• Journal of the Korean Institute of Intelligent Systems
• /
• v.14 no.7
• /
• pp.847-851
• /
• 2004

In the original mixture density network(MDN), which was introduced by Bishop and Nabney, the parameters of the conditional probability density function are represented by the output vector of a single multi-layer perceptron. Among the recent modification of the MDNs, there is the so-called modified mixture density network, in which each of the priors, conditional means, and covariances is represented via an independent multi-layer perceptron. In this paper, we consider a further simplification of the modified MDN, in which the conditional means are linear with respect to the input variable together with the development of the MATLAB program for the simplification. In this paper, we first briefly review the original mixture density network, then we also review the modified mixture density network in which independent multi-layer perceptrons play an important role in the learning for the parameters of the conditional probability, and finally present a further modification so that the conditional means are linear in the input. The applicability of the presented method is shown via an illustrative simulation example.

• Proceedings of the Optical Society of Korea Conference
• /
• 1991.06a
• /
• pp.23-28
• /
• 1991

A single-layer neural network with 4$\times$4 input neurons and 4 output neurons is optically implemented. Holographic lenslet arrays are used for the e optical interconnection topology, a liquid crystal light valve(LCLV) is used for controlling optical interconection weights. Using a Perceptron learning rule, it classifics input patterns into 4 different categories. It is shown that the performance of the adaptive neural network depends on the learning rate, the correlation of input patterns, and the nonlinear characteristic properties of the liquid crystal light valve.

• Journal of Electrical Engineering and Technology
• /
• v.4 no.1
• /
• pp.28-34
• /
• 2009

In a Gas Insulated Substation (GIS), Very Fast Transients (VFTs) are generated mainly due to switching operations. These transients may cause internal faults, i.e., layer-to-layer faults in a capacitively graded bushing as it is one of the most important terminal equipment for GIS. The healthiness of the bushing is generally verified by measuring its leakage current. However, the change in current magnitude/pattern is only marginal for different types of fault conditions. Leakage current monitoring (LCM) systems generate large amounts of data and computer aided interpretation of defects may be of great assistance when analyzing this data. In view of the above, ANN techniques have been used in this study for identification of these minor faults. A single layer perceptron network, a two layer feed-forward back propagation network and cascade correlation (CC) network models are used to identify interlayer faults in the bushing. The effectiveness of the CC network over perceptron and back propagation networks in identification of a fault has been analysed as part of the paper.

• International Journal of Contents
• /
• v.9 no.4
• /
• pp.11-15
• /
• 2013

"Protein Folding Problem" is considered to be one of the "Great Challenges of Computer Science" and prediction of disordered protein is an important part of the protein folding problem. Machine learning models can predict the disordered structure of protein based on its characteristic of "learning from examples". Among many machine learning models, we investigate the possibility of multilayer perceptron (MLP) as the predictor of protein disorder. The investigation includes a single hidden layer MLP, multi hidden layer MLP and the hierarchical structure of MLP. Also, the target node cost function which deals with imbalanced data is used as training criteria of MLPs. Based on the investigation results, we insist that MLP should have deep architectures for performance improvement of protein disorder prediction.

• 제어로봇시스템학회:학술대회논문집
• /
• 1991.10a
• /
• pp.280-285
• /
• 1991

Neural network control has many innovative potentials for fast, accurate and intelligent adaptive control. In this paper, a learning control architecture for the dynamic control of a robot manipulator is developed using inverse dynamic neurocontroller and linear neurocontroher. The inverse dynamic neurocontrouer consists of a MLP (multi-layer perceptron) and the linear neurocontroller consists of SLPs (single layer perceptron). Compared with the previous type of neurocontroller which is using an inverse dynamic neurocontroller and a fixed PD gain controller, proposed architecture shows the superior performance over the previous type of neurocontroller because linear neurocontroller can adapt its gain according to the applied task. This superior performance is tested and verified through the control of PUMA 560. Without any knowledge on the dynamic model, its parameters of a robot , (The robot is treated as a complete black box), the neurocontroller, through practice, gradually and implicitly learns the robot's dynamic properties which is essential for fast and accurate control.

• The Journal of the Acoustical Society of Korea
• /
• v.28 no.1
• /
• pp.32-43
• /
• 2009

In this paper, we improve the performance of cardiac disorder classification by continuous heart sound signals using automatic segmentation and extreme learning machine (ELM). The accuracy of the conventional cardiac disorder classification systems degrades because murmurs and click sounds contained in the abnormal heart sound signals cause incorrect or missing starting points of the first (S1) and the second heart pulses (S2) in the automatic segmentation stage, In order to reduce the performance degradation due to segmentation errors, we find the positions of the S1 and S2 pulses, modify them using the time difference of S1 or S2, and extract a single period of heart sound signals. We then obtain a feature vector consisting of the mel-scaled filter bank energy coefficients and the envelope of uniform-sized sub-segments from the single-period heart sound signals. To classify the heart disorders, we use ELM with a single hidden layer. In cardiac disorder classification experiments with 9 cardiac disorder categories, the proposed method shows the classification accuracy of 81.6% and achieves the highest classification accuracy among ELM, multi-layer perceptron (MLP), support vector machine (SVM), and hidden Markov model (HMM).

• Laser Solutions
• /
• v.2 no.3
• /
• pp.19-31
• /
• 1999

This paper introduces the monitoring scheme of laser welding quality using neural network. The developed monitoring scheme detects light signal emitting from plasma formed above the weld pool with optic sensor and DSP-based signal processor, and analyzes to give a guidance about the weld quality. It can automatically detect defects of laser weld and further give an information about what kind of defects it is, specially partial penetration and porosity among the interior defects. Those could be detected only by naked eyes or X-ray after welding, which needs more processes and costs in mass production. The monitoring scheme extracts four feature vectors from signal processing results of optical measuring data. In order to classify pattern for extracted feature vectors and to decide defects, it uses single-layer neural network with perceptron learning. The monitoring result using only the first feature vector shows confidence rate in recognition of 90%($\pm$5) and decides whether normal status or defects status in real time.

• Journal of information and communication convergence engineering
• /
• v.2 no.1
• /
• pp.36-39
• /
• 2004

In this paper, a method of improving the learning speed and convergence rate is proposed to exploit the advantages of artificial neural networks and neuro-fuzzy systems. This method is applied to the XOR problem, n bit parity problem, which is used as the benchmark in the field of pattern recognition. The method is also applied to the recognition of digital image for practical image application. As a result of experiment, it does not always guarantee convergence. However, the network showed considerable improvement in learning time and has a high convergence rate. The proposed network can be extended to any number of layers. When we consider only the case of the single layer, the networks had the capability of high speed during the learning process and rapid processing on huge images.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.29B no.8
• /
• pp.35-43
• /
• 1992

We have presented the tracing algorithm for center pixel of character image. Character image was read by scanner device. Performing the tracing process, it can be possible to detect feature points, such as branch point, stroke of 4 directions. So, the tracing process covers the thinning and feature point detection process for improving the processing time. Usage of suggested tracing algorithm instead of thinning that is the preprocessing of character recognition increases speed up to 5 times. The preprocessing chip has been designed by using single layer perceptron algorithm.

• Proceedings of the Optical Society of Korea Conference
• /
• 1991.06a
• /
• pp.29-34
• /
• 1991

An optoelectronic implementation of analog and non-volatile synaptic weights of neural networks is proposed by using the doping modulated amophous silicon multilayer. The persistent photoconductivity(PPC) of the multilayer induced by a short illumination is characterized in experiment and implemented to the non-volatile synaptic weights. An optoelectronic processor with the single layer perceptron algorithm is also proposed. Some learning equations of the processor and the results of simulation are presented.