• The Transactions of the Korean Institute of Electrical Engineers A
• /
• v.50 no.3
• /
• pp.105-111
• /
• 2001

The research presented in this paper focuses on a method for the detection of High Impedance Fault(HIF). The method will use the wavelet transform and neural network system. HIF on the multi-grounded three-phase four-wires primary distribution power system cannot be detected effectively by existing over current sensing devices. These paper describes the application of discrete wavelet transform to the various HIF data. These data were measured in actual 22-9kV distribution system. Wavelet transform analysis gives the frequency and time-scale information. The neural network system as a fault detector was trained to discriminate HIF from the normal status by a gradient descent method. The proposed method performed very well by proving the right state when it was applied staged fault data and normal load mimics HIF, such as arc-welder.

• Proceedings of the KIEE Conference
• /
• 2003.07d
• /
• pp.2187-2189
• /
• 2003

This paper presents a design method of the wavelet neural network(WNN) controller based on a direct adaptive control scheme for the intelligent control of chaotic systems. The conventional control methods such as optimal control, adaptive control and robust control may not be feasible when an explicit, faithful mathematical model cannot be constructed. Therefore, an intelligent control system that is an on-line trained WNN controller based on a direct adaptive control method is proposed to control chaotic systems whose mathematical models are not available. The gradient-descent method is used for training a wavelet neural network controller. Finally, the effectiveness and feasibility of the proposed control method is demonstrated with applications to the chaotic system.

• Proceedings of the Korean Information Science Society Conference
• /
• 2004.04b
• /
• pp.721-723
• /
• 2004

A trust-region method is a quite attractive optimization technique. It is, in general, faster than the steepest descent method and is free of a learning rate unlike the gradient-based methods. In addition to its convergence property (between linear and quadratic convergence), ifs stability is always guaranteed, in contrast to the Newton's method. In this paper, we present an efficient implementation of the maximum likelihood independent component analysis (ICA) using the trust-region method, which leads to trust-region-based ICA (TR-ICA) algorithms. The useful behavior of our TR-ICA algorithms is confimed through numerical experimental results.

• Korean Journal of Artificial Intelligence
• /
• v.5 no.1
• /
• pp.29-37
• /
• 2017

The range of problems that can be handled by the activation of big data and the development of hardware has been rapidly expanded and machine learning such as deep learning has become a very versatile technology. In this paper, mnist data set is used as experimental data, and the Cross Entropy function is used as a loss model for evaluating the efficiency of machine learning, and the value of the loss function in the steepest descent method is We applied the Gradient Descent Optimize algorithm to minimize and updated weight and bias via backpropagation. In this way we analyze optimal reliability value corresponding to the number of exercises and optimal reliability value without overfitting. And comparing the overfitting time according to the number of data changes based on the number of training times, when the training frequency was 1110 times, we obtained the result of 92%, which is the optimal reliability value without overfitting.

• International Journal of Advanced Culture Technology
• /
• v.5 no.2
• /
• pp.82-89
• /
• 2017

The range of problems that can be handled by the activation of big data and the development of hardware has been rapidly expanded and machine learning such as deep learning has become a very versatile technology. In this paper, mnist data set is used as experimental data, and the Cross Entropy function is used as a loss model for evaluating the efficiency of machine learning, and the value of the loss function in the steepest descent method is We applied the GradientDescentOptimize algorithm to minimize and updated weight and bias via backpropagation. In this way we analyze optimal reliability value corresponding to the number of exercises and optimal reliability value without overfitting. And comparing the overfitting time according to the number of data changes based on the number of training times, when the training frequency was 1110 times, we obtained the result of 92%, which is the optimal reliability value without overfitting.

• Smart Media Journal
• /
• v.3 no.2
• /
• pp.9-13
• /
• 2014

The Levenberg-Marquardt (LM) algorithm is the most widely used fitting algorithm. It outperforms simple gradient descent and other conjugate gradient methods in a wide variety of problems. Based on the work of paper[1], we propose a modified Levenberg-Marquardt algorithm for better performance of mobile system. The LM parameter at the $k_{th}$ iteration is chosen ${\mu}=A{\bullet}{\parallel}f(x){\parallel}{\bullet}I$ where f is the residual function, and J is the Jacobi of f. In this paper, we show this method is more efficient than traditional method under the situation that the system iteration is limited.

• Proceedings of the Korea Electromagnetic Engineering Society Conference
• /
• 2005.11a
• /
• pp.153-156
• /
• 2005

This paper concerns an efficient iterative approach for eliminating coherent interference signals in linearly constrained adaptive arrays. The Alternate Mainbeam Nulling Algorithm[1] is implemented iteratively to find an optimum weight vector. The convergence parameters in the unit gain and null constraints are calculated using steepest descent method with gradient estimation. The nulling performance of the proposed method is compared with that of conventional ones. It is shown that the proposed method performs better than conventional ones when the power of the coherent signals is large compared with a desired signal. Also, it performs consistently well for more number of interferences.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.53 no.4
• /
• pp.226-234
• /
• 2004

A rapid Prototyping system that laser-cuts and laminates thick layers can fabricate 3D objects promptly with a variety of materials. Building such a system must consider the surface distortions due to both vertical-cut layers and triangular surfaces. We developed a tangential layer-cutting algorithm by rearranging tangential lines such that they reconstruct 3D surfaces more closely and also constitute smoother laser trajectories. An energy function that reflects the surface-closeness with the tangential lines was formulated and then the energy was minimized by a gradient descent method. Since this simple method tends to cause many local minima for complex 3D objects, we tried to solve this problem by adding a simulated annealing process to the proposed method. To view and manipulate 3D objects, we also implemented a 3D visual environment. Under this environment, experiments on various 3D objects showed that our algorithm effectively approximates 3D surfaces and makes laser-trajectory feasibly smooth.

• Proceedings of the KIEE Conference
• /
• 2000.11d
• /
• pp.621-623
• /
• 2000

In this paper, the effective design method of the predictive controller using fuzzy neural networks(FNNs) is presented for the Intelligent control of chaotic nonlinear systems. In our design method of controller, predictor parameters are tuned by the error value between the actual output of a chaotic nonlinear system and that of a fuzzy neural network model. And the parameters of predictive controller using fuzzy neural network are tuned by the gradient descent method which uses control error value between the actual output of a chaotic nonlinear system and the reference signal. In order to evaluate the performance of our controller, it is applied to the Duffing system which are the representative continuous-time chaotic nonlinear systems and the Henon system which are representative discrete-time chaotic nonlinear systems.

• Proceedings of the IEEK Conference
• /
• 1998.06a
• /
• pp.933-935
• /
• 1998

The HMM-Net is an architecture for a neural network that implements a hidden markov model (HMM). The architecture is developed for the purpose of combining the discriminant power of neural networks with the time-domain modeling capability of HMMs. Criteria used for learning HMM-Net classifiers are maximum likelihood(ML) and minimization of mean squared error(MMSE). In this paper we propose an efficient learning method of HMM_Net classifiers using a ML-MMSE hybrid criterion and report the results of an experimental study comparing the performance of HMM_Net classifiers trained by the gradient descent algorithm with the above criteria. Experimental results for the isolated numeric digits from /0/ to /9/ show that the performance of the proposed method is better than the others in the repects of learning and recognition rates.