• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.66 no.12
• /
• pp.1772-1781
• /
• 2017

In this study, the design methodology for alleviating the overfitting problem of Polynomial Neural Networks(PNN) is realized with the aid of two kinds techniques such as L2 regularization and Sum of Squared Coefficients (SSC). The PNN is widely used as a kind of mathematical modeling methods such as the identification of linear system by input/output data and the regression analysis modeling method for prediction problem. PNN is an algorithm that obtains preferred network structure by generating consecutive layers as well as nodes by using a multivariate polynomial subexpression. It has much fewer nodes and more flexible adaptability than existing neural network algorithms. However, such algorithms lead to overfitting problems due to noise sensitivity as well as excessive trainning while generation of successive network layers. To alleviate such overfitting problem and also effectively design its ensuing deep network structure, two techniques are introduced. That is we use the two techniques of both SSC(Sum of Squared Coefficients) and $L_2$ regularization for consecutive generation of each layer's nodes as well as each layer in order to construct the deep PNN structure. The technique of $L_2$ regularization is used for the minimum coefficient estimation by adding penalty term to cost function. $L_2$ regularization is a kind of representative methods of reducing the influence of noise by flattening the solution space and also lessening coefficient size. The technique for the SSC is implemented for the minimization of Sum of Squared Coefficients of polynomial instead of using the square of errors. In the sequel, the overfitting problem of the deep PNN structure is stabilized by the proposed method. This study leads to the possibility of deep network structure design as well as big data processing and also the superiority of the network performance through experiments is shown.

• The Bulletin of The Korean Astronomical Society
• /
• v.41 no.1
• /
• pp.75.3-75.3
• /
• 2016

We present a new survey of nearby galaxies to obtain deep wide-field images of 200 nearby bright galaxies in the southern hemisphere using Korea Microlensing Telescope Network (KMTNet). We are taking very deep and wide-field images, spending 4.5 hours for the B and R filters for each object. Using this dataset, we will look for diffuse, low-surface brightness structures including outer disks, truncated disks, tidal features and stellar streams, and faint companions. The multicolor data will enable us to estimate the incidence and star formation history of those features. We present an outline of the data reduction pipeline, and preliminary results from the commissioning data.

• Imaging Science in Dentistry
• /
• v.50 no.4
• /
• pp.331-337
• /
• 2020

Purpose: As cone-beam computed tomography (CBCT) has become the most widely used 3-dimensional (3D) imaging modality in the dental field, storage space and costs for large-capacity data have become an important issue. Therefore, if 3D data can be stored at a clinically acceptable compression rate, the burden in terms of storage space and cost can be reduced and data can be managed more efficiently. In this study, a deep learning network for super-resolution was tested to restore compressed virtual CBCT images. Materials and Methods: Virtual CBCT image data were created with a publicly available online dataset (CQ500) of multidetector computed tomography images using CBCT reconstruction software (TIGRE). A very deep super-resolution (VDSR) network was trained to restore high-resolution virtual CBCT images from the low-resolution virtual CBCT images. Results: The images reconstructed by VDSR showed better image quality than bicubic interpolation in restored images at various scale ratios. The highest scale ratio with clinically acceptable reconstruction accuracy using VDSR was 2.1. Conclusion: VDSR showed promising restoration accuracy in this study. In the future, it will be necessary to experiment with new deep learning algorithms and large-scale data for clinical application of this technology.

• The Bulletin of The Korean Astronomical Society
• /
• v.43 no.1
• /
• pp.42.1-42.1
• /
• 2018

Deep Ecliptic Patrol of the Southern Sky (DEEP-South) observation is being made during the off-season for exoplanet survey, using Korea Microlensing Telescope Network (KMTNet). An optimal combination of its prime focus optics and the 0.3 billion pixel CCD provides a four square degrees field of view with 0.4 arcsec/pixel plate scale which is also best suited for small body studies. Normal operation of KMTNet started in October 2015, and a significant portion of the allocated telescope time for DEEP-South is dedicated to targeted observation, Opposition Census (OC), of near-Earth asteroids for physical and taxonomic characterization. This is effectively achieved through multiband, time series photometry using Johnson-Cousins BVRI filters. Uninterrupted monitoring of the southern sky with KMTNet is optimized for spin characterization of a broad spectrum of asteroids ranging from the near-Earth space to the main-belt, including binaries, asteroids with satellites, slow/fast- and non-principal axis-rotators, and thus is expected to facilitate the debiasing of previously reported lightcurve observations. Our software subsystem consists of an automated observation scheduler, a pipelined data processing system for differential photometry, and an easy-to-use lightcurve analysis toolkit. Lightcurves, spin periods and provisional determination of class of asteroids to which the lightcurve belongs will be presented, using the dataset from first year operation of KMTNet. Our new taxonomic classification scheme for asteroids will also be summarized.

• The Bulletin of The Korean Astronomical Society
• /
• v.46 no.1
• /
• pp.42.1-42.1
• /
• 2021

In this study, we present a visual explanation of a deep learning solar flare forecast model and its relationship to physical parameters of solar active regions (ARs). For this, we use full-disk magnetograms at 00:00 UT from the Solar and Heliospheric Observatory/Michelson Doppler Imager and the Solar Dynamics Observatory/Helioseismic and Magnetic Imager, physical parameters from the Space-weather HMI Active Region Patch (SHARP), and Geostationary Operational Environmental Satellite X-ray flare data. Our deep learning flare forecast model based on the Convolutional Neural Network (CNN) predicts "Yes" or "No" for the daily occurrence of C-, M-, and X-class flares. We interpret the model using two CNN attribution methods (guided backpropagation and Gradient-weighted Class Activation Mapping [Grad-CAM]) that provide quantitative information on explaining the model. We find that our deep learning flare forecasting model is intimately related to AR physical properties that have also been distinguished in previous studies as holding significant predictive ability. Major results of this study are as follows. First, we successfully apply our deep learning models to the forecast of daily solar flare occurrence with TSS = 0.65, without any preprocessing to extract features from data. Second, using the attribution methods, we find that the polarity inversion line is an important feature for the deep learning flare forecasting model. Third, the ARs with high Grad-CAM values produce more flares than those with low Grad-CAM values. Fourth, nine SHARP parameters such as total unsigned vertical current, total unsigned current helicity, total unsigned flux, and total photospheric magnetic free energy density are well correlated with Grad-CAM values.

• Journal of Aerospace System Engineering
• /
• v.16 no.4
• /
• pp.17-27
• /
• 2022

In this study, a deep learning-based network that can predict the aerodynamic characteristics of airfoils was designed, and the feasibility of the proposed network was confirmed by applying aerodynamic data generated by Xfoil. The prediction of aerodynamic characteristics according to the variation of airfoil thickness was performed. Considering the angle of attack, the coordinate data of an airfoil is converted into image data using signed distance function. Additionally, the distribution of the pressure coefficient on airfoil is expressed as reduced data via proper orthogonal decomposition, and it was used as the output of the proposed network. The test data were constructed to evaluate the interpolation and extrapolation performance of the proposed network. As a result, the coefficients of determination of the lift coefficient and moment coefficient were confirmed, and it was found that the proposed network shows benign performance for the interpolation test data, when compared to that of the extrapolation test data.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.48 no.4
• /
• pp.295-301
• /
• 2020

When a spacecraft in deep space falls into an abnormal state, an emergency communication channel between ground and the spacecraft is essential in order to perform analysis to the cause of the anomaly, and to remedy the spacecraft from the distressed state. Because the recovery actions generally comprises of long and complicated sequences of commands, the transmission of the recovery commands may require a reliable and a delay tolerant networking technology based on bundle routing. While the delay tolerant networking protocol becomes a prominent method interfacing ground and space into a internet-like Solar system network because it can address the issues of the severe communication problems in deep space, the communication system on the spacecraft which based on space packet protocol cannot use the delay tolerant networking technology directly. So a community of the consultative committee for space data systems starts a discussion of the first-hop last-hop mechanism to establish a feasible concept and standardization. This paper presents an enhanced concept of the first-hop last-hop by applying it a virtual cislunar communication environment, and we believe this contributes to make a way applicable to an interoperable relay concept of the first hop last hop between the delay tolerant networking and space packet protocol standard.

• Journal of the Society of Disaster Information
• /
• v.6 no.1
• /
• pp.92-119
• /
• 2010

As the necessity to utilize deep-placed underground spaces is increasing, we have to seriously consider the safety problems arising from the U/G spaces which is a restricted environment. Due to the higher cost of land compensation for above ground area and environmental issues, the plan to utilize deep-placed U/G spaces is currently only being established for the construction of U/G road network and GTX. However it is also expected that the U/G spaces are to be used as a living space because of the growing desires to change the above ground areas into the environmentally green spaces. Accordingly it is necessary to protect the U/G environments which is vulnerable against desasters caused by fire, explosion, flooding, terrorism, electric power failure, etc. properly. We want to introduce the principles of the Quantitative Risk Assessment(QRA) method for preparedness against the desasters arising from U/G environments, and also want to introduce an example of QRA which was implemented for the GOTTHARD tunnel which is the longest one in Europe.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.6
• /
• pp.2612-2633
• /
• 2020

Due to the explosive growth of multimedia speech data, how to protect the privacy of speech data and how to efficiently retrieve speech data have become a hot spot for researchers in recent years. In this paper, we proposed an encrypted speech retrieval scheme based on long short-term memory (LSTM) neural network and deep hashing. This scheme not only achieves efficient retrieval of massive speech in cloud environment, but also effectively avoids the risk of sensitive information leakage. Firstly, a novel speech encryption algorithm based on 4D quadratic autonomous hyperchaotic system is proposed to realize the privacy and security of speech data in the cloud. Secondly, the integrated LSTM network model and deep hashing algorithm are used to extract high-level features of speech data. It is used to solve the high dimensional and temporality problems of speech data, and increase the retrieval efficiency and retrieval accuracy of the proposed scheme. Finally, the normalized Hamming distance algorithm is used to achieve matching. Compared with the existing algorithms, the proposed scheme has good discrimination and robustness and it has high recall, precision and retrieval efficiency under various content preserving operations. Meanwhile, the proposed speech encryption algorithm has high key space and can effectively resist exhaustive attacks.

• Journal of Astronomy and Space Sciences
• /
• v.34 no.2
• /
• pp.139-151
• /
• 2017

This paper presents an overview of deep space orbit determination software (DSODS), as well as validation and verification results on its event prediction capabilities. DSODS was developed in the MATLAB object-oriented programming environment to support the Korea Pathfinder Lunar Orbiter (KPLO) mission. DSODS has three major capabilities: celestial event prediction for spacecraft, orbit determination with deep space network (DSN) tracking data, and DSN tracking data simulation. To achieve its functionality requirements, DSODS consists of four modules: orbit propagation (OP), event prediction (EP), data simulation (DS), and orbit determination (OD) modules. This paper explains the highest-level data flows between modules in event prediction, orbit determination, and tracking data simulation processes. Furthermore, to address the event prediction capability of DSODS, this paper introduces OP and EP modules. The role of the OP module is to handle time and coordinate system conversions, to propagate spacecraft trajectories, and to handle the ephemerides of spacecraft and celestial bodies. Currently, the OP module utilizes the General Mission Analysis Tool (GMAT) as a third-party software component for high-fidelity deep space propagation, as well as time and coordinate system conversions. The role of the EP module is to predict celestial events, including eclipses, and ground station visibilities, and this paper presents the functionality requirements of the EP module. The validation and verification results show that, for most cases, event prediction errors were less than 10 millisec when compared with flight proven mission analysis tools such as GMAT and Systems Tool Kit (STK). Thus, we conclude that DSODS is capable of predicting events for the KPLO in real mission applications.