• Journal of KIISE:Databases
• /
• v.28 no.4
• /
• pp.732-741
• /
• 2001

In the relational database systems the join operation is one of the most time-consuming query operations. Many parallel join algorithms have been developed 개 reduce the execution time Multiple hash join algorithm using allocation tree is one of the most efficient ones. However, it may have some delay on the processing each node of allocation tree, which is occurred in tuple-probing phase by the difference between one page reading time of outer relation and the processing time of already read one. This delay problem was solved by using the concept of synchronization of page execution time with we had proposed In this paper the effects of the performance improvements in each node of the allocation tree are extended to the whole allocation tree and the performance evaluation about that is processed. In addition we propose an efficient algorithm for multiple hash joins in limited number of processor environments according to the relationship between the number of input relations in the allocation tree and the number of processors allocated to the tree. Finally. we analyze the performance by building the analytical cost model and verify the validity of it by various performance comparison with previous method.

• Resources Recycling
• /
• v.29 no.6
• /
• pp.3-14
• /
• 2020

This study aims to introduce and review on the recovery technologies of rare earth elements(REEs) from coal ash. Many researchers have been carried out by various beneficiation processes, such as particle size separation, magnetic separation, specific gravity, and flotation to recover rare earth elements from coal ash generated from Pulverized Coal(PC) boiler. Through the beneficiation process, it was confirmed that concentration of rare earth elements was much lower than the 4,700 ppm, and that additional enrichment treatment through wet process was needed for the products recovered after the beneficiation process. It was confirmed that the rare earth elements contained in coal ash were applied to the leaching process after pretreatment such as alkali-fusion to improve leaching efficiency. Although beneficiation and leaching methods have been studied, its optimum recovery technologies for rare earth elements not been confirmed up to now, research on the recovery of rare earth contained in coal ash is reported to continue. In case of Korea, the technology for the recovery of rare earth elements from coal ash and coal by-product could not been confirmed up to present. In these reasons, it is urgent to develop technologies such as beneficiation and leaching process continuously.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.10 no.4
• /
• pp.81-93
• /
• 2000

In a methodological approach to improve the processing performance of modulo exponentiation which is the primary arithmetic in RSA crypto algorithm, we present a new RSA hardware architecture based on high-radix modulo multiplication and CRT(Chinese Remainder Theorem). By implementing the modulo multiplier using radix-16 arithmetic, we reduced the number of PE(Processing Element)s by quarter comparing to the binary arithmetic scheme. This leads to having the number of clock cycles and the delay of pipelining flip-flops be reduced by quarter respectively. Because the receiver knows p and q, factors of N, it is possible to apply the CRT to the decryption process. To use CRT, we made two s/2-bit multipliers operating in parallel at decryption, which accomplished 4 times faster performance than when not using the CRT. In encryption phase, the two s/2-bit multipliers can be connected to make a s-bit linear multiplier for the s-bit arithmetic operation. We limited the encryption exponent size up to 17-bit to maintain high speed, We implemented a linear array modulo multiplier by projecting horizontally the DG of Montgomery algorithm. The H/W proposed here performs encryption with 15Mbps bit-rate and decryption with 1.22Mbps, when estimated with reference to Samsung 0.5um CMOS Standard Cell Library, which is the fastest among the publications at present.

• Journal of IKEEE
• /
• v.26 no.4
• /
• pp.633-638
• /
• 2022

Rencently, neuromorphic systems of spiking neural networks (SNNs) that imitate the human brain have attracted attention. Neuromorphic technology has the advantage of high speed and low power consumption in cognitive applications and processing. Resistive random-access memory (RRAM) for SNNs are the most efficient structure for parallel calculation and perform the gradual switching operation of spike-timing-dependent plasticity (STDP). RRAM as synaptic device operation has low-power processing and expresses various memory states. However, the integration of RRAM device causes high switching voltage and current, resulting in high power consumption. To reduce the operation voltage of the RRAM, it is important to develop new materials of the switching layer and metal electrode. This study suggested a optimized new structure that is the Metal/Al₂O₃/HfO_x/SWCNTs/N+silicon (MOCS) with single-walled carbon nanotubes (SWCNTs), which have excellent electrical and mechanical properties in order to lower the switching voltage. Therefore, we show an improvement in the gradual switching behavior and low-power I/V curve of SWCNTs-based memristors.

• Journal of the Korean Society of Radiology
• /
• v.17 no.5
• /
• pp.633-640
• /
• 2023

With the recent development of static and dynamic modulated brachytherapy methods in brachytherapy, which use radiation shielding to modulate the dose distribution to deliver the dose, the amount of parameters and data required for dose calculation in inverse treatment planning and treatment plan optimization algorithms suitable for new directional beam intensity modulated brachytherapy is increasing. Although intensity-modulated brachytherapy enables accurate dose delivery of radiation, the increased amount of parameters and data increases the elapsed time required for dose calculation. In this study, a GPU-based CUDA-accelerated dose calculation algorithm was constructed to reduce the increase in dose calculation elapsed time. The acceleration of the calculation process was achieved by parallelizing the calculation of the system matrix of the volume of interest and the dose calculation. The developed algorithms were all performed in the same computing environment with an Intel (3.7 GHz, 6-core) CPU and a single NVIDIA GTX 1080ti graphics card, and the dose calculation time was evaluated by measuring only the dose calculation time, excluding the additional time required for loading data from disk and preprocessing operations. The results showed that the accelerated algorithm reduced the dose calculation time by about 30 times compared to the CPU-only calculation. The accelerated dose calculation algorithm can be expected to speed up treatment planning when new treatment plans need to be created to account for daily variations in applicator movement, such as in adaptive radiotherapy, or when dose calculation needs to account for changing parameters, such as in dynamically modulated brachytherapy.

• Fire Science and Engineering
• /
• v.17 no.1
• /
• pp.8-20
• /
• 2003

Natural gas has been supplied through underground pipelines and valve stations as a new city gas in Seoul. In contrast to its handiness the natural gas has very substantial hazards due to fires and explosions occurring from careless treatments or malfunctions of the transporting system. The main objectives of this study are to identify major hazards and to perform risk assessments after assessing reliabilities of the composing units in dealing with typical pipeline networks. there-fore two method, fault tree analysis ;1nd event tree analysis, are used here. Random valve stations are selected and considered its situation in location. The value of small leakage, large rupture, and no supply of liquefied natural gas is estimated as that of top event. By this calculation the values of small leakage are 3.29 in I)C valve station, 1.41 in DS valve station, those of large rup-lure are $1.90Times10_{-2}$ in DC valve station, $2.32$\times$10^{-2}$ in DS valve station, and those of no supply of LNG to civil gas company are $2.33$\times$10 ^{-2}$ , $2.89$\times$10^{-2}$ in each valve station. And through minimal cut set we can find the parts that is important and should be more important in overall system. In DC valve station one line must be added between basic event 26,27 because the potential hazard of these parts is the highest value. If it is added the failure rate of no supply of LNG is reduced to one fourth. In DS valve station the failure rate of basic event 4 is 92eye of no supply of LNG. Therefore if the portion of this part is reduced (one line added) the total failure rate can be decreased to one tenth. This analytical study on the risk assessment is very useful to prepare emergency actions or procedures in case of gas accidents around underground pipeline networks and to establish a resolute gas safety management system for loss prevention in Seoul metropolitan area.

• Progress in Medical Physics
• /
• v.19 no.2
• /
• pp.102-112
• /
• 2008

We developed a high-resolution micro-CT system based on rotational gantry and flat-panel detector for live mouse imaging. This system is composed primarily of an x-ray source with micro-focal spot size, a CMOS (complementary metal oxide semiconductor) flat panel detector coupled with Csl (TI) (thallium-doped cesium iodide) scintillator, a linearly moving couch, a rotational gantry coupled with positioning encoder, and a parallel processing system for image data. This system was designed to be of the gantry-rotation type which has several advantages in obtaining CT images of live mice, namely, the relative ease of minimizing the motion artifact of the mice and the capability of administering respiratory anesthesia during scanning. We evaluated the spatial resolution, image contrast, and uniformity of the CT system using CT phantoms. As the results, the spatial resolution of the system was approximately the 11.3 cycles/mm at 10% of the MTF curve, and the radiation dose to the mice was 81.5 mGy. The minimal resolving contrast was found to be less than 46 CT numbers on low-contrast phantom imaging test. We found that the image non-uniformity was approximately 70 CT numbers at a voxel size of ${\sim}55{\times}55{\times}X100\;{\mu}^3$. We present the image test results of the skull and lung, and body of the live mice.

• The KIPS Transactions:PartC
• /
• v.13C no.1 s.104
• /
• pp.83-94
• /
• 2006

This paper describes ATM cell encipherment method using Rijndael Algorithm adopted as an AES(Advanced Encryption Standard) by NIST in 2001. ISO 9160 describes the requirement of physical layer data processing in encryption/decryption. For the description of ATM cell encipherment method, we implemented ATM data encipherment equipment which satisfies the requirements of ISO 9160, and verified the encipherment/decipherment processing at ATM STM-1 rate(155.52Mbps). The DES algorithm can process data in the block size of 64 bits and its key length is 64 bits, but the Rijndael algorithm can process data in the block size of 128 bits and the key length of 128, 192, or 256 bits selectively. So it is more flexible in high bit rate data processing and stronger in encription strength than DES. For tile real time encryption of high bit rate data stream. Rijndael algorithm was implemented in FPGA in this experiment. The boundary of serial UNI cell was detected by the CRC method, and in the case of user data cell the payload of 48 octets (384 bits) is converted in parallel and transferred to 3 Rijndael encipherment module in the block size of 128 bits individually. After completion of encryption, the header stored in buffer is attached to the enciphered payload and retransmitted in the format of cell. At the receiving end, the boundary of ceil is detected by the CRC method and the payload type is decided. n the payload type is the user data cell, the payload of the cell is transferred to the 3-Rijndael decryption module in the block sire of 128 bits for decryption of data. And in the case of maintenance cell, the payload is extracted without decryption processing.

• Journal of Intelligence and Information Systems
• /
• v.24 no.1
• /
• pp.205-225
• /
• 2018

Convolutional Neural Network (ConvNet) is one class of the powerful Deep Neural Network that can analyze and learn hierarchies of visual features. Originally, first neural network (Neocognitron) was introduced in the 80s. At that time, the neural network was not broadly used in both industry and academic field by cause of large-scale dataset shortage and low computational power. However, after a few decades later in 2012, Krizhevsky made a breakthrough on ILSVRC-12 visual recognition competition using Convolutional Neural Network. That breakthrough revived people interest in the neural network. The success of Convolutional Neural Network is achieved with two main factors. First of them is the emergence of advanced hardware (GPUs) for sufficient parallel computation. Second is the availability of large-scale datasets such as ImageNet (ILSVRC) dataset for training. Unfortunately, many new domains are bottlenecked by these factors. For most domains, it is difficult and requires lots of effort to gather large-scale dataset to train a ConvNet. Moreover, even if we have a large-scale dataset, training ConvNet from scratch is required expensive resource and time-consuming. These two obstacles can be solved by using transfer learning. Transfer learning is a method for transferring the knowledge from a source domain to new domain. There are two major Transfer learning cases. First one is ConvNet as fixed feature extractor, and the second one is Fine-tune the ConvNet on a new dataset. In the first case, using pre-trained ConvNet (such as on ImageNet) to compute feed-forward activations of the image into the ConvNet and extract activation features from specific layers. In the second case, replacing and retraining the ConvNet classifier on the new dataset, then fine-tune the weights of the pre-trained network with the backpropagation. In this paper, we focus on using multiple ConvNet layers as a fixed feature extractor only. However, applying features with high dimensional complexity that is directly extracted from multiple ConvNet layers is still a challenging problem. We observe that features extracted from multiple ConvNet layers address the different characteristics of the image which means better representation could be obtained by finding the optimal combination of multiple ConvNet layers. Based on that observation, we propose to employ multiple ConvNet layer representations for transfer learning instead of a single ConvNet layer representation. Overall, our primary pipeline has three steps. Firstly, images from target task are given as input to ConvNet, then that image will be feed-forwarded into pre-trained AlexNet, and the activation features from three fully connected convolutional layers are extracted. Secondly, activation features of three ConvNet layers are concatenated to obtain multiple ConvNet layers representation because it will gain more information about an image. When three fully connected layer features concatenated, the occurring image representation would have 9192 (4096+4096+1000) dimension features. However, features extracted from multiple ConvNet layers are redundant and noisy since they are extracted from the same ConvNet. Thus, a third step, we will use Principal Component Analysis (PCA) to select salient features before the training phase. When salient features are obtained, the classifier can classify image more accurately, and the performance of transfer learning can be improved. To evaluate proposed method, experiments are conducted in three standard datasets (Caltech-256, VOC07, and SUN397) to compare multiple ConvNet layer representations against single ConvNet layer representation by using PCA for feature selection and dimension reduction. Our experiments demonstrated the importance of feature selection for multiple ConvNet layer representation. Moreover, our proposed approach achieved 75.6% accuracy compared to 73.9% accuracy achieved by FC7 layer on the Caltech-256 dataset, 73.1% accuracy compared to 69.2% accuracy achieved by FC8 layer on the VOC07 dataset, 52.2% accuracy compared to 48.7% accuracy achieved by FC7 layer on the SUN397 dataset. We also showed that our proposed approach achieved superior performance, 2.8%, 2.1% and 3.1% accuracy improvement on Caltech-256, VOC07, and SUN397 dataset respectively compare to existing work.

• Journal of Intelligence and Information Systems
• /
• v.26 no.1
• /
• pp.135-149
• /
• 2020

Artificial intelligences are changing world. Financial market is also not an exception. Robo-Advisor is actively being developed, making up the weakness of traditional asset allocation methods and replacing the parts that are difficult for the traditional methods. It makes automated investment decisions with artificial intelligence algorithms and is used with various asset allocation models such as mean-variance model, Black-Litterman model and risk parity model. Risk parity model is a typical risk-based asset allocation model which is focused on the volatility of assets. It avoids investment risk structurally. So it has stability in the management of large size fund and it has been widely used in financial field. XGBoost model is a parallel tree-boosting method. It is an optimized gradient boosting model designed to be highly efficient and flexible. It not only makes billions of examples in limited memory environments but is also very fast to learn compared to traditional boosting methods. It is frequently used in various fields of data analysis and has a lot of advantages. So in this study, we propose a new asset allocation model that combines risk parity model and XGBoost machine learning model. This model uses XGBoost to predict the risk of assets and applies the predictive risk to the process of covariance estimation. There are estimated errors between the estimation period and the actual investment period because the optimized asset allocation model estimates the proportion of investments based on historical data. these estimated errors adversely affect the optimized portfolio performance. This study aims to improve the stability and portfolio performance of the model by predicting the volatility of the next investment period and reducing estimated errors of optimized asset allocation model. As a result, it narrows the gap between theory and practice and proposes a more advanced asset allocation model. In this study, we used the Korean stock market price data for a total of 17 years from 2003 to 2019 for the empirical test of the suggested model. The data sets are specifically composed of energy, finance, IT, industrial, material, telecommunication, utility, consumer, health care and staple sectors. We accumulated the value of prediction using moving-window method by 1,000 in-sample and 20 out-of-sample, so we produced a total of 154 rebalancing back-testing results. We analyzed portfolio performance in terms of cumulative rate of return and got a lot of sample data because of long period results. Comparing with traditional risk parity model, this experiment recorded improvements in both cumulative yield and reduction of estimated errors. The total cumulative return is 45.748%, about 5% higher than that of risk parity model and also the estimated errors are reduced in 9 out of 10 industry sectors. The reduction of estimated errors increases stability of the model and makes it easy to apply in practical investment. The results of the experiment showed improvement of portfolio performance by reducing the estimated errors of the optimized asset allocation model. Many financial models and asset allocation models are limited in practical investment because of the most fundamental question of whether the past characteristics of assets will continue into the future in the changing financial market. However, this study not only takes advantage of traditional asset allocation models, but also supplements the limitations of traditional methods and increases stability by predicting the risks of assets with the latest algorithm. There are various studies on parametric estimation methods to reduce the estimated errors in the portfolio optimization. We also suggested a new method to reduce estimated errors in optimized asset allocation model using machine learning. So this study is meaningful in that it proposes an advanced artificial intelligence asset allocation model for the fast-developing financial markets.