• The Journal of Information Technology
• /
• v.5 no.4
• /
• pp.155-164
• /
• 2002

MPLS is a integrated technology by using routing function and label swapping in the network layer. Based on the previous forwarding equivalence classes, it adds the fixed length label in ingress of the MPLS domain. For the routing, without the packet header information, it uses label for the forwarding decisions. In this paper, traffic engineering requirements in the MPLS internet will be setup. The traffic engineering function have to be performed previously with the network topology. In addition to, we presents the IP network topology and main function with MPLS signaling protocol.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2006.05a
• /
• pp.986-989
• /
• 2006

앞으로의 네트워크는 Convergence와 Integrate라는 키워드를 기반으로 IP 중심의 통합된 망으로 발전되고 있다. IP 중심의 망은 각각의 다양한 access 기술을 포함하며, voice, multimedia, 그리고 새롭게 정의되는 서비스가 모두 제공될 수 있어야 한다. 따라서 점차 인터넷을 중심으로 하는 하나의 통합된 망의 형태로 진화 될 것이다 이러한 차세대 네트워크상의 단말은 소형화, 지능화, 그리고 이동성을 지니고 있으며, 다양한 access 기술을 사용하기 위한 multiple 인터페이스를 가진 멀티호밍 환경에 놓여있다. 따라서 이 네트워크는 기본적으로 이동성과 멀티호밍을 지원해야 하는 것이다. 그러나 현재 인터넷의 핵심인 Internet Protocol 구조는 이를 지원하지 못한다. 현재 IP 주소는 최종 단말의 식별자 (Identifier)와 단말의 위치 식별자 (locator)의 의미를 함께 사용하고 있어, 통신 중인 단말이 이동하면 IP 주소도 변경되어 통신 중인 세션이 끊기는 문제가 발생한다. 멀티호밍 환경에서도 역시 통신 중인 노드들의 경로를 바꾸게 되면 세션이 끊기게 되는 문제가 발생한다. 본 논문은 이와 같은 린 구조의 근본적인 문제를 해결하기 위해 Identifier와 locator를 분리하며, 단순하게 단말에 스택으로 존재하는 L3SHIM을 소개하고, 모든 단말에 이 기능이 지원되었을 때 기존의 인터넷 프로토콜의 확장과 인터넷의 구조에 어떤 영향을 줄 수 있는지에 대해 선행 연구를 하였다.

• The Transactions of the Korea Information Processing Society
• /
• v.6 no.7
• /
• pp.1756-1769
• /
• 1999

Parallel paths in an interconnection network have some significance in that message transmission time can be reduced because message is divided into packets and transmitted in parallel through several paths, and also an whose nodes has 2n binary bit string, is an interconnection network which has a lower network cost than hypercube and its variation. In this paper, we analyze node disjoint parallel path in Folded Hyper-Star graph FHS(2n,n) proposed as the topology of parallel computers and, using the result, prove that the fault diameter of a Folded Hyper-Star graph FHS(2n,n) is 2n-1.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2018.06a
• /
• pp.322-325
• /
• 2018

영상의 관심 영역 검출은 영상처리 및 컴퓨터 비전 응용 분야에서 꾸준하게 사용되고 있는 기법이다. 특히, 근래 심층신경망 연구의 급격한 발전에 힘입어 심층신경망을 이용한 관심 영역 검출 기법에 대한 연구가 활발하게 진행되고 있다. 한편 Fully Convolutional Network(이하 FCN)은 본래 심층 예측(Dense Prediction)을 통한 의미론적 영상 분할(Semantic Segmentation)을 수행하기 위해 제안된 심층신경망 구조이다. FCN을 영상의 관심 영역 검출에 활용하여도 기존 관심 영역 검출 기법과 비교하여 충분히 좋은 성능을 발휘할 수 있다. 그러나 FCN에 사용되는 convolution 층의 수가 많고, 이에 따른 가중치(weight)의 개수도 기하급수적으로 늘어나 검출에 필요한 시간 복잡도가 매우 크다는 문제점이 있다. 따라서 본 논문에서는 기존 FCN이 가진 검출 시간 복잡도의 문제점을 convolution 층의 가중치 관점에서 해결하고자 이를 조절하여 FCN의 관심 영역 검출 속도를 향상시키는 방법을 제안한다. 적절한 convolution 층의 가중치를 조절함으로써, MSRA10K 데이터셋 환경에서 검출 정확도를 크게 저하시키지 않고도 최대 약 20.5%만큼 검출 속도를 향상시킬 수 있었다.

• Journal of Software Assessment and Valuation
• /
• v.15 no.1
• /
• pp.43-53
• /
• 2019

BigCloneBench has recently been used for performance evaluation of code clone detection tool using machine learning. However, since BigCloneBench is not a benchmark that is optimized for machine learning, incorrect learning data can be created. In this paper, we have shown through experiments using machine learning that the set of Type-4 clone methods provided by BigCloneBench can additionally be found. Experimental results using Tree-Based Convolutional Neural Network show that our proposed method is effective in improving BigCloneBench's dataset.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.10
• /
• pp.2461-2469
• /
• 1997

Target detection and recognition problems, in which neural networks are widely used, require translation invariant and real-time processing in addition to the requirements that general pattern recognition problems need. This paper presents a novel architecture that meets the requirements and explains effective methodology to train the network. The proposed neural network is an architectural extension of the shared-weight neural network that is composed of the feature extraction stage followed by the pattern recognition stage. Its feature extraction stage performs correlational operation on the input with a weight kernel, and the entire neural network can be considered a nonlinear correlation filter. Therefore, the output of the proposed neural network is correlational plane with peak values at the location of the target. The architecture of this neural network is suitable for implementing with parallel or distributed computers, and this fact allows the application to the problems which require realtime processing. Net training methodology to overcome the problem caused by unbalance of the number of targets and non-targets is also introduced. To verify the performance, the proposed network is applied to detection and recognition problem of a specific automobile driving around in a parking lot. The results show no false alarms and fast processing enough to track a target that moves as fast as about 190 km per hour.

• Journal of Digital Convergence
• /
• v.16 no.9
• /
• pp.11-20
• /
• 2018

The Q&A board is widely used as a means of communicating service enquiries, and the need for efficient management of the enquiry system has risen because certain questions are being repeatedly and frequently registered. This study aims to construct a student-centered FAQ, centered on the unstructured data posted on the university homepage's Q&A board. We extracted major keywords from 690 postings registered in the recent 3 years, and conducted the semantic network analysis to find the relationship between the keywords and the centrality analysis in order to carry out network visualization. The most central keywords found through the analysis, in order of centrality, were application, curriculum, credit point, completion, graduation, approval, period, major, portal, department. Also, the major keywords were classified into 8 groups of course, register, student life, scholarship, library, dormitory, IT and commute. If the most frequent questions are organized into these areas to form the FAQ, based on the results above, it is expected to contribute to user convenience and the efficiency of administration by simplifying the service enquiry process for repeated questions, as well as enabling smooth two-way communication among the members of the university.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.19 no.2
• /
• pp.89-103
• /
• 2020

As we face an aging society, the demand for personal mobility for disabled and aged people is increasing. In fact, as of 2017, the number of electric wheelchair in the country continues to increase to 90,000. However, people with disabilities and seniors are more likely to have accidents while driving, because their judgment and coordination are inferior to normal people. One of the causes of the accident is the interference of personal vehicle steering control due to unbalanced road surface conditions. In this paper, we introduce a encoder type semantic segmentation algorithm that can recognize road conditions at high speed to prevent such accidents. To this end, more than 1,500 training data and 150 test data including road surface damage were newly secured. With the data, we proposed a deep neural network composed of encoder stages, unlike the Auto-encoding type consisting of encoder and decoder stages. Compared to the conventional method, this deep neural network has a 4.45% increase in mean accuracy, a 59.2% decrease in parameters, and an 11.9% increase in computation speed. It is expected that safe personal transportation will be come soon by utilizing such high speed algorithm.

• Journal of the Korean Society for Library and Information Science
• /
• v.57 no.1
• /
• pp.93-114
• /
• 2023

Information extraction can facilitate the intensive analysis of documents by providing semantic triples which consist of named entities and their relations recognized in the texts. However, most of the research so far has been carried out separately for named entity recognition and relation extraction as individual studies, and as a result, the effective performance evaluation of the entire information extraction systems was not performed properly. This paper introduces two models of end-to-end information extraction that can extract various entity names in clinical records and their relationships in the form of semantic triples, namely pipeline and joint models and compares their performances in depth. The pipeline model consists of an entity recognition sub-system based on bidirectional GRU-CRFs and a relation extraction module using multiple encoding scheme, whereas the joint model was implemented with a single bidirectional GRU-CRFs equipped with multi-head labeling method. In the experiments using i2b2/VA 2010, the performance of the pipeline model was 5.5% (F-measure) higher. In addition, through a comparative experiment with existing state-of-the-art systems using large-scale neural language models and manually constructed features, the objective performance level of the end-to-end models implemented in this paper could be identified properly.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.23 no.6
• /
• pp.494-499
• /
• 2013

Based on the state-feedback adaptive neuro-control algorithm for a SISO nonaffine pure-feedback nonlinear system proposed in [15], an output-feedback controller is proposed in this paper. The output-feedback adaptive neural-net controller for the considered nonlinear system has not been previously proposed in any other literatures yet. The proposed output-feedback controller inherits all the advantages of [15] such that it does not adopt backstepping and this results in relatively simple control and adapting laws. Only one neural network is required for the proposed adaptive controller. The proposed neural-net control scheme expands the applicable class of nonlinear systems.