• Annual Conference on Human and Language Technology
• /
• 2002.10e
• /
• pp.176-183
• /
• 2002

전자우편과 같이 일정한 질의 형식을 가지고 있는 긴 자연어 질의에 대해서 사용자 질의 단어에 가중치를 부과하는 방법과 질의에 대한 정답을 기존의 질의응답 집합에서 유사한 질의를 검색하여 그 정답을 사용자에게 제공하는 전자우편 질의응답 시스템을 제안한다. 사용자의 긴 자연어 질의가 주어지면 질의의 범주와 문장의 중요도 정보를 이용하여 질의에서 사용된 단어가 주제어로 쓰였을 확률을 계산하고, 계산된 확률에 기반하여 중요도를 할당하는 질의생성 모델을 제안한다. 또한 사용자 질의와 기존에 문의되어진 전자우편 질의의 유사도를 단어의 빈도를 고려한 어휘유사도, 한글 시소러스(Thesaurus)를 이용한 의미유사도와 본 논문에서 제안한 질의생성 모델을 이용한 주제 유사도를 이용하여 계산한다. 실험을 위하여 실세계에서 사용 중인 질의응답 집합을 이용하여 실험을 하였으며 각 유사도 계산 방법의 기여도를 비교 평가하고 제안한 질의생성모델이 성능향상에 미치는 영향을 평가하였다.

• Annual Conference of KIPS
• /
• 2015.10a
• /
• pp.1579-1580
• /
• 2015

본 논문에서는 3차원 메쉬 모델의 중요 영역을 표현하는 메쉬 돌출맵(mesh saliency map)을 생성하기 위하여 다중 스케일 평균 곡률 (multi-scale mean curvature)을 기반으로 정의된 전역 희소치(global rarity)를 이용하는 방법을 제안한다. 제안 방법에서는 우선, 메쉬 모델의 지역 영역 특성을 정의하기 위하여 기존 관련 연구들에서 많이 사용하고 있는 가우시안 가중치 평균곡률(Gaussian-weighted mean curvature)을 5단계 서로 다른 스케일에서 정의하고, 메쉬의 각 정점(vertex)에 대하여 중심주변 연산자(center-surround operator)를 적용하여 5단계 지역 돌출특성(local saliency)을 정의한다. 주어진 메쉬 모델의 전역 희소치를 구하기 위하여 메쉬의 모든 정점쌍 (vertex pair)에 대하여 5단계 지역 돌출 특성 공간에서의 거리를 계산하고, 각 정점별로 5단계 지역 돌출 특성 공간에서의 다른 정점과의 거리의 합으로 전역 희소치를 정의한다. 이러한 전역 희소치를 각 정점의 메쉬 돌출치로 정의한다. 서로 다른 형태의 3차원 모델에 대하여 제안방법에 의한 메쉬 돌출맵과 지역 특성만을 고려한 기존 메쉬 돌출맵을 생성하여 중요 영역 표현 결과를 비교 분석한다.

• The Journal of the Acoustical Society of Korea
• /
• v.40 no.2
• /
• pp.130-138
• /
• 2021

This paper presents an effective structure by applying various normalization to Convolutional Neural Networks (CNN) for seismic event classification. Normalization techniques can not only improve the learning speed of neural networks, but also show robustness to noise. In this paper, we analyze the effect of input data normalization and hidden layer normalization on the deep learning model for seismic event classification. In addition an effective model is derived through various experiments according to the structure of the applied hidden layer. As a result of various experiments, the model that applied input data normalization and weight normalization to the first hidden layer showed the most stable performance improvement.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.16 no.1
• /
• pp.75-80
• /
• 2021

Artificial intelligence is becoming an important part of our lives providing incredible benefits. In this respect, facial expression recognition has been one of the hot topics among computer vision researchers in recent decades. Classifying small dataset of low resolution images requires the development of a new small scale deep CNN model. To do this, we propose a method suitable for small datasets. Compared to the traditional deep CNN models, this model uses only a fraction of the memory in terms of total learnable weights, but it shows very similar results for the FER2013 and FERPlus datasets.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.05a
• /
• pp.635-637
• /
• 2021

Cybersecurity assessment is the process of assessing the risk level of a system through threat and vulnerability analysis to take appropriate security measures. Accurate security evaluation models are needed to prepare for the recent increase in cyberattacks and the ever-developing intelligent security threats. Therefore, we present a risk assessment model through a matrix-based security assessment model analysis that scores by assigning weights across security equipment, intervals, and vulnerabilities. The factors necessary for cybersecurity evaluation can be simplified and evaluated according to the corporate environment. It is expected that the evaluation will be more appropriate for the enterprise environment through evaluation by security equipment, which will help the cyber security evaluation research in the future.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2023.07a
• /
• pp.493-494
• /
• 2023

본 논문은 벚꽃나무 영상 데이터를 활용하여 벚꽃의 상태(개화, 만개, 낙화)를 실시간으로 분류하는 연구를 소개한다. 이 연구의 목적은, 실시간으로 취득되는 벚꽃나무의 영상 데이터를 사전에 학습된 CNN 기반 이미지 분류 모델을 통해 벚꽃의 상태에 따라 분류하는 것이다. 약 1,000장의 벚꽃나무 이미지를 활용하여 CNN 모델을 학습시키고, 모델이 새로운 이미지에 대해 얼마나 정확하게 벚꽃의 상태를 분류하는지를 평가하였다. 학습데이터는 훈련 데이터와 검증 데이터로 나누었으며, 개화, 만개, 낙화 등의 상태별로 폴더를 구분하여 관리하였다. 또한, ImageNet 데이터셋에서 사전 학습된 ResNet50 가중치를 사용하는 전이학습 방법을 적용하여 학습 과정을 더 효율적으로 수행하고, 모델의 성능을 향상시켰다.

• Journal of the Korean Society for Library and Information Science
• /
• v.47 no.4
• /
• pp.335-361
• /
• 2013

This study tries to develop the so-called LCSI (Library Customer Satisfaction Index) Lite for public library which can be used easily in the fields as a simplified model. Its conceptual model is developed from the former research about library service quality and customer satisfaction. Based on the analyses on the twenty Delphi-type AHP questionnaires both from the library researchers and from the public library practitioners, and on the analysis and testing of the structural equation model using the data from the more than 800 user questionnaires, it suggests a model which consists of total 15 items - namely 8 items for service quality (2 for personnel, 4 for library resources and services, 2 for facilities and environment), 3 items for customer satisfaction, and 2 items for loyalty. Library Satisfaction Index for public library will be calculated by the sum total of service quality (50%), customer satisfaction (40%), and loyalty (10%).

• The KIPS Transactions:PartA
• /
• v.12A no.1 s.91
• /
• pp.13-22
• /
• 2005

Similarly to the edges defined in a 2D image, we can define the geometric features representing the boundary of the distinctive parts appearing on 3D meshes. The geometric features have been used as basic primitives in several applications such as mesh simplification, mesh deformation, and mesh editing. In this paper, we propose geometric livewire and geometric livelane for extracting geometric features in a 3D mesh, which are the extentions of livewire and livelane methods in images. In these methods, approximate curvatures are adopted to represent the geometric features in a 3D mesh and the 3D mesh itself is represented as a weighted directed graph in which cost functions are defined for the weights of edges. Using a well-known shortest path finding algorithm in the weighted directed graph, we extracted geometric features in the 3D mesh among points selected by a user. In this paper, we also visualize the results obtained from applying the techniques to extracting geometric features in the general meshes modeled after human faces, cows, shoes, and single teeth.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.17 no.6
• /
• pp.1399-1408
• /
• 2013

This paper proposes a new $H_{\infty}$ T-S fuzzy controller with a novel integral control for induction motors which have nonlinear dynamics. The $H_{\infty}$ T-S fuzzy controller is used for the nonlinearity and robustness and weighted integral is used for tracking problem and control performance. A T-S Fuzzy controller is the fuzzy combination of local linear controllers considering the overall stability, and LMI(Linear Matrix Inequlity) is used for determining the gains of linear controllers. The tracking problem of an induction motor is changed into regulator problem by introducing the integral control technique with weighting factor, diminishing the conservatism of $H_{\infty}$ T-S fuzzy controller.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.21 no.11
• /
• pp.2133-2138
• /
• 2017

The analysis of existing XML documents and other documents was centered on words. It can be implemented using a morpheme analyzer, but it can classify many words in the document and cannot grasp the core contents of the document. In order for a user to efficiently understand a document, a paragraph containing a main word must be extracted and presented to the user. The proposed system retrieves keyword in the normalized XML document. Then, the user extracts the paragraphs containing the keyword inputted for searching and displays them to the user. In addition, the frequency and weight of the keyword used in the search are informed to the user, and the order of the extracted paragraphs and the redundancy elimination function are minimized so that the user can understand the document. The proposed system can minimize the time and effort required to understand the document by allowing the user to understand the document without reading the whole document.