• Journal of Digital Contents Society
• /
• v.18 no.5
• /
• pp.969-976
• /
• 2017

This paper proposes the learning method of an artificial neural network and a convolutional neural network using the WFSO algorithm developed as an optimization algorithm. Since the optimization algorithm searches based on a number of candidate solutions, it has a drawback in that it is generally slow, but it rarely falls into the local optimal solution and it is easy to parallelize. In addition, the artificial neural networks with non-differentiable activation functions can be trained and the structure and weights can be optimized at the same time. In this paper, we describe how to apply WFSO algorithm to artificial neural network learning and compare its performances with error back-propagation algorithm in multilayer artificial neural networks and convolutional neural networks.

• Bulletin of the Korean Chemical Society
• /
• v.34 no.1
• /
• pp.107-111
• /
• 2013

A new green-emitting material with donor-acceptor architecture, 3,7-bis(1'-phenylbenzimidazole-2'-yl)-10-phenylphenothiazine (BBPP) was synthesized and its thermal, optical, and electroluminescent characteristics were investigated. Organic light-emitting diodes (OLEDs) with four different multilayer structures were prepared using BBPP as an emitting layer. The optimized device with the structure of [ITO/2-TNATA (40 nm)/BBPP (30 nm)/TPBi (30 nm)/Alq3 (10 nm)/LiF (1 nm)/Al (100 nm)] exhibited efficient green emission. Enhanced charge carrier balance and electron mobility in the organic layers enabled the device to demonstrate a maximum luminance of 31,300 cd/$m^2$, a luminous efficiency of 6.83 cd/A, and an external quantum efficiency of 1.62% with the CIE 1931 chromaticity coordinates of (0.21, 0.53) at a current density of 100 mA/$cm^2$.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.53 no.8
• /
• pp.551-560
• /
• 2004

In this paper, we propose a new architecture of Self-Organizing Fuzzy Polynomial Neural Networks (SOFPNN) that is based on a genetically optimized multilayer perceptron with fuzzy polynomial neurons (FPNs) and discuss its comprehensive design methodology involving mechanisms of genetic optimization, especially genetic algorithms (GAs). The proposed SOFPNN gives rise to a structurally optimized structure and comes with a substantial level of flexibility in comparison to the one we encounter in conventional SOFPNNs. The design procedure applied in the construction of each layer of a SOFPNN deals with its structural optimization involving the selection of preferred nodes (or FPNs) with specific local characteristics (such as the number of input variables, the order of the polynomial of the consequent part of fuzzy rules, and a collection of the specific subset of input variables) and addresses specific aspects of parametric optimization. Through the consecutive process of such structural and parametric optimization, an optimized and flexible fuzzy neural network is generated in a dynamic fashion. To evaluate the performance of the genetically optimized SOFPNN, the model is experimented with using two time series data(gas furnace and chaotic time series), A comparative analysis reveals that the proposed SOFPNN exhibits higher accuracy and superb predictive capability in comparison to some previous models available in the literatures.

• Proceedings of the KIEE Conference
• /
• 2003.07d
• /
• pp.2519-2521
• /
• 2003

In this paper, we propose Self-Organizing Fuzzy Polynomial Neural Networks(SOFPNN) architecture for optimal model identification and discuss a comprehensive design methodology supporting its development. It is shown that this network exhibits a dynamic structure as the number of its layers as well as the number of nodes in each layer of the SOFPNN are not predetermined (as this is the case in a popular topology of a multilayer perceptron). As the form of the conclusion part of the rules, especially the regression polynomial uses several types of high-order polynomials such as linear, quadratic, and modified quadratic. As the premise part of the rules, both triangular and Gaussian-like membership function are studied and the number of the premise input variables used in the rules depends on that of the inputs of its node in each layer. We introduce two kinds of SOFPNN architectures, that is, the basic and modified one with both the generic and the advanced type. The superiority and effectiveness of the proposed SOFPNN architecture is demonstrated through nonlinear function numerical example.

• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.31 no.6
• /
• pp.408-411
• /
• 2018

We fabricated highly flexible Mn-doped $SnO_2$ (MTO)/Ag/MTO/polydimethylsiloxane (PDMS)/MTO multilayer transparent conducting films. To reduce refractive-index mismatching of the MTO/Ag/MTO/polyethylene terephthalate (PET), index-matching layers were inserted between the oxide-metal-oxide-structured films and the PET substrate. The PDMS layer was deposited by spin-coating after adjusting the mixing ratio of PDMS and hexane. We investigated the effects of the index-matching layer on the color and reflectance differences with different PDMS dilution ratios. As the dilution ratio increased from 1:100 to 1:130, the color difference increased slightly, while the reflectance difference decreased from 0.62 to 0.32. The MTO/Ag/MTO/PDMS/MTO film showed a transmittance of 87.18~87.68% at 550 nm. The highest value of the Haacke figure of merit was $47.54{\times}10^{-3}{\Omega}^{-1}$ for the dilution ratio of 1:130.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2013.08a
• /
• pp.278.1-278.1
• /
• 2013

Among the prerequisites for stable neural interfacing are the long-term stability of electrical performance of and the excellent biocompatibility of conducting materials in implantable neural electrodes. Reduced graphene oxide offers a great potential for a variety of biomedical applications including biosensors and, particularly, neural interfaces due to its superb material properties such as high electrical conductivity, decent optical transparency, facile processibility, and etc. Nonetheless, there have been few systematic studies on the graphene-based neural interfaces in terms of biocompatibility of electrode materials and long term stability in electrical characteristics. In this research, we prepared the primary culture of rat hippocampal neurons directly on reduced graphene oxide films which is chosen as a model electrode material for the neural electrode. We observed that the viability of primary neuronal culture on the present structure is minimally affected by nanoscale graphene flakes below. These results implicate that the multilayer films of reduced graphene oxides can be utilized for the next-generation neural interfaces with decent biocompatibility and outstanding electrical performance.

• Journal of the Korean Physical Society
• /
• v.73 no.9
• /
• pp.1279-1282
• /
• 2018

This paper reports the deposition of a metal line using a multilayer stack and laser-induced forward transfer (LIFT) using a low cost continuous wave blue laser with a wavelength of 450 nm. The donor structure was composed of a light-to-heat (LTH) layer, a release layer, and a transfer layer in series. Amorphous silicon as the LTH layer absorbs photon energy and converts it to heat. A release layer was melted so that a silver transfer layer would be transferred to the receiver substrate. The transferred silver layer showed reasonable physical and electrical characteristics. A low cost fine linewidth metal layer could be achieved using this modified LIFT technique and blue laser.

• Journal of Surface Science and Engineering
• /
• v.52 no.4
• /
• pp.211-226
• /
• 2019

Optical thin films were deposited by using a reactive pulsed DC magnetron sputtering method with a high density plasma(HDP). In this study, the effect of sputtering process conditions on the microstructure and optical properties of $SiO_2$, $TiO_2$, $Nb_2O_5$ thin films was clarified. These thin films had flat and dense microstructure, stable stoichiometric composition at the optimal conditions of low working pressure, high pulsed DC power and RF power(HDP). Also, the refractive index of the $SiO_2$ thin films was almost constant, but the refractive indices of $TiO_2$ and $Nb_2O_5$ thin films were changed depending on the microstructure of these films. Antireflection films of $Air/SiO_2/Nb_2O_5/SiO_2/Nb_2O_5/SiO_2/Nb_2O_5/Glass$ structure designed by Macleod program were manufactured by our developed sputtering system. Transmittance and reflectance of the manufactured multilayer films showed outstanding value with the level of 95% and 0.3%, respectively, and also had excellent durability.

• Journal of the Korean Magnetics Society
• /
• v.5 no.6
• /
• pp.963-967
• /
• 1995

Ultra thin layers of NiFeCo or NiFe were inserted at the interfaces of Ni and Cu to form a multilayer structure. In case of inserting a NiFe layer, the magnetoresistance was about 6%, the saturation magnetic field was 50 Oe and the hysteresis of R-H (resistance-magnetic field) was very small. In case of inserting a NiFeCo layer, the magnetoresistance increased to about 7% but the saturation magnetic field and hysteresis were also increased. The increase of the output under biased magnetic field was much larger in case of inserting a NiFe layer because of relatively smaller hysteresis in R-H behavior.

• Korean Journal of Metals and Materials
• /
• v.48 no.4
• /
• pp.353-356
• /
• 2010

InSb thin films with a thickness of approximately 300 nm were prepared using single- and double-source vacuum evaporation methods and their structures and properties were investigated in terms of a heat treatment procedure. The double-source InSb films, prepared by the alternate stacking of In and Sb, were polycrystalline in structure and included small amounts of unreacted In and Sb phases. After annealing at elevated temperatures below the melting point of InSb (525$^{\circ}C$), the films changed into the InSb phase and were found to contain small amounts of unreacted In. The formation capability of the InSb compound was slightly lower for multilayer films than for single-layer films. The electrical and galvanomagnetic properties were found to be strongly related to the microstructures of the films. The maximum value of the Hall mobility and the magnetoresistance were determined to be $4.3{\times}10^3cm^2$/Vs and 70%, respectively, for the single-layer films, while these values for the alternately stacked films were respectively $2.9{\times}10^3cm^2$/Vs and 29% for the $[Sb(2.5)/In(2.5)]_{60}$ films, and $3.1{\times}10^3cm^2$/Vs and 10% for the $[Sb(150)/In(150)]_1$ films.