• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.33 no.5
• /
• pp.761-773
• /
• 2023

This paper proposes a method to increase the power-analysis resistance of the neural network model's feedforward process by replacing the exponential-based activation function, used in the deep-learning field, with an approximated function especially at the multi-layer perceptron model. Due to its nature, the feedforward process of neural networks calculates secret weight and bias, which already trained, so it has risk of exposure of internal information by side-channel attacks. However, various functions are used as the activation function in neural network, so it's difficult to apply conventional side-channel countermeasure techniques, such as masking, to activation function(especially, to exponential-based activation functions). Therefore, this paper shows that even if an exponential-based activation function is replaced with approximated function of simple form, there is no fatal performance degradation of the model, and than suggests a power-analysis resistant feedforward neural network with exponential-based activation function, by masking approximated function and whole network.

• Proceedings of the Korea Society of Information Technology Applications Conference
• /
• 2007.05a
• /
• pp.103-108
• /
• 2007

다항지수 신뢰도 함수(multinomial-exponential reliability function ; MERF) 는 소프트웨어의 고장/수정 공정을 세밀하게 수행하는 중에 개발되는 관계에 있다. 후에 MERF는 좀더 매우 단순화한 지수 신뢰도 함수(exponential reliability function ; EARF)로 근사화되는 공정을 거치게 된다. 이는 MERF의 특성을 대부분 가지고 있어서 두 개의 함수가 하나의 신뢰도 함수로 단일화되도록 한다. 신뢰도 모델 MERF/EARF는 소프트웨어 고장 공정을 NHPP로, 수정공정을 다항분포로 고려한다. 이 모텔은 두 개의 공정 모두가 통계적 독립인 것으로 간주한다. 본 논문에서는 모델의 이론적인 기준, 수학적 특성, 소프트웨어 신뢰도에의 응용을 검토한다. 이는 물리적 인 시스템을 검사하고 유지보수하는 선도적인 모델응용이다. 본 논문에는 소프트웨어 신뢰도 분석에 응용하는 하나의 수치 예를 포함한다.

• Journal of Korean Society of Forest Science
• /
• v.102 no.4
• /
• pp.467-476
• /
• 2013

Site index information given from forest site map only exist in the sampled locations. In this study, site index for unsampled locations were estimated using kriging interpolation method which can interpolate values between point samples to generate a continuous surface. Site index of Pinus densiplora in Danyang area were calculated using Chapman-Richards model by plot unit. Then site index for unsampled locations were interpolated by theoretical variogram models and ordinary kriging. Also in order to assess parameter selection, cross-validation was performed by calculating mean error (ME), average standard error (ASE) and root mean square error (RMSE). In result, gaussian model was excluded because of the biggest relative nugget (37.40%). Then spherical model (16.80%) and exponential model (8.77%) were selected. Site index estimates of Pinus densiplora throughout the entire area in Danyang showed 4.39~19.53 based on exponential model, and 4.54~19.23 based on spherical model. By cross-validation, RMSE had almost no difference. But ME and ASE from spherical model were slightly lower than exponential model. Therefore site index prediction map from spherical model were finally selected. Average site index from site prediction map was 10.78. It can be expected that regional variance can be considered by site index prediction map in order to estimate forest biomass which has big spatial variance and eventually it is helpful to improve an accuracy of forest carbon estimation.

• Journal for History of Mathematics
• /
• v.24 no.2
• /
• pp.47-59
• /
• 2011

Various mathematical models have been widely studied recently in both fields of mathematics and ecology since they help us understand the dynamical process of population changes in biological species living in a certain habitat and give useful predictions. The world population model proposed by Malthus, a British economist, in his work 'An Essay on the Principle of Population' published in the period of 1789~1826 is one of the early mathematical models on population changes. Malthus' models and the carrying capacity models of Verhulst in 1845 were based on exponential type functions. The independent research field of mathematical ecology has been started from Lotka's works in 1920's. Since then various different mathematical models has been proposed and examined. This article mainly deals with single species population change models expressed in terms of ordinary differential equations.

• Journal of Broadcast Engineering
• /
• v.26 no.2
• /
• pp.184-196
• /
• 2021

This paper proposes statistics adaptive linear regression-based object size prediction method for object detection. YOLOv2 and YOLOv3, which are typical deep learning-based object detection algorithms, designed the last layer of a network using statistics adaptive exponential regression model to predict the size of objects. However, an exponential regression model can propagate a high derivative of a loss function into all parameters in a network because of the property of an exponential function. We propose statistics adaptive linear regression layer to ease the gradient exploding problem of the exponential regression model. The proposed statistics adaptive linear regression model is used in the last layer of the network to predict the size of objects with statistics estimated from training dataset. We newly designed the network based on the YOLOv3tiny and it shows the higher performance compared to YOLOv3 tiny on the UFPR-ALPR dataset.

• The KIPS Transactions:PartD
• /
• v.11D no.1
• /
• pp.163-172
• /
• 2004

In the software development project applying object-oriented development methodology, the research on the UCP(Use Case Point) as a method to estimate development effort is being carried on. The existing research proposes the linear model calculating the development effort that multiplies an invariant on AUCP(Adjusted Use Case Point) which applied technical and environmental factors. However, the statistical model that estimates the development effort using AUCP and UUCP(Unadjusted Use Case Point) is not being studied. The irrelevant relationship of the linear regression model, whose development period is increasing tremendously as the software size increases, is confirmed. Moreover, during the UCP calculating process, there can be errors in FP by applying the TCF(Technical Complexity Factor) and EF(Environmental Factor). This paper presents a non-linear regression model, that does not consider the TCF and EF, and that estimate the development effort from UUCP directly by utilizing the exponential function. An exponential function is selected among the linear, logarithm, polynomial, power, and exponential model via statistical evaluations of the models mentioned above.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.9 no.2
• /
• pp.133-140
• /
• 2016

Software reliability in the software development process is an important issue. Software process improvement helps in finishing with reliable software product. Infinite failure NHPP software reliability models presented in the literature exhibit either constant, monotonic increasing or monotonic decreasing failure occurrence rates per fault. In this paper, we were proposed the reliability model with the exponential and inverse exponential distribution, which made out efficiency application for software reliability. Algorithm to estimate the parameters used to maximum likelihood estimator and bisection method, model selection based on mean square error (MSE) and coefficient of determination($R^2$), for the sake of efficient model, were employed. Analysis of failure, using real data set for the sake of proposing the exponential and inverse exponential distribution, was employed. This analysis of failure data compared with the exponential and inverse exponential distribution property. In order to insurance for the reliability of data, Laplace trend test was employed. In this study, the inverse exponential distribution model is also efficient in terms of reliability because it (the coefficient of determination is 80% or more) in the field of the conventional model can be used as an alternative could be confirmed. From this paper, the software developers have to consider life distribution by prior knowledge of the software to identify failure modes which can be able to help.

• The Proceedings of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.6 no.3
• /
• pp.43-49
• /
• 1992

Forecasting of electric power demand has been a basic element for electric power system operation and system development, and it's accuracy has very strong influence on reliability and economical efficience of power supply. So, in this paper, long―term maximum electric power demand has been forecasted by using the triple exponential smoothing method initiated R.G.Brown. It has been regarded this method as high accuracy and operational convenience. The smoothing function is a liner combination of all past observations and the weight given to previous observations decreases geometrically with age.

• Journal of Internet Computing and Services
• /
• v.8 no.2
• /
• pp.105-115
• /
• 2007

The up-to-moment dataset is built by combining the past dataset and the recent dataset. The proposal is to compute association rules in real time. This study proposed the model, $EM_{past'}$ and algorithm that is sensitive to time. It can be utilized in real time by applying partitioned combination law after dividing the past dataset into(k-1). Also, we suggested $EM^{ES}_{past}$ applying the exponential smoothing method to $EM^p_{past'}$ When the association rules of $EM_{past'}\;EM^w_{past'\;and\;EM^{ES}_{past}$ were compared, The simulation results showed that $EM^{ES}_{past}$ is most accurate for testing dataset than $EM_{past}$ and $EM^w_{past}$ in huge dataset.

• Proceedings of the KIEE Conference
• /
• 1990.07a
• /
• pp.51-54
• /
• 1990

A parametrization for a linear system is presented to design a direct model reference adaptive pole placement controler. This parametrized model is one of the structured nonminimal models. The exponentially weighted least-squres algorithm is employed to estimate the control parameters. The direct adaptive controller has the exponential weighting properties by the proposed method of selecting the characteristic polynomials of the sensitivity function filters in connection with the reference models.