• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.17 no.12
• /
• pp.1465-1474
• /
• 1992

This paper describes a new overlapped object recognition method using extended local features. At first, we extract the extended local features consisting of corners, arcs, parallel-lines, and corner-arcs from the images consisting of model objects. Based on the extended local features we construct a knowledge-base. In order to match objects, we also extract the extended local features from the input image, and then check the compatibility between the extracted features and the features in the knowledge-base. From the set of compatible features, we compute geometric transforms. If any geometric transforms are clustered, we generate the hypothesis of the objects as the centers of the clusters, and then verify the hypothesis by a reverse geometric transform. An experiment shows that the proposed method increases the recognition rate and the accuracy as compared with existing methods.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.12 no.3
• /
• pp.124-131
• /
• 1998

This paper proposed the alarm processing system to improve the efficiency of monitoring method by applying the neural network and troubleshooting knowledge base for IADAPS(Intelligent Alarm Diagnosis And Processing System) method in an receiving and distributing board of Building complex. This IADAPS is abased on the cumulative generalized delta rule of backpropagation in neural network. It was used to infer the minimum alarms among multi-fired alarms, and the inferred alarm can be displayed maintenance information of facility by using a pre-defined troubleshoot knowledge base. For validating the proposed monitoring method, he method of simulation used to the five case of virtual scenario. As comparison results, a proposed method in this paper could be proved the applied possibility of an neural network and utilized in fields of facilities maintenance, if needed, be operated by non-expertise.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2008.11a
• /
• pp.52-55
• /
• 2008

The paper aims at showing the subsumption relations of the Wordnet can be captured automatically from a dynamic encyclopedia such as Wikipedia with a meaningful precision. The idea behind the proposal is that a knowledge base in the form of Wordnet can be dynamically obtained and maintained accordingly to the online dictionaries so that the scalability of knowledge base construction may be achieved to some degree. To show the plausibility of dynamic ISA construction, we have tested how well the ISA relations among the 100 technology terms selected from the Wordnet can be saved from the ISA construction by the wiki definitions of the selected terms. As a result the wiki definition led to the ISA relations of the Wordnet with the precision of 80%.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.15 no.3
• /
• pp.691-699
• /
• 2011

Recently, diverse approaches retrieval services based on social network are suggested. Although existing recommendation systems can retrieve experts of specific fields, profiles and evaluations about experts that users want to be recommended are in a system. The proposed expert recommendation system can automatize collection of evaluation to evaluate experts and experts' profiles in separate systems by using the Knowledge Base and XMDR. We also attempt to construct system which can recommend a number of experts by dynamically constructing Social Network by using diverse resources distributed 로컬ly and composed of heterogeneous data sources. To resolve these problems efficiently, there is a need to provide constructed resources between heterogeneous systems with transparency and independence and provide users with a singular interface. Therefore, the proposed system in this paper uses Knowledge Base and XMDR for extracting distributed experts' profiles and designs expert recommendation system connecting Knowledge Base with Social Network.

• KIPS Transactions on Software and Data Engineering
• /
• v.5 no.11
• /
• pp.549-554
• /
• 2016

ConceptNet is a common sense knowledge base which is formed in a semantic graph whose nodes represent concepts and edges show relationships between concepts. As it is difficult to make knowledge base integrity, a knowledge base often suffers from incompleteness problem. Therefore the quality of reasoning performed over such knowledge bases is sometimes unreliable. This work presents neural tensor networks which can alleviate the problem of knowledge bases incompleteness by reasoning new assertions and adding them into ConceptNet. The neural tensor networks are trained with a collection of assertions extracted from ConceptNet. The input of the networks is two concepts, and the output is the confidence score, telling how possible the connection between two concepts is under a specified relationship. The neural tensor networks can expand the usefulness of ConceptNet by increasing the degree of nodes. The accuracy of the neural tensor networks is 87.7% on testing data set. Also the neural tensor networks can predict a new assertion which does not exist in ConceptNet with an accuracy 85.01%.

• Annual Conference on Human and Language Technology
• /
• 2010.10a
• /
• pp.177-178
• /
• 2010

단어 의미 중의성은 자연언어처리 분야에서의 주요 관심 분야이다. 한국어에서의 단어 의미 중의성 문제는 다른 언어에 비하여 연구가 미흡한 상태이다. 기존 연구에서는 빈도 수에 기반한 공기 정보 벡터를 이용한 방법에서 처리되지 못하는 경우가 발생하였다. 또한 사전에 기반한 상위어 추출 시에 정형화된 형태가 아닌 경우에 어려움이 발생하였다. 본 논문에서는 상호정보량을 추가하여 공기 정보 처리 과정 시에 발생하는 오류를 최소화 하였다. 또한 대상 명사의 상위어 추출 문제를 해결하기 위해 어휘 지식 베이스를 적용하였다.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 1993.04b
• /
• pp.242-246
• /
• 1993

설계경험이 적은 설계자를 지원하기 위하여, 범용 전문가시스템 쉘인Nexpert을 이용하여, 선체구조설계를 위한 전문가 시스템을 구현하였다. 지식베이스를 구축하기 위하여, 선급협회의 규정집과 실제로 설계실무에 이용되고 있는 기존의 선박구조 설계프로그램으로 부터 지식을 추출하였으며, 객체지향 개념을 이용하여 이 지식들을 조직 화 하였다. 또한, 공학설계 작업을 위해 추가로 요구되는 기능들을 접속하여 시스템을 구성하였다. 추가된 기능은, 종강도의 계산, 실적선 데이타베이스, 그래픽 사용자 인터페이스와 설계결과를 가시화하는 부분이다. 이 연구를 통해 범용전문가 시스템이 선박설계에 이용될수 있는 가능성을 보였으며, 그 추구해야할 방향을 설정할수 있었다. 특히, 어떤 설계결정에 이르는데 관여한 규칙들을 보여주는 기능의 유용성과, 개발된 시스템이 자주 개정되는 선급규정에 맞추어 쉽게 수정될수 있음이 관찰되었다.

• Annual Conference on Human and Language Technology
• /
• 2021.10a
• /
• pp.582-587
• /
• 2021

한국어 통합 지식베이스를 생성하기 위해 필수적인 분류체계(taxonomy)를 구축하는 방식을 제안한다. 위키데이터를 기반으로 분류 후보군을 추출하고, 상하위 관계를 통해 방향 비순환 그래프(Directed Acyclic Graph)를 구성한 뒤, 국부적 도달 중심성(local reaching centrality) 등의 정보를 활용하여 정제함으로써 246 개의 분류와 314 개의 상하위 관계를 갖는 분류체계를 생성한다. 워드넷(WordNet), 디비피디아(DBpedia) 등 기존 링크드 오픈 데이터의 분류체계 대비 깊이 있는 계층 구조를 나타내며, 다중 상위 분류를 지닐 수 있는 비트리(non-tree) 구조를 지닌다. 또한, 위키데이터 속성에 기반하여 위키데이터 정보가 있는 인스턴스(instance)에 자동으로 분류를 부여할 수 있으며, 해당 방식으로 실험한 결과 99.83%의 분류 할당 커버리지(coverage) 및 99.81%의 분류 예측 정확도(accuracy)를 나타냈다.

• Journal of Internet Computing and Services
• /
• v.15 no.3
• /
• pp.125-135
• /
• 2014

In the medical field, for the purpose of diagnosis and treatment of diseases, building knowledge base has received a lot of attention. The most important thing to build a knowledge base is representing the knowledge accurately. In this paper we suggest a knowledge representation method using Ontology technique with the datasets obtained from the domestic papers on Alzheimer disease that has received a lot of attention recently in the medical field. The suggested Ontology for Alzheimer disease defines all the possible classes: lexical information from journals such as 'author' and 'publisher' research subjects extracted from 'title', 'abstract', 'keywords', and 'results'. It also included various semantic relationships between classes through the Ontology properties. Inference can be supported since our Ontology adopts hierarchical tree structure for the classes and transitional characteristics of the properties. Therefore, semantic representation based query is allowed as well as simple keyword query, which enables inference based knowledge query using an Ontology query language 'SPARQL'.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2001.12a
• /
• pp.208-212
• /
• 2001

This paper is for the optimized management of the knowledge abstracted from the World-Wide Web(WWW) in which we assume the infinite knowledge-base. Though we can abstract various useful knowledge such as the facts and the rules from the WWW pages, they may include many noisy knowledge. Therefore we have to reasonably reject them from the knowledge-base which is composed of knowledge abstracted from the WWW. To do this, we propose the oblivious memorization concept. This concept is characterized by the memorization based on the oblivion mechanism of human being. We assume the memorization is the function of the concern for any knowledge, oblivion ability and time. That is, the more concern for my knowledge the ore memorizable. And, the more oblivious and the more tine spent the less memorizable by exponentially. Where, tie assume the oblivion is the function of the degree of previous memorization, memorization ability md the number of knowledge stimulation. That is, the more previously memorized, the greater memorizing ability and the more frequently stimulated by any knowledge the less knowledge oblivious.