• 한국정보과학회:학술대회논문집
• /
• 한국정보과학회 1998년도 가을 학술발표논문집 Vol.25 No.2 (2)
• /
• pp.327-329
• /
• 1998

Up-Propagation is an algorithm for inverting and learning neural network generative models. Sensory input is processed by inverting a model that generates patterns from hidden variables using top-down connections. The inversion process is iterative, utilizing a negative feedback loop that depends on an error signal propagated by bottom-up connections. The error signal is also used to learn the generative model from examples. the algorithm is benchmarked against principal component analysis in experiments on images of handwritten digits.

• 한국농공학회지
• /
• 제18권1호
• /
• pp.4021-4037
• /
• 1976

This paper describes the actual experiments of the error propagation and studies of analytical photogrammetry using the simulation method in which we can find the causes of the errors. These studies and the results give the valuable data which are very effective for systematically controlling the errors in aerial triangulation. The main contents in my paper are as follows: 1. Dose the scale errors in the successive models in the form of normal distribution appear when the observation errors propagate in the form of normal distribution\ulcorner 2. In what form does this scale error propagation in the actual model appear\ulcorner 3. When the causes of the scale error propagation are found, is the evaluation standard determined normally\ulcorner 4. What degree of influence is there form the constant errors\ulcorner I have done several experiments using the simulation method technique to solve the complicated error propgation of aerial triangulation which is the effective means to research the relations between cause and effect. In this paper, the studies have concentrated on the following points of simulation experiments. (1) The first part descries how we can produce the soft program of the simulation experiment. (2) The second part is the method propagating the errors in the simulation models and the kinds of errors. (3) The final part is the most important; that is the analyzing and evaluation of control during actual work. From the above-mentioned points, it is said that these studies are a very important development in the field of controlling and managing aerial photogrammetry and especially in the case of error propagation, we can clearly find the causes of the errors and steps and parts of errors generated when we use these techniques.

• 한국군사과학기술학회지
• /
• 제18권2호
• /
• pp.167-172
• /
• 2015

In this paper, we have compared some radio propagation models in order to verify the performance of W.C.Y LEE propagation model in mountain area. The four propagation models, which are Okumura-Hata, ITU-R P.525, Egli and W.C.Y. LEE, are analyzed by comparing the differences between measured values and propagation loss estimation values. And a correction method for W.C.Y LEE model is suggested to improve the performance of W.C.Y. LEE model with measured data in mountain area. Simulation results show that the estimation error using W.C.Y LEE model is the lowest among four propagation models. Also, the results show that the corrected W.C.Y LEE model with suggested method improves the performance of propagation loss estimation.

• 대한토목학회논문집
• /
• 제12권3호
• /
• pp.231-238
• /
• 1992

계산시간(計算時間)의 단축(短縮)을 위하여 EVP(Error Vector Propagation) 방법(方法)을 사용하여 타원형(楕圓形) 완경사방정식(緩傾斜方程式)을 해석(解析)하였다. 수치실험(數値實驗)은 수중(水中)에 타원형(楕圓形) 여울이 존재하는 완경사(緩傾斜) 해역(海域)에서 수행하였으며, 포물선형(抛物線形) 모형(模型) 및 쌍곡선형(雙曲線形) 모형(模型)을 같이 계산하여 각각의 결과(結果)를 수리실험(水理實驗) 결과(結果)와 비교(比較)하였다. 또한 이안제(離岸堤)가 설치된 파랑장(波浪場)의 경우에도 쌍곡선형(雙曲線形) 모형(模型)의 결과(結果) 및 수리실험(水理實驗) 결과(結果)와 비교(比較)하였다. 적용결과(適用結果) 계산시간(計算時間) 면에서는 다른 모형(模型)에 비하여 만족스럽게 단축(短縮)할 수 있었으며, 해(解)의 정확성(正確性)에서는 약간의 진동현상(振動現象)이 나타나지만 그 경향(傾向)은 잘 일치하였다.

• 한국정보과학회논문지:정보통신
• /
• 제31권4호
• /
• pp.375-383
• /
• 2004

최근에 사용 편이성으로 인해 다양한 무선 이동 네트워크들이 널리 보급되면서, 무선 네트워크성능을 향상시키기 위한 연구가 활발히 진행되고 있다. 무선 네트워크에서의 패킷 손실은 유선 네트워크의 혼잡이 아닌, 전파 오류로 인해 빈번히 발생되기 때문에, 시뮬레이션에서 무선 네트워크의 성능을 정확히 평가하기 위해서는 알맞은 무선 채널 모델을 채택해야 한다. 적합한 채널 모델은 사용 주파수 영역, 신호출력, 방해물 존재 유무, 평가하는 프로토콜의 비트 오류에 대한 민감성 둥 여러 가지 변수를 고려하여 선택해야 한다. 본 논문에서는 센서(Sensor) 채널의 고 전파 오류 특성을 분석하고, 센서 채널에 알맞은 채널 모델을 결정한다. 또한 센서 네트워크에서 수집한 비트 오류 데이타와 다양한 이론적 무선 채널 모델링 방식을 이용하여 링크계층 FEC(Forward Error Correction) 알고리즘과 TCP 성능 변화를 평가한다. 10일간의 센서 채널 트레이스와의 비교 분석에 의하면, CM(Chaotic Map) 모델은 센서 채널의 BER 편차와 PER(Packet Error Rate) 같을 각각 3배와 10배 이내의 오차 범위에서, 다른 모델은 수십 배 이상 오차범위에서 예측한다. FEC 알고리즘과 세가지 TCP (Tahoe, Reno, 그리고 Vegas) 시뮬레이션 실험에서도 CM 모델은 트레이스와 유사한 성능 변화를, 다른 모델은 최대 10배 이상의 오차를 보인다.

• 생물정신의학
• /
• 제4권2호
• /
• pp.194-197
• /
• 1997

Neural network models, also known as connectionist models or PDP models, simulate some functions of the brain and may promise to give insight in understanding the cognitive brain functions. The models composed of neuron-like elements that are linked into circuits can learn and adapt to its environment in a trial and error fashion. In this article, the history and principles of the neural network modeling are briefly reviewed, and its applications to psychiatry are discussed.

• 한국농공학회지
• /
• 제45권2호
• /
• pp.45-57
• /
• 2003

An artificial neural network model was developed to analyze and forecast Short-term river runoff from the Naju watershed, in Korea. Error back propagation neural networks (EBPN) of hourly rainfall and runoff data were found to have a high performance In forecasting runoff. The number of hidden nodes were optimized using total error and Bayesian information criterion. Model forecasts are very accurate (i.e., relative error is less than 3% and $R^2$is greater than 0.99) for calibration and verification data sets. Increasing the time horizon for application data sets, thus mating the model suitable for flood forecasting. decreases the accuracy of the model. The resulting optimal EBPN models for forecasting hourly runoff consists of ten rainfall and four runoff data(ANN0410 model) and ten rainfall and ten runoff data(ANN1010 model). Performances of the ANN0410 and ANN1010 models remain satisfactory up to 6 hours (i.e., $R^2$is greater than 0.92).

• 한국지능시스템학회논문지
• /
• 제14권1호
• /
• pp.118-124
• /
• 2004

In this paper, we present a indirect adaptive control method using a wavelet neural network (WNN) for the control of chaotic nonlinear systems without precise mathematical models. The proposed indirect adaptive control method includes the off-line identification and on-line control procedure for chaotic nonlinear systems. In the off-line identification procedure, the WNN based identification model identifies the chaotic nonlinear system by using the serial-parallel identification structure and is trained by the gradient-descent method. And, in the on-line control procedure, a WNN controller is designed by using the off-line identification model and is trained by the error back-propagation algorithm. Finally, the effectiveness and feasibility of the proposed control method is demonstrated with applications to the chaotic nonlinear systems.

• Structural Engineering and Mechanics
• /
• 제70권6호
• /
• pp.671-681
• /
• 2019

This paper examines the applicability of artificial neural network (ANN) and multivariate adaptive regression splines (MARS) to predict the compressive strength of bacteria incorporated geopolymer concrete (GPC). The mix is composed of new bacterial strain, manufactured sand, ground granulated blast furnace slag, silica fume, metakaolin and fly ash. The concentration of sodium hydroxide (NaOH) is maintained at 8 Molar, sodium silicate ($Na_2SiO_3$) to NaOH weight ratio is 2.33 and the alkaline liquid to binder ratio of 0.35 and ambient curing temperature ($28^{\circ}C$) is maintained for all the mixtures. In ANN, back-propagation training technique was employed for updating the weights of each layer based on the error in the network output. Levenberg-Marquardt algorithm was used for feed-forward back-propagation. MARS model was developed by establishing a relationship between a set of predictors and dependent variables. MARS is based on a divide and conquers strategy partitioning the training data sets into separate regions; each gets its own regression line. Six models based on ANN and MARS were developed to predict the compressive strength of bacteria incorporated GPC for 1, 3, 7, 28, 56 and 90 days. About 70% of the total 84 data sets obtained from experiments were used for development of the models and remaining 30% data was utilized for testing. From the study, it is observed that the predicted values from the models are found to be in good agreement with the corresponding experimental values and the developed models are robust and reliable.

• 천문학회보
• /
• 제37권2호
• /
• pp.124.1-124.1
• /
• 2012

We have made a comparison of the WSA-ENLIL CME propagation model with three cone types and an empirical model using 29 halo CMEs from 2001 to 2002. These halo CMEs have cone model parameters from Michalek et al. (2007) as well as their associated interplanetary (IP) shocks. For this study we consider three different cone models (an asymmetric cone model, an ice-cream cone model and an elliptical cone model) to determine CME cone parameters (radial velocity, angular width and source location), which are used for input parameters of the WSA-ENLIL CME propagation model. The mean absolute error (MAE) of the arrival times at the Earth for the elliptical cone model is 10 hours, which is about 2 hours smaller than those of the other models. However, this value is still larger than that (8.7 hours) of an empirical model by Kim et al. (2007). We are investigating several possibilities on relatively large errors of the WSA-ENLIL cone model, which may be caused by CME-CME interaction, background solar wind speed, and/or CME density enhancement.