• Journal of Korea Multimedia Society
• /
• v.6 no.1
• /
• pp.136-147
• /
• 2003

To achieve an optimal performance for a given problem, we need an integrative process of the parameter optimization via learning and the structure optimization via model selection. In this paper, we propose an efficient optimization method for improving generalization performance by considering the property of each sub-method and by combining them with common theoretical properties. First, weight parameters are optimized by natural gradient teaming with adaptive regularization, which uses a diverse error function. Second, the network structure is optimized by eliminating unnecessary parameters with natural pruning. Through iterating these processes, candidate models are constructed and evaluated based on the Bayesian Information Criterion so that an optimal one is finally selected. Through computational experiments on benchmark problems, we confirm the weight parameter and structure optimization performance of the proposed method.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.14 no.6
• /
• pp.2618-2623
• /
• 2013

A hybrid model is proposed to easily predict the effective thermal conductivity of composites with aligned- and coated-short fibers, whose aspect ratio is not constant. The thermal conductivities of coated fillers are computed by using the generalized self-consistent model, resulting in that composites are simply simulated by the matrix with the equivalent short fibers. Finally, the thermal conductivity of the composites is predicted using the modified Eshelby model. The predicted results by the representative models and hybrid model are compared for the composite with aligned- and coated-short fibers of single aspect ratio. It is demonstrated that the hybrid model can be applied to the composite with aligned- and short-fibers of aspect ratios, 2 and 10, without any difficulties.

• Information and Communications Magazine
• /
• v.33 no.9
• /
• pp.3-7
• /
• 2016

딥러닝 기술은 수많은 입력 데이터에 내재하고 있는 특징을 추출 및 합성함으로써 복잡한 특징공간을 모델링할 수 있는 강점을 가지지만, 테스트 환경에서 나타날 수 있는 특정 데이터 분포에 대하여 일반화가 잘 되지 않을 경우에는 해당 데이터를 이용하여 주어진 환경에 모델을 적응시킬 수 있는 기술을 필요로 한다. 이 글에서는 DNN 모델의 adaptation 기술 연구가 가장 활발하게 진행되고 있는 음향모델링에서의 다양한 adaptation 기술을 통해 연구 동향을 알아본다.

• Proceedings of the KIEE Conference
• /
• 2009.04b
• /
• pp.76-78
• /
• 2009

Spoke type 영구자석 전동기는 자속이 집중되어 높은 공극자속밀도를 가지는 이점이 있으나 코깅토크가 비교적 크다는 단점이 있다. 소음 진동의 원인이 되는 코깅 토크를 저감하기 위해 모델1, 모델 2를 통해 회전자의 형상을 변화시키는 일반화 계수를 찾고 이를 모델3에 적용시켜 확인하고자 한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2024.05a
• /
• pp.846-849
• /
• 2024

다양한 분야에서 QR 코드가 급속도로 확산되면서, QR 코드를 악용하여 사용자를 악성 웹사이트로 리디렉션하는 '큐싱(Qshing)'이라는 새로운 형태의 사이버 범죄가 등장했다. 이에 본 연구에서는 일반화 성능을 향상시키기 위해 교차 검증(CV)을 활용하여 QR 코드 스캔과 관련된 악성 URL을 탐지하도록 설계된 스태킹 앙상블 모델을 제안한다. 이러한 통합은 실제 애플리케이션에서 높은 성능을 기대할 수 있도록 설계되었다. 본 연구는 이 모델이 기존의 연구보다 QR 코드 관련 사이버 위협에 대처하는 보다 효과적인 수단을 제공할 것으로 기대한다.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.18 no.2
• /
• pp.1-15
• /
• 2015

Urban flooding threatens human beings and facilities with chemical and physical hazards since the beginning of human civilization. Recent studies have emphasized the integration of data and models for effective urban flood inundation modeling. However, the model set-up process is tend to be time consuming and to require a high level of data processing skill. Furthermore, in spite of the use of high resolution grid data, inundation depth and velocity are varied with building treatment methods in 2-D inundation model, because undesirable grids are generated and resulted in the reliability decline of the simulation results. Thus, it requires building generalization process or enhancing building orthogonality to minimize the distortion of building before converting building footprint into grid data. This study aims to develop building generalization method for 2-dimensional inundation analysis to enhance the model reliability, and to investigate the effect of building generalization method on urban inundation in terms of geographical engineering and hydraulic engineering. As a result to improve the reliability of 2-dimensional inundation analysis, the building generalization method developed in this study should be adapted using Digital Building Model(DBM) before model implementation in urban area. The proposed building generalization sequence was aggregation-simplification, and the threshold of the each method should be determined by considering spatial characteristics, which should not exceed the summation of building gap average and standard deviation.

• KIPS Transactions on Computer and Communication Systems
• /
• v.6 no.11
• /
• pp.445-452
• /
• 2017

Electric pole is a supporting beam used for power transmission/distribution which accelerometer are used for measuring a external force. The meteorological condition has various effects on the external forces of electric pole. One of them is the elasticity change of the aerial wire. It is very important to perform modelling. The acceleration sensor is converted into a pitch and a roll angle. The meteorological condition has a high correlation between variables, and selecting significant explanatory variables for modeling may result in the problem of over-fitting. We constructed high deviance explained model considering multicollinearity using the Generalized Additive Model which is one of the machine learning methods. As a result of the Variation Inflation Factor Test, we selected and fitted the significant variable as temperature, precipitation, wind speed, wind direction, air pressure, dewpoint, hours of daylight and cloud cover. It was noted that the Hours of daylight, cloud cover and air pressure has high explained value in explonatory variable. The average coefficient of determination (R-Squared) of the Generalized Additive Model was 0.69. The constructed model can help to predict the influence on the external forces of electric pole, and contribute to the purpose of securing safety on utility pole.

• The Journal of the Acoustical Society of Korea
• /
• v.33 no.6
• /
• pp.383-391
• /
• 2014

In order to detect underwater acoustic signals, various SONAR array types have been developed, including towed line array SONAR system (TASS). However, the TASS suffers from performance degradation which is caused by aperture deformation during a turn, because the TASS have a long-aperture array. A parabolic array model for turning TASS have been developed to solve the degradation problem occurred during a turn. In this paper, adaptive beamforming system is developed using the parabolic TASS model to cancel interference signals. The developed beamforming system is based on generalized sidelobe canceller (GSC) structure and self-tuning adaptive algorithm.

• Proceedings of the Korean Information Science Society Conference
• /
• 2005.07a
• /
• pp.898-900
• /
• 2005

모듈러 멱승은 양수 x, E, N에 대하여 $x^Emod$ N로 정의된다. 모듈러 멱승 연산은 대부분의 공개키 암호화 알고리즘과 전자서명 프로토콜에서 핵심적인 연산으로 사용되고 있으므로, 그 효율성은 암호 프로토콜의 성능에 직접적인 영향을 미친다. 따라서 모듈러 멱승 연산에 필요한 곱셈 수를 감소시키기 위하여, 슬라이딩 윈도우를 적용한 CLNW 방법이나 VLNW 방법이 가장 널리 사용되고 있다. 본 논문에서는 조합론(combinatorics)에서 많이 응용되는 그래프 모델을 모듈러 멱승 연산에 적용할 수 있음을 보이고, 일반화된 그래프 모델을 통하여 VLNW 방법보다 더 적은 곱셈 수로 모듈러 멱승을 수행하는 방법을 설명한다. 본 논문이 제안하는 방법은 전체 곱셈 수를 감소시키는 새로운 블록들을 일반화된 그래프 모델의 초기 블록 테이블에 추가할 수 있는 초기 블록 테이블의 두 가지 확장 방법들로써, 접두사 블록의 확장과 덧셈 사슬 블록의 확장이다. 이 방법들은 새로운 블록을 초기 블록 테이블에 추가하기 위해 필요한 곱셈의 수와 추가한 뒤의 전체 곱셈 수를 비교하면서 초기 블록 테이블을 제한적으로 확장하므로, 지수 E에 non-zero bit가 많이 나타날수록 VLNW 방법에 비해 좋은 성능을 보이며 이는 실험을 통하여 검증하였다.

• Korean Journal of Environmental Biology
• /
• v.20 no.4
• /
• pp.378-385
• /
• 2002

Growth of Pacific oystey, Crassostrea gigas, in Kamak Bay, Korea was modeled using Von Bertalanffy growth function, seasonal Von Bertalanffy growth function and generalized growth equation of Schnute and Richards' growth model, based on shell length and wet weight frequency data of 9208 oysters. Carrying capacity in the oyster culture ground was also estimated using Schaefer's and Fox's surplus production model. The present results suggest that the generalized growth equation of Schnute and Richards' model is fitter to describe the length growth pattern of C. gigas than Von Bertalanffy growth functions. This results also suggest that the current number of culture facility per unit area in 2000 is similar to the number of facility that produces the maximum production of oyster per unit area.