• Journal of Information Technology Services
• /
• v.21 no.2
• /
• pp.85-95
• /
• 2022

This study analyzed the translation error factors of machine translation services such as Naver Papago and Google Translate through Self-Attention path visualization. Self-Attention is a key method of the Transformer and BERT NLP models and recently widely used in machine translation. We propose a method to explain translation error factors of machine translation algorithms by comparison the Self-Attention paths between ST(source text) and ST'(transformed ST) of which meaning is not changed, but the translation output is more accurate. Through this method, it is possible to gain explainability to analyze a machine translation algorithm's inside process, which is invisible like a black box. In our experiment, it was possible to explore the factors that caused translation errors by analyzing the difference in key word's attention path. The study used the XLM-RoBERTa multilingual NLP model provided by exBERT for Self-Attention visualization, and it was applied to two examples of Korean-Chinese and Korean-English translations.

• Proceedings of the KSR Conference
• /
• 2000.05a
• /
• pp.100-107
• /
• 2000

This paper proposes a new model for harmonic analysis in 2${\times}$25㎸ traction power supply system including inverted feeder, contact line, rails and auto-transformer. The system model is based on four-port representation which is an extension of two-port network theory. In order to verify the proposed approach, we have analysed and tested real traction power feeding system focused on the amplification of harmonic current. The calculation results front tile proposed approach and the measurement data from the test are widely described in the paper.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.05a
• /
• pp.415-417
• /
• 2021

Since the words and words in Chinese sentences are continuous and the length of vocabulary is huge, Chinese NER(Named Entity Recognition) always based on character representation. In recent years, many Chinese research has been reconsidered how to integrate the word information into the Chinese NER model. However, the traditional sequence model has complex structure, the slow inference speed, and an additional dictionary information is needed, which is difficult to implement in the industry. The approach in this paper has the state of the art and parallelizable, which is integrated the char-word embeddings, so that the model learns word information. The proposed model is easy to implement, and outperforms traditional model in terms of speed and efficiency, which is improved f1-score on two dataset.

• 제어로봇시스템학회:학술대회논문집
• /
• 2004.08a
• /
• pp.1042-1047
• /
• 2004

Voltage control is an essential part of the electric energy transmission and distribution system to maintain proper voltage limit at the consumer's terminal. Besides the generating units that provide the basic voltage control, there are many additional voltage-controlling agents e.g., shunt capacitors, shunt reactors, static VAr compensators, regulating transformers mentioned in [1], [2]. The most popular one, among all those agents for controlling voltage levels at the distribution and transmission system, is the on-load tap changer transformer. It serves two functions-energy transformation in different voltage levels and the voltage control. Artificial Neural Network (ANN) has been realized as a convenient tool that can be used in controlling the on load tap changer in the distribution transformers. Usage of the ANN in this area needs suitable training and testing data for performance analysis before the practical application. This paper briefly describes a procedure of processing the data to train an Artificial Neural Network (ANN) to control the tap changer operating decision of parallel transformers for a closed primary bus. The data set are used to train a two layer ANN using three different neural net learning algorithms, namely, Standard Backpropagation [3], Bayesian Regularization [4] and Scaled Conjugate Gradient [5]. The experimental results are presented including performance analysis.

• The Proceedings of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.7 no.6
• /
• pp.34-41
• /
• 1993

This paper describes an effort to develop a microcomputer-based parameter identification system for three phase and two phase brushless DC motor. Back EMF equation is derived from back EMF waveform of three phase and two phase brushless DC motor. In this paper, a new identification algorithm for the brushless DC motor parameters by Pasek's technique is de veloped. It is found that Pasek's equation is valid for the brushless DC motor, too. The results obtained clearly shows that it is possible to implement the identification system for the determination of the brushless DC motor parameters. To minimize errors due to the ripple component in the measured armature current, dlgital averaging filter is employed. The whole identification process of signal generation, measuring, parameter determination is fully automated. The use of the pmpased method in the parameter identification system shows that the averaged current curve is in excellent agreement with the estimated current curve. Therefore, this close agreement conf i i the validity of this technique.

• The Transactions of the Korean Institute of Power Electronics
• /
• v.7 no.6
• /
• pp.596-603
• /
• 2002

In this paper a new harmonic reduction method of three-phase diode rectifiers is proposed to improve input current performance using the zero-sequence harmonics injection technique. The proposed mothed, based on the third-harmonic injection, employs two half-bridge inverters and two single-phase transformers to independently shape the positive and negative dc rail currents of the diode rectifier. The actively shaped zero-sequence harmonic currents are t]ten circulated through the ac side of the rectifier using a zigzag transformer This results in pure sinusoidal input currents in the three-phase diode rectifier. Experimental results on a 1.5kVA prototype are provided to validate the proposed technique.

• The Transactions of the Korean Institute of Power Electronics
• /
• v.23 no.3
• /
• pp.208-216
• /
• 2018

In this paper, an LLC resonant converter employing two coupled inductors on the secondary side of the converter is proposed. The conventional LLC converter exhibits serious power loss during secondary winding of the transformer because of generation of tremendous output current ripples. To overcome this problem, an LLC resonant converter with a current doubler as a rectifying circuit was recently proposed. However, the current-doubler rectifying circuit requires coupled inductors with a high coupling ratio to retain the designed resonance characteristics. Therefore, an additional hardware filter is required at the output stage to address large output current ripples. Additional design procedures are also necessary because the inductance component of the added filter affects the designed resonant network. To solve this issue, an LLC resonant converter employing two coupled inductors is proposed in this paper. Mathematical analysis shows that the proposed secondary-side current-doubler circuit does not affect the designed resonance characteristics. The operating principles and theoretical analyses are proven through a simulation and experiments with a 54 V/28 A prototype.

• Proceedings of the KIEE Conference
• /
• 1989.07a
• /
• pp.202-206
• /
• 1989

A new algorithm is suggested to solve the optimal reactive power control(optimal VAR control) problem. An efficient computer program based on the latest achievements in the sparse matrix/vector techniques has been developed for this purpose. The model minimizes the real power losses in the system. The constraints include the reactive power limits of the generators, limits on the bus voltages and the operating limits of control variables- the transformer tap positions, generator terminal voltages and switchable reactive power sources. The method developed herein employs linearized sensitivity relationships of power systems to establish both the objective function for minimizing the system losses and the system performance sensitivities relating dependent and control variables. The algorithm consists of two modules, i.e. the Q-V module for reactive power-voltage control, Load flow module for computational error adjustments. In particular, the acceleration factor technique is introduced to enhance the convergence property in Q-module, The combined use of the afore-mentioned two modules ensures more effective and efficient solutions for optimal reactive power dispatch problems. Results of the application of the method to the sample system and other worst-case system demonstrated that the algorithm suggested herein is compared favourably with conventional ones in terms of computation accuracy and convergence characteristics.

• KIEE International Transaction on Electrical Machinery and Energy Conversion Systems
• /
• v.5B no.4
• /
• pp.343-349
• /
• 2005

This paper proposes a novel zero-voltage-switching (ZVS) Push-pull DC-DC Converter for high input voltage and high power applications. This topology utilizes two switches in series to replace one switch in conventional push-pull converter, and two clamping diodes are introduced. The voltage stress of the switches is the input voltage, and the switches can realize ZVS with the use of the leakage inductance of the transformer. Furthermore, secondary full-wave rectifier with a clamping capacitor is used to eliminate the voltage oscillation and spike of the rectifier diodes due to the reverse recovery. Therefore, the electromagnetic interference is reduced effectively. The operation principle of the proposed converter is analyzed theoretically. The output characteristic, ZVS condition and design principle of the clamping capacitor are discussed. Experimental results obtained from a 270V input 2kW prototype with $95.8\%$ high efficiency confirms the design.

• The Transactions of the Korean Institute of Electrical Engineers P
• /
• v.57 no.1
• /
• pp.1-5
• /
• 2008

The output voltage value of AC high voltage source has been usually obtained by multiplying low voltage value measured at both terminals of low voltage resistor by the dividing ratio of the high voltage capacitive divider. From the dividing ratio determined of each 200 kV capacitive divider, we have developed step-up method for measuring the output voltage up to 400 kV using two same type of 200 kV capacitive dividers connected in series. The theoretical dividing ratio of 400 kV capacitive dividers connected in series coincides with that of manufacturer's certification within measurement uncertainty. Thus, this developed step-up method makes it possible to extend the range of output voltage from 200 kV to 400 kV. Furthermore, The dividing ratio of divider under test obtained using this step-up method is consistent with that obtained using one 200 kV high voltage divider within corresponding uncertainties.