• Communications for Statistical Applications and Methods
• /
• 제27권3호
• /
• pp.313-326
• /
• 2020

Deep Learning is the most important key to the development of Artificial Intelligence (AI). There are several distinguishable architectures of neural networks such as MLP, CNN, and RNN. Among them, we try to understand one of the main architectures called Recurrent Neural Network (RNN) that differs from other networks in handling sequential data, including time series and texts. As one of the main tasks recently in Natural Language Processing (NLP), we consider Neural Machine Translation (NMT) using RNNs. We also summarize fundamental structures of the recurrent networks, and some topics of representing natural words to reasonable numeric vectors. We organize topics to understand estimation procedures from representing input source sequences to predict target translated sequences. In addition, we apply multiple translation models with Gated Recurrent Unites (GRUs) in Keras on English-Korean sentences that contain about 26,000 pairwise sequences in total from two different corpora, colloquialism and news. We verified some crucial factors that influence the quality of training. We found that loss decreases with more recurrent dimensions and using bidirectional RNN in the encoder when dealing with short sequences. We also computed BLEU scores which are the main measures of the translation performance, and compared them with the score from Google Translate using the same test sentences. We sum up some difficulties when training a proper translation model as well as dealing with Korean language. The use of Keras in Python for overall tasks from processing raw texts to evaluating the translation model also allows us to include some useful functions and vocabulary libraries as well.

• Journal of the Korean Wood Science and Technology
• /
• 제37권2호
• /
• pp.148-154
• /
• 2009

This research examined the applicability of using an ultrasonic test to reconstruct CT images of an ancient wooden building. Most of the columns in the building are severely deteriorated due to termite attacks or the effect of weathering. Ultrasonic CT images of the columns were used to create highly accurate digital reconstructions, despite a lack of the data caused by parts of the building walls being buried. Another semi-NDE technique, a drilling test based on resistography, was applied in order to verify the ultrasonic test results. The discrepancy in detection between two methods is believed to be due to the fundamental differences between two methods. The performance of the ultrasonic test was hindered by poor surface conditions and this technique tended to over-estimate the size of cavities produced by termites or other insects. Nevertheless, the deterioration detected was in many ways congruent with the drilling test results

• Journal of international Conference on Electrical Machines and Systems
• /
• 제3권3호
• /
• pp.317-324
• /
• 2014

The forces involved in the firing of the electromagnetic rail gun may be analyzed from Amperian, Maxwellian and Einsteinian approaches. This paper discusses these different paradigms with regard to rail gun performance modeling relating to the generation and balance of the forces caused by the currents and their induced magnetic fields. Recent experimental work on model rail guns, where the armature is held static, shows very little recoil upon the rails, thereby indicating a possible violation of Newton's Third Law of Motion. Dynamic testing to show this violation, as suggested by the authors in an earlier paper, has inherent technical difficulties. A purpose-built finite element C/C++ simulator that models that suspended rail gun firing action shows a net force acting upon the entire rail gun system. A new effect in physics, universal in scope, is thus indicated: a current circulating in an asymmetric and rigid circuit causes a net force to act upon the circuit for the duration of the current. This conclusion following from computer simulation based upon Maxwellian electrodynamics as opposed to the more modern relativistic quantum electrodynamics needs to be supported by unambiguous experimental validation.

• 한국주거학회논문집
• /
• 제27권5호
• /
• pp.65-72
• /
• 2016

The purpose of the study was to investigate defect cases in residential interior design, to identify the state of defects, and to categorize types of defect and disputes. In addition, consumer appraisal to residential interior design service were analyzed. The results of this study could provide fundamental information regarding the defects and claims in residential interior design. First, we did literature review for defect disputes in architectural design and interior design. We identify the definition of defects by building life cycle, by state of construction, by activity, and by design performance. Second, we analysed interior design defects cases reported in Korea Consumer Agency & Ministry of Land, Infrastructure and Transport. A total of 49 cases of defect disputes in residential interior design from 2000 to 2015 were investigated. As a result, many defects appeared during the construction stage. A majority of defects cases fall into insulation, water-proofing/leakage work. In terms of design aspects claim, functional and aesthetic defects were claimed. Third, from Consumer Market Evaluation Indicators, raw data from 500 respondents were investigated for the housing repair and interior design. It is found that information comparability, responsiveness to consumer claim, price, and safety are important factors for consumer satisfaction in interior design.

• 한국수소및신에너지학회논문집
• /
• 제24권5호
• /
• pp.421-427
• /
• 2013

We report on the time dependent degradation of cell in dye-sensitized solar cells (DSSC). The photovoltaic performance of DSSC over a period of time was investigated in liquid electrolyte based on triiodide/iodide during six days. It was found that the short circuit current density ($j_{sc}$) of the cell dropped from 9.9 to $7mA/cm^2$ while efficiency (${\eta}$) of the cell decreased from 4.4 to 3.3%. The parameters corresponding to fundamental electronic and ionic processes in a working DSSC are determined from the electrochemical impedance spectrascopy (EIS) at open-circuit potential ($V_{oc}$). EIS study of the DSSC in the this work showed that the electron life time ${\tau}_r$ and chemical capacitance $C_{\mu}$ decreased significantly after six days. It was correlated the $j_{sc}$ and efficiency decreased after six days.

• 전기학회논문지
• /
• 제65권6호
• /
• pp.1108-1114
• /
• 2016

The cause of failure in the AT feeding system is divided into grounding, short-circuit of feeding circuit and internal faults of the railway substation. Since the fault current is very high, real-time current is detected and the failure must be immediately removed. In this paper, a new DC offset elimination filter that can remove component to decrease in the form of exponential function using low-pass filter was proposed in order to extract the fundamental wave from distorted fault current. In order to confirm the performance of the proposed filter method, AT feeder system was modelled by simulation tool and simulations were performed under various conditions such as fault location, fault resistance and fault voltage phase angle in case of trolley-rail short-circuit fault. When applying the proposed DC-offset removal method, it can be seen that the phase delay and gain error did not appear.

• 전기학회논문지
• /
• 제58권12호
• /
• pp.2498-2504
• /
• 2009

The fundamental research for the mobile robot navigation using the numerical optimization method is presented. We propose an image-based visual servo navigation algorithm for a wheeled mobile robot utilizing a ceiling mounted camera. For the image-based visual servoing, we define the composite image Jacobian which represents the relationship between the speed of wheels of a mobile robot and the robot's overall speed in the image plane. The rotational speed of wheels of a mobile robot can be directly related to the overall speed of a mobile robot in the image plane using the composite image Jacobian. We define the mobile robot navigation problem as an unconstrained optimization problem to minimize the cost function with the image error between the goal position and the position of a mobile robot. In order to avoid the obstacle, the modified cost function is proposed which is composed of the image error between the position of a mobile robot and the goal position and the distance between the position of a mobile robot and the position of the obstacle. The performance was evaluated using the simulation.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2008년도 제39회 하계학술대회
• /
• pp.1291-1292
• /
• 2008

ITO is widely used as an anode material in OLED, because of its good transparency, low electrical resistivity, high work function, and efficient hole injection properties. The interface between the electrode and the organic layer in the OLED effects the charge injection process and may influence the electrical and the luminance properties. Surface treatment of ITO, such as an Aquaregia treatment has been shown to enhance the efficiency, and luminance characteristics of the OLED. In this study, we investigated the effect of Aquaregia treatment. The fundamental structures of the OLED were ITO/NPB/$Alq_3$/LiF/Al. The current density-voltage-luminance, efficiency, and lifetime characteristics were measured with untreated and Aquaregia-treated ITO substrates. The Aquaregia treatment was found to enhance the performance of OLED. For the Aquaregia treated device, the maximum luminance and efficiency were increased by about 2 times compared to the untreated device. The mechanism of the Aquaregia treatment is discussed based on AFM analyses.

• Geomechanics and Engineering
• /
• 제14권3호
• /
• pp.257-269
• /
• 2018

Analysis of slope stability failures, as one of the complex natural hazards, is one of the important research issues in the field of civil engineering. Present paper adopts and investigates four soft computing-based techniques for this problem: Patient Rule-Induction Method (PRIM), M5' algorithm, Group Method of data Handling (GMDH) and Multivariate Adaptive Regression Splines (MARS). A comprehensive database consisting of 168 case histories is used to calibrate and test the developed models. Six predictive variables including slope height, slope angle, bulk density, cohesion, angle of internal friction, and pore water pressure ratio were considered to generate new models. The results of test studies are used for feasibility, effectiveness and practicality comparison of techniques with each other, and with the other available well-known methods in the literature. Results show that all methods not only are feasible but also result in better performance than previously developed soft computing based predictive models and tools. It is shown that M5' and PRIM algorithms are the most effective and practical prediction models.

• 한국전산유체공학회지
• /
• 제14권3호
• /
• pp.69-75
• /
• 2009

The objective of this study is to investigate the suitable design for a domestic Circulating water pump(CWP), which is used in cooling-water intakes for the unit 3 and 4 of Yeonggwang nuclear power plant. All the simulations are performed, using CFD method with a commercial code STAR-CCM+ version 3.02. After modeling a present design of the pump, the flow around the rotating blade was calculated by using quasi-static method and sliding mesh method with the almost same condition as an actual state. Based on fundamental simulations with various depth of sea water, the reference pressure for the boundary condition of the present study was decided. To verify the reliability of the calculation results, the suction flow rate of the data was compared with that of the experimental data. As a result of this comparison, it is confirmed that two results are fairly consistent. For the improvement of the suction flow rate, computational analysis was done by changing a flow channel and blade shapes. It is shown that the suction flow rate of the new pump was improved.