularized agent components, and it enhances the reusability and composability of the entity. Consequently, time, costs, and efforts that are required to develop a new simulation system is able to be reduced by the enhancement. In the case of ground engagement simulation, it is very important to reflect environmental effects. Synthetic battlefield of the proposed architecture has environmental data of the battlefield and interacts with entities in the simulation system. The proposed architecture based simulation system can build swiftly various simulation models by the objectives and derive reasonable results from behaviors of entities that include environmental effects. This paper contains the construction of an example system based on the proposed architecture to verify the advantages of the architecture.

• Annual Conference on Human and Language Technology
• /
• 2008.10a
• /
• pp.17-22
• /
• 2008

본 논문에서는 영역별로 자동으로 개체명을 추출하기 위하여 씨앗단어를 이용하고, 웹페이지와 개체명 후보들 간의 상호 중요도를 측정하여 개체명 후보들의 순위를 정하는 방식을 제안한다. 제안된 방식은 크게 세 단계에 의해서 수행되어 지는데 먼저 씨앗단어 정보를 이용하여 웹페이지를 검색하고, 검색되어진 웹 페이지와 씨앗단어 정보를 이용하여 패턴 규칙을 추출한다. 추출된 패턴 규칙을 웹페이지에 적용하여 개체명 후보들을 추출하고 추출된 후보들과 웹페이지 사이의 상호 중요도를 재귀적으로 계산하여 최종적으로 개체명 후보들의 순위가 정해 진다. 한국어와 영어 개체명 영역에 제안된 기법을 적용하여 실험한 결과 한국어에서는 78.72%의 MAP를 얻을 수 있었고, 영어에서는 96.48%의 MAP를 얻었다. 특히 영어 개체명 인식에서의 성능은 구글에서 제공하고 있는 구글셋의 결과보다도 높은 성능을 보였다.