• Proceedings of the KIEE Conference
• /
• 1999.07g
• /
• pp.2932-2934
• /
• 1999

In this paper, a scheme for recognition of handwritten digits using a multilayer neural network trained with the back-propagation algorithm using generalized delta rule is proposed. The neural network is trained with hand written digit data of different writers and different styles. One of the purpose of the work with neural networks is the minimization of the mean square error(MSE) between actual output and desired one. The back-propagation algorithm is an efficient and very classical method. The back-propagation algorithm for training the weights in a multilayer net uses the steepest descent minimization procedure and the sigmoid threshold function. As an error rate is reduced, recognition rate is improved. Therefore we propose a method that is reduced an error rate.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.40 no.2
• /
• pp.224-233
• /
• 1991

The multilayer expanson of single layer NN (Neural Network) was needed to solve the linear seperability problem as shown by the classic example using the XOR function. The EBP (Error Back Propagation ) learning rule is often used in multilayer Neural Networks, but it is not without its faults: 1)D.Rimmelhart expanded the Delta Rule but there is a problem in obtaining Ca from the linear combination of the Weight matrix N between the hidden layer and the output layer and H, wich is the result of another linear combination between the input pattern and the Weight matrix M between the input layer and the hidden layer. 2) Even if using the difference between Ca and Da to adjust the values of the Weight matrix N between the hidden layer and the output layer may be valid is correct, but using the same value to adjust the Weight matrixd M between the input layer and the hidden layer is wrong. Recognitron III was proposed to solve these faults. According to simulation results, since Recognitron III does not learn the three layer NN itself, but divides it into several single layer NNs and learns these with learning patterns, the learning time is 32.5 to 72.2 time faster than EBP NN one. The number of patterns learned in a EBP NN with n input and output cells and n+1 hidden cells are 2**n, but n in Recognitron III of the same size. [5] In the case of pattern generalization, however, EBP NN is less than Recognitron III.

• 제어로봇시스템학회:학술대회논문집
• /
• 1997.10a
• /
• pp.1561-1564
• /
• 1997

In the study, simulation result was studied by connecting PID controller in series to the established Neural Networks Controller. Neural Network model is composed of two layers to evaluate tracking performance improvement. The reqular dynamic characteristics was also studied for the expected error to be minimized by using Widrow-Hoff delta rule. As a result of the study, We identified that tracking performance inprovement was developed more in case of connecting PID than Neural Network Contoller and that tracking plant parameter in 251 sample was approached rapidly case of time variable.

• Journal of Ocean Engineering and Technology
• /
• v.7 no.2
• /
• pp.131-140
• /
• 1993

To study the corrosion fatigue begavior of T.M.C.P. steel, the rotary bending fatigue test with the change of concentration of NaCl solution was carried out. Fatigue life in the corrosion environment is decreased markedly in comparision with that in the air. Fatigue limit in the air was about 225 MPa. In case of 3.5% NaCl solution fatigue life could be expressed as .sigma./sub f/=10,392 * (N/sub f/)/sup -o.2923 . According to the paris's rule, crack growth rates could be expressed as da/dN=2.62.*10/sup -7/ .DELTA. K/sup 1.09/(3.5% NaCl solution), da/dN=1.95 *10/sup -7 .DELTA. K/sup 1.05/(1% NaCl solution), da/dN=2.62 * 10/sup -7/.DELTA./sup 0.72/(0.01% NaCl solution) with da/dN expressed in mm/cycle and .DELTA.K in MPa.GAMMA.m. The crack growth rate in the corrosion environment was highest under 3.5% NaCl solution.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2005.07a
• /
• pp.589-590
• /
• 2005

BSCCO:2212-2223 thin films were fabricated by using the ion beam sputter with a evaporation method at various substrate temperatures, $T_{sub}$, and ozone gas pressures, $pO_3$. The correlation diagrams of the BSCCO phases with Tsub and $pO_3$ are established in the 2212 and 2223 compositional films. In spite of 2212 compositional sputtering, Bi2201 and Bi2223 as well as Bi2212 phases come out as stable phases depending on Tsub and $pO_3$. From these results, the thermodynamic evaluation of ${\Delta}H$ and ${\Delta}S$, which are related with Gibbs' free energy change for single Bi2212 or Bi2223 phase, was performed.

• Korean Journal of Materials Research
• /
• v.24 no.4
• /
• pp.201-206
• /
• 2014

TiC-21 mol% Mo solid solution (${\delta}$-phase) and TiC-99 mol% Mo solid solution (${\beta}$-phase), and TiC-(80~90) mol% Mo hypo-eutectic composite were deformed by compression in a temperature range from room to 2300 K and in a strain rate range from $4.9{\times}10^{-5}$ to $6.9{\times}10^{-3}/s$. The deformation behaviors of the composites were analyzed from the strengths of the ${\delta}$- and ${\beta}$-phases. It was found that the high strength of the eutectic composite is due primarily to solution hardening of TiC by Mo, and that the ${\delta}$-phase undergoes an appreciable plastic deformation at and above 1420 K even at 0.2% plastic strain of the composite. The yield strength of the three kinds of phase up to 1420 K is quantitatively explained by the rule of mixture, where internal stresses introduced by plastic deformation are taken into account. Above 1420 K, however, the calculated yield strength was considerably larger than the measured strength. The yield stress of ${\beta}$-phase was much larger than that of pure TiC. A good linear relationship was held between the yield stress and the plastic strain rate in a double-logarithmic plot. The deformation behavior in ${\delta}$-phase was different among the three temperature ranges tested, i.e., low, intermediate and high. At an intermediate temperature, no yield drop occurred, and from the beginning the work hardening level was high. At the tested temperature, a good linear relationship was held in the double logarithmic plot of the yield stress against the plastic strain rate. The strain rate dependence of the yield stress was very weak up to 1273 K in the hypo-eutectic composite, but it became stronger as the temperature rose.

• The Korean Journal of Ceramics
• /
• v.6 no.1
• /
• pp.9-14
• /
• 2000

Glass transition temperature (T/sub g/) is proportional to the quadrupole splitting(Δ) of Fe(III) obtained from the /sup 57/Fe Mossbauer spectra (T/sub g/-Δ rule (1990)). The values of Δ reflect the distortion of Fe(III) atoms, which occupy the sites of network-forming atoms. Heat treatment of potassium vanadate and calcium gallate glasses at around the individual T/sub g/ causes a structural relaxation, accompanying a linear decrease of T/sub g/ and Δ values. These experimental results prove that T/sub g/ decreases with a decrease in the distortion of VO₄, GaO₄, and FeO₄tetrahedra, as the T/sub g/-Δ rule predicted.

• Journal of the Korean Data and Information Science Society
• /
• v.16 no.4
• /
• pp.815-821
• /
• 2005

Among the various artificial neural networks the backpropagation network (BPN) has become a standard one. One of the components in a neural network is an activating function or a transfer function of which a representative function is a sigmoid. We have discovered that by updating the slope parameter of a sigmoid function simultaneous with the weights could improve performance of a BPN.

• Journal of the Korean Society of Fisheries and Ocean Technology
• /
• v.28 no.2
• /
• pp.216-227
• /
• 1992

The corrosion rate, behavior of corrosion fatigue and characteristic of cathodic protection for SB41 were investigated by corrosion and corrosion control tests in seawater at laboratory and coast. The main result obtained are as the following; 1) The corrosion rate of base metal (BM) is about 28-37 mg/dm super(2) day in seawater of coast. 2) The correlation between the stress intensity factor range $\Delta$K and crack propagation rate da/dN for weldment follows paris' rule in seawater : da/dN=C($\Delta$K) super(m) where m is the slope of the correlation, and is 2.02 for BM and 1.75 for heat affected zone (HAZ) respectively. 3) The corrosion sensitivity of HAZ is more sensitive than that of BM under the low region of $\Delta$K. 4) With increase of bared surace area of cathode, cathodic protection potential is increased sharply.

• Journal of the Korea Society of Computer and Information
• /
• v.10 no.5 s.37
• /
• pp.11-16
• /
• 2005

Recently, there are a lot of endeavor to implement a fuzzy theory to artificial neural network. Goh proposed the fuzzy single layer perceptron algorithm and advanced fuzzy perceptron based on the generalized delta rule to solve the XOR Problem and the classical Problem. However, it causes an increased amount of computation and some difficulties in application of the complicated image recognition. In this paper, we propose an enhanced fuzzy single layer Perceptron using the dynamic adjustment of threshold. This method is applied to the XOR problem, which used as the benchmark in the field of pattern recognition. The method is also applied to the recognition of digital image for image application. In a result of experiment, it does not always guarantee the convergence. However, the network show improved the learning time and has the high convergence rate.