• Proceeding of EDISON Challenge
• /
• 2014.03a
• /
• pp.211-224
• /
• 2014

금속을 포함한 분자에 대한 양자계산은 정확하고 일관된 결과를 얻기가 힘들 뿐만 아니라 상당한 컴퓨터 자원을 소비하며 많은 시간이 소요된다. 본 연구에서는 복잡한 양자계산의 근사를 위한 방법으로 본래 정성적인 구조 예측에 사용되는 닮은 궤도함수분석(Isolobal Analysis)을 정량적인 측면에서 접근해보고, 이를 통해 닮은 궤도(Isolobal) 구조를 가지고 있는 단위들(radical 등)에 대해서 계산을 근사할 수 있는 방법에 대해 논의한다. $CH_3$, $CH_2$와 닮은 궤도 구조를 가진 전형 원소를 중심으로 하는 분자들에 대해 가장 기초적인 근사계산인 Hartree-Fock 양자계산을 수행하였다. $(CUH_5){_2}^{2-}$를 표적으로 결합 구조를 예측하기 위한 경향성을 계산한 결합 성질로부터 파악한다. 분석 결과 동일한 주기에 대해서는 원자반지름(Atomic radii)에 대해 조화 형태의 결합에너지가 얻어졌으며, 동일한 족에 대해서는 좋은 근사가 되지 않았다. 파악된 경향성을 바탕으로 금속의 결합을 근사한 에너지에 대해서는 -1054.1875 kJ/mol로 비교적 큰 오차를 보였으나, 오차 항에 대한 분석이 가능해 추가적인 계들에 대한 계산으로 근사를 교정할 수 있을 것으로 보인다.

• Proceedings of the IEEK Conference
• /
• 1999.06a
• /
• pp.486-491
• /
• 1999

본 논문에서는 웨이브렛 신경회로망을 사용하여 알려지지 않은 비선형 시스템을 안정하게 적응 제어하는 문제를 다룬다. 비선형 시스템의 정확한 제어는 함수를 근사화하는 데 사용된 함수 근사화기의 정확성과 효율성에 의존한다. 이에 비선형 시스템 제어에 기준 함수의 선택이 자유롭고 함수 근사화 능력이 뛰어난 웨이브렛 신경회로망을 사용한다. 초기 웨이브렛 신경회로망 제어기 설정은 웨이브렛 신경회로망 변수인 신축과 이동 값을 제어기 입력의 시-주파수 특성을 분석해서 구하고, 연결강도는 Lyapunov 안정성 이론에 기초한 적응 법칙을 사용하여 조절한다. 이를 비선형 시스템인 역 진자 시스템에 적용한다.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.41 no.5
• /
• pp.381-389
• /
• 2017

The kernel density was determined based on sampling points obtained in a Markov chain simulation and was assumed to be an important sampling function. A Kriging metamodel was constructed in more detail in the vicinity of a limit state. The failure probability was calculated based on importance sampling, which was performed for the Kriging metamodel. A pre-existing method was modified to obtain more sampling points for a kernel density in the vicinity of a limit state. A stable numerical method was proposed to find a parameter of the kernel density. To assess the completeness of the Kriging metamodel, the possibility of changes in the calculated failure probability due to the uncertainty of the Kriging metamodel was calculated.

• The Korean Journal of Applied Statistics
• /
• v.18 no.1
• /
• pp.183-196
• /
• 2005

In this paper we studied the saddlepoint approximations to the distribution of non-homogeneous quadratic forms in normal variables. The results are the extension of Kuonen's which provide the same approximations to homogeneous quadratic forms. The CGF of interested statistics and related properties are derived for applications of saddlepoint techniques. Simulation results are also provided to show the accuracy of saddlepoint approximations.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.22 no.2
• /
• pp.191-198
• /
• 2011

In this paper, a method to approximate a multi-layer Green's function is proposed based on a FDTD scheme and a rational function approximation. For a given horizontal propagation wavenumber, time domain response is calculated and then Fourier transformed to the spectral domain Green's function. Using the rational function approximation, the pole and residue of the Green's function can be estimated, which are crucial for a calculation of a path loss. The proposed method can provide a wideband Green's function, while the conventional normal mode method can be applied to a single frequency problem. To validate the proposed method, We consider two problems, one of which has a analytical solution. The other is about multi-layer case, for which the proposed method is compared with the known normal mode solution, Kraken.

• 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.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.3
• /
• pp.747-753
• /
• 1997

Function approximation from a set of input-output parirs has numerous applications in scientiffc and engineer-ing areas.Multiayer feedforward neural networks have been proposed as a good approximator of noninear function.The back propagation (BP) algorithm allows muktiayer feedforward neural networks oro learn input-output mappongs from training samples.However, the mapping acquired through the BP algorithm nay be cor-rupt when errorneous trauning data are employed.In this paper we propose a robust BP learning algorithm that is resistant to the errormeous data and is capable of rejecting gross errors during the approximation process.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.13 no.3
• /
• pp.316-321
• /
• 2003

Recently, support vector learning attracts great interests in the areas of pattern classification, function approximation, and abnormality detection. It is well-known that among the various support vector learning methods, the so-called no-versions are particularly useful in cases that we need to control the total number of support vectors. In this paper, we consider the problem of function approximation utilizing both predetermined basis functions and a no-version support vector learning called $\nu-SVR$. After reviewing $\varepsilon-SVR$, $\nu-SVR$, and a semi-parametric approach, this paper presents an extension of the conventional $\nu-SVR$ method toward the direction that can utilize Predetermined basis functions. Moreover, the applicability of the presented method is illustrated via an example.

• Tunnel and Underground Space
• /
• v.28 no.5
• /
• pp.426-441
• /
• 2018

Recently, the generalized Hoek-Brown (GHB) failure criterion has been actively employed in various rock engineering calculations, but the analytical form of the corresponding Mohr failure envelope is not available, making it difficult to extend the application of the GHB criterion. In order to overcome this disadvantage, this study proposes a new method to express the tangential friction angle as an explicit function of normal stress by invoking the polynomial best-fitting to the relationship between normal stress and tangent friction angle implied in the GHB failure function. If this normal stress - tangential friction angle relationship is best-fitted with linear or quadratic polynomial function, it is possible to find the analytical root for tangential friction angle. Subsequently, incorporating the root into the relationship between shear stress and tangential friction angle accomplishes the derivation of the approximate Mohr envelope for the GHB criterion. It is demonstrated that the derived approximate Mohr failure envelopes are very accurate in the entire range of GSI value.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.16 no.2
• /
• pp.198-202
• /
• 2006

In this paper, we introduce Baldwin's approximate reasoning with fuzzy logic and some truth function mappings usually used in Baldwin's method. And we introduce some assessment criteria for approximate reasonings and we define some truth function mappings which satisfy more criteria than those which are already known.