• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2015.05a
• /
• pp.598-599
• /
• 2015

For the wireless mesh networks, efficient routing protocol is required to transmit data from source to destination node through multi-hop transmission. Multipath routing enable the stable data transmission even in the case of traffic congestion or unstable link. In this paper, we propose partial multipath routing which does not make a full multipath but partial for some parts of unstable links. This partial multipath routing reduce the number of nodes for routing and minimized the required power.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.19 no.8
• /
• pp.1917-1924
• /
• 2015

Efficient routing algorithm is required to transmit data from source to destination by multi-hop transmission in wireless sensor networks. In the multi-hop transmission, multipath routing can be one of the solutions to cope with the traffic congestion and unstable link condition. In this paper, we propose partial multipath routing which does not establish a secondary full routing path but a partial multipath to complement some poor links, and it can enable stable data transmission and reduce the number of nodes in routing path and the required total power compared with conventional multipath routing.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2008.10a
• /
• pp.655-658
• /
• 2008

In this paper, we explain the combined scheme of multi-code modulation and trans-orthogonal modulation. The name of the explained scheme is the multi-code trans-orthogonal modulation. Additionally, we propose a multi-code trans-biorthogonal modulation scheme, which is a little more improved multi-code trans-orthogonal modulation. We show the performance of the proposed schemes and compare the performance of the proposed schemes with the orthogonal modulation by using the computer simulation. According to the computer simulations, we can know that the bandwidth efficiency of the proposed schemes is better than that of the conventional orthogonal modulation. The proposed schemes can be used in physical layer of digital wireless telecommunication system to provide high speed data rate.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.8
• /
• pp.319-330
• /
• 2021

Recently, in the field of open domain natural language question answering, multi-task, multi-hop question answering has been studied extensively. In this paper, we propose a novel deep neural network model using hierarchical graphs to answer effectively such multi-task, multi-hop questions. The proposed model extracts different levels of contextual information from multiple paragraphs using hierarchical graphs and graph neural networks, and then utilize them to predict answer type, supporting sentences and answer spans simultaneously. Conducting experiments with the HotpotQA benchmark dataset, we show high performance and positive effects of the proposed model.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.48 no.1
• /
• pp.44-50
• /
• 2011

We propose a multi-class AdaBoost algorithm for en efficient classification of multi-class data in this paper. Traditional AdaBoost algorithm is basically a binary classifier and it has limitations when applied to multi-class data problems even though multi-class versions are available. In order to overcome the problems on the AdaBoost algorithm for multi-class classification problems, we devise an AdaBoost architecture with a training algorithm that utilizes multi-class classifiers for its weak classifiers instead of series of binary classifiers. Experiments on a image classification problem using collected Caltech Image Database are preformed. The results show that the proposed AdaBoost architecture can reduce its training time while maintaining its classification accuracy competitive when compared to Adaboost.M2.

• The KIPS Transactions:PartB
• /
• v.14B no.5
• /
• pp.337-342
• /
• 2007

Multi-level has advantage to express image in all levels with different images. This paper proposes digital watermarking built-in technique to transform color image to YCbCr color space to guarantee robustness and imperceptibility of the watermark in the various expression of color images, and to hide multi-level data which shows spread spectrum from low resolution to whole resolution for the Y-signal of multi-level. In color signal, Y-signal and low resolution built-in watermark has risk to be visible, but it can guarantee the robustness of watermark in various colors and transformed images. As a result of the experiment, wavelet compression image with built-in watermark showed robustness and imperceptibility of watermark.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.25 no.6
• /
• pp.523-528
• /
• 2015

In this paper, we propose a new multiresolutional and dynamic approach of the EM algorithm. EM is a very popular and powerful clustering algorithm. EM, however, has problems that indexes multiresolution data and requires a priori information on a proper number of clusters in many applications, To solve such problems, the proposed EM algorithm can impose a multiresolution kd-tree structure in the E-step and allocates a cluster based on sequential data. To validate clusters, we use a merge criteria for cluster merging. We demonstrate the proposed EM algorithm outperforms for texture image segmentation.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.5 no.4
• /
• pp.621-629
• /
• 2001

In this thesis a wireless data transmission system has been proposed and analysed that uses the multi-carrier technique with the adaptive modulation scheme. In general, the OFDM(Orthogonal Frequency Division Multiplexing) system is assigning a same amount of information to all sub-carriers in a wireless data transmission. In the proposed system, the different amount of information is assigned to each sub-carrier depending on the state of channel and the target probability of error of system. With the proposed scheme, the transmission rate can be maximized with the fixed power and the required power to transmit the information can be minimized with the target probability of error of system.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.05a
• /
• pp.36-39
• /
• 2019

데이터 마이닝에서 클러스터링은 서로 유사한 특징을 갖는 데이터들을 동일한 클래스로 분류하는 방법이다. 클러스터링에는 다양한 방법이 존재하지만 대표적으로 집합으로 표현된 데이터들의 유사도를 측정하기 위해서는 자카드 유사도(Jaccard Similarity)를 이용한다. 자카드 유사도는 서로 다른 집합 간의 공통된 부분을 상대적으로 평가하여 유사도를 측정하는 방법이다. 그러나 최근에는 데이터를 저장할 수 있는 기술과 매체의 발전으로 표현할 수 있는 데이터의 영역과 범위는 발전되고 있기 때문에 많은 연산과 시간의 비용이 발생하게 된다. 이를 해결하기 위해서 두 데이터의 표본의 유사도를 통해 실제 데이터들의 유사도를 추정할 수 있는 Min-Hash 가 제안되었다. 본 논문에서는 이를 활용하여 집합의 영역을 다중 집합(Multiset)으로 확장하여 중복되는 값을 가질 수 있는 두 데이터 간의 유사도를 효율적으로 추정할 수 있는 Min-Max Hash 를 제안한다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2005.11b
• /
• pp.217-219
• /
• 2005

최근 들어 무선 인터넷 및 모바일 기술이 급속한 발달을 이루면서 이동 객체의 위치에 기반 한 많은 서비스들이 개발되고 있다. 이 서비스에 사용되는 않은 어플리케이션들은 비교적 용량이 큰 공간 정보를 사용하여 최근에는 기존 디스크 기반 데이터베이스 관리 시스템이 제공할 수 있는 처리 속도보다 더욱 빠른 트랜잭션 처리를 요구하고 있다. 따라서 공간 데이터와 같은 대용량 데이터의 효율적인 처리와 폭주 하는 여러 사용자들에게 빠른 응답시간을 제공하여 주는 공간 DBMS가 요구되고 있다. 기존 디스크 기반의 공간 DBMS는 공간데이터와 같은 대용량의 데이터 관리가 가능하지만, 빠른 응답속도를 요구하는 여러 어플리케이션을 지원하기에는 무리가 있다. 반면에 메인 메모리 기반의 공간 DBMS는 불필요한 디스크 I/O를 없앰으로써 더욱 빠른 트랜잭션 처리를 지원하지만, 메인 메모리의 저장 한계로 대용량 처리에는 한계가 있다. 이러한 이유로 디스크 공간 DBMS의 장점과 메인 메모리 공간 DBMS의 장점으로 이루어진 다중레벨 공간 DBMS를 제안한다. 다중레벨 공간 DBMS는 디스크 기반의 공간 DBMS인 GMS시스템에 메인 메모리 데이터베이스와 그와 관련된 여러 컴포넌트들을 추가하여 개발 하였다. 제안된 시스템은 디스크 데이터베이스 기반의 대용량 데이터의 효율적인 관리와 메모리 데이터베이스 기반의 빠른 트랜잭션 처리를 보장한다.