• Electronics and Telecommunications Trends
• /
• v.33 no.4
• /
• pp.1-13
• /
• 2018

Deep learning (DL) models have been widely applied to AI applications such image recognition and language translation with big data. Recently, DL models have becomes larger and more complicated, and have merged together. For the accelerated training of a large-scale deep learning model, model parallelism that partitions the model parameters for non-shared parallel access and updates across multiple machines was provided by a few distributed deep learning frameworks. Model parallelism as a training acceleration method, however, is not as commonly used as data parallelism owing to the difficulty of efficient model parallelism. This paper provides a comprehensive survey of the state of the art in model parallelism by comparing the implementation technologies in several deep learning frameworks that support model parallelism, and suggests a future research directions for improving model parallelism technology.

• Proceedings of the IEEK Conference
• /
• 1998.06a
• /
• pp.106-109
• /
• 1998

Wireless communication systems, in particular, must operate in a crowded electro-magnetic environmnet where in-band undesired signals are treated as noise by the receiver. These interfering signals are often random but not Gaussian Due to nongaussian noise, the distribution of the observables cannot be specified by a finite set of parameters; instead r-dimensioal sample space (pure noise samples) is equiprobably partitioned into a finite number of disjointed regions using quantiles and a vector quantizer based on training samples. If we assume that the detected symbols are correct, then we can observe the pure noise samples during the training and transmitting mode. The algorithm proposed is based on a piecewise approximation to a regression function based on quantities and conditional partition moments which are estimated by a RMSA (Robbins-Monro Stochastic Approximation) algorithm. In this paper, we develop a diversity combiner with modified detector, called Non-Linear Detector, and the receiver has a differential phase detector in each diversity branch and at the combiner each detector output is proportional to the second power of the envelope of branches. Monte-Carlo simulations were used as means of generating the system performance.

• The Journal of the Acoustical Society of Korea
• /
• v.15 no.4E
• /
• pp.28-36
• /
• 1996

In this paper, we propose a new evaluation methodology which can more clearly show the performance of the allophone modeling algorithm generally used in large vocabulary speech recognition. The proposed evaluation method shows the running characteristics and limitations of the modeling algorithm by testing how the variation of phonetic environments of training data affects the recognition performance and the desirable number of free parameters to be estimated. Using the method, we experiment results, we conclude that, in vocabulary-independent recognition task, the phonetic diversity of training data greatly affects the robustness of model, and it is necessary to develop a proper measure which can determine the number of states compromizing the robustness and the precision of the HMnet better than the conventional modeling efficiency.

• Journal of Institute of Control, Robotics and Systems
• /
• v.11 no.7
• /
• pp.565-571
• /
• 2005

We present a novel speed estimation of a Permanent Magnet Synchronous Motor(PMSM) based on Support Vector Regression(SVR). The proposed method can estimate wide speed range, including 0.33Hz with full load, accurately in the steady and transient states where motor parameters variations are known without parameter estimator. Moreover, the method does not need offline training previously but is trained on-line. The training starts with the PMSM operation simultaneously and estimates the speed in real time. The experimental results shows the validity and the usefulness of the proposed algorithm for the 0.4Kw PMSM DSP(TMS320VC33) drive system.

• JSTS:Journal of Semiconductor Technology and Science
• /
• v.6 no.3
• /
• pp.182-188
• /
• 2006

An artificial neural network model for the microwave characteristics of an InGaAs/InP hemt for 70 nm gate length has been developed. The small-signal microwave parameters have been evaluated to determine the transconductance and drain-conductance. We have further investigated the frequency characteristics of the device. The neural network training have been done using the three layer architecture using Levenberg-Marqaurdt Backpropagation algorithm. The results have been compared with the experimental data, which shows a close agreement and the validity of our proposed model.

• The Bulletin of The Korean Astronomical Society
• /
• v.42 no.2
• /
• pp.60.1-60.1
• /
• 2017

In this study we apply Convolution Neural Network(CNN) to solar flare occurrence prediction with various parameter options using the 00:00 UT MDI images from 1996 to 2010 (total 4962 images). We assume that only X, M and C class flares correspond to "flare occurrence" and the others to "non-flare". We have attempted to look for the best options for the models with two CNN pre-trained models (AlexNet and GoogLeNet), by modifying training images and changing hyper parameters. Our major results from this study are as follows. First, the flare occurrence predictions are relatively good with about 80 % accuracies. Second, both flare prediction models based on AlexNet and GoogLeNet have similar results but AlexNet is faster than GoogLeNet. Third, modifying the training images to reduce the projection effect is not effective. Fourth, skill scores of our flare occurrence model are mostly better than those of the previous models.

• Structural Engineering and Mechanics
• /
• v.51 no.2
• /
• pp.299-313
• /
• 2014

In the first part of this paper, the influences of some of crack parameters on natural frequencies of a cracked cantilever Functionally Graded Beam (FGB) are studied. A cantilever beam is modeled using Finite Element Method (FEM) and its natural frequencies are obtained for different conditions of cracks. Then effect of variation of depth and location of cracks on natural frequencies of FGB with single and multiple cracks are investigated. In the second part, two Multi-Layer Feed Forward (MLFF) Artificial Neural Networks (ANNs) are designed for prediction of FGB's Cracks' location and depth. Particle Swarm Optimization (PSO) and Back-Error Propagation (BEP) algorithms are applied for training ANNs. The accuracy of two training methods' results are investigated.

• Proceedings of the Korean Society of Soil and Groundwater Environment Conference
• /
• 2004.09a
• /
• pp.307-311
• /
• 2004

The unsaturated zone is a significant pathway of the surface contaminant movement and is a highly heterogeneous medium. Therefore, there are limitations in applying conventional convection-dispersion equation(CDE). Artificial neural network(ANN) is considered to be a versatile tool for approximating complex functions. For evaluating the applicability of ANN, numerical tests using ANN were conducted with training set generated by HYDRUS-2D which is based on CDE. The results represent that ANN can estimate the solute transport and the choice of network parameters and generation of training set patterns are important for efficient estimation.

• Journal of the Korean Society for Aviation and Aeronautics
• /
• v.28 no.1
• /
• pp.59-65
• /
• 2020

This paper presents comparative results of configuration modeling and aerodynamic analysis to single-engine airplanes such as C-172, SR-20, and DA40NG. The software OpenVSP was used as an airplane configuration modeling tool. OpenVSP can provide the fastest method to get three-dimensional aircraft configuration from given basic data and drawings of aircraft. Parametric design input in OpenVSP, from given aircraft geometric parameters, was applied to small airplanes mentioned. New aircraft models in this study were reversely designed to coincide with the publicly obtained dimensions of the original aircraft. The basic aerodynamic analysis of newly designed modeling aircraft was performed by the vortex lattice method. Results are shown that the similarity of aerodynamic data obtained except for the lack of DA40NG. In conclusion, the modeling process applied to this work is valuable to obtain conceptual design insight in the reverse design from the small airplanes currently in use for flight training.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2004.05b
• /
• pp.1285-1289
• /
• 2004

Elman Discrete Recurrent Neural Networks Model(EDRNNM) was used to be a suitable short-term hydrological forecasting tool yielding a very high degree of flood stage forecasting accuracy at Musung station of Wi-stream one of IHP representative basins in South Korea. A relative new approach method has recurrent feedback nodes and virtual small memory in the structure. EDRNNM was trained by using two algorithms, namely, LMBP and RBP The model parameters, optimal connection weights and biases, were estimated during training procedure. They were applied to evaluate model validation. Sensitivity analysis test was also performed to account for the uncertainty of input nodes information. The sensitivity analysis approach could suggest a reduction of one from five initially chosen input nodes. Because the uncertainty of input nodes information always result in uncertainty in model results, it can help to reduce the uncertainty of EDRNNM application and management in small catchment.