• Bulletin of the Korean Chemical Society
• /
• v.21 no.10
• /
• pp.1049-1051
• /
• 2000

A chromia-pillared clay has been prepared by ion exchange type intercalation reaction between the sodium ion in montmorillonite and the trimeric chromium oxyformate (TCF) ion, and by subsequent heat-treatment. The structural and thermal properties have been systematically studied by thermal analysis, powder XRD, IR spec-troscopy, and XAS. The gallery height of~6.8 $\AA$ upon intercalation of the TCF ion suggests that the $Cr_3O$ plane is parallel to the aluminosilicate layers. Even though the basal spacing of TCF intercalated clay decreases slightly upon heating, the layer structure was retained up to $550^{\circ}C$ as confirmed by XRD and TG/DTA. Ac-cording to the EXAFS spectroscopic analysis, it is identified that the (Cr-Cr) distance of 3.28 $\AA$ between vertex-linked CrO6 octahedra in TCF splits into 2.64 $\AA$, 2.98 $\AA$, and 3.77 $\AA$ due to the face-, edge-, and corner-shared CrO6 octahedra after heating at $400^{\circ}C$, implying that a nano-sized chromium oxide phase was stabilized within the interlayer space of clay.

• Pediatric Gastroenterology, Hepatology & Nutrition
• /
• v.13 no.1
• /
• pp.75-80
• /
• 2010

Intestinal duplication cysts are characterized by the attachment to some part of the gastrointestinal tract with which a blood supply is shared, and have an epithelial lining resembling some part of the alimentary tract. A 15-month-old female was admitted to our hospital with cyclic irritability, vomiting, and blood-tinged stool. The results of an ultrasound showed an ileocolic intussusception and a 1.3 cm cystic mass had double-wall sign and a Y-configuration with an adjacent ileal loop. She had a past history of two ileocolic intussusceptions. The cystic mass was considered to be a pathologic lead point, so resection and end-to-end anastomosis was performed. The gross and histologic evaluation of the specimen demonstrated a $2.4{\times}2.4$ cm cystic mass containing yellow mucoid fluid and the cyst wall was lined with intestinal and gastric mucosa and enclosed by a layer of muscle, which was shared with the adjacent ileum.

• Journal of Intelligence and Information Systems
• /
• v.22 no.2
• /
• pp.127-142
• /
• 2016

Deep learning model is a kind of neural networks that allows multiple hidden layers. There are various deep learning architectures such as convolutional neural networks, deep belief networks and recurrent neural networks. Those have been applied to fields like computer vision, automatic speech recognition, natural language processing, audio recognition and bioinformatics where they have been shown to produce state-of-the-art results on various tasks. Among those architectures, convolutional neural networks and recurrent neural networks are classified as the supervised learning model. And in recent years, those supervised learning models have gained more popularity than unsupervised learning models such as deep belief networks, because supervised learning models have shown fashionable applications in such fields mentioned above. Deep learning models can be trained with backpropagation algorithm. Backpropagation is an abbreviation for "backward propagation of errors" and a common method of training artificial neural networks used in conjunction with an optimization method such as gradient descent. The method calculates the gradient of an error function with respect to all the weights in the network. The gradient is fed to the optimization method which in turn uses it to update the weights, in an attempt to minimize the error function. Convolutional neural networks use a special architecture which is particularly well-adapted to classify images. Using this architecture makes convolutional networks fast to train. This, in turn, helps us train deep, muti-layer networks, which are very good at classifying images. These days, deep convolutional networks are used in most neural networks for image recognition. Convolutional neural networks use three basic ideas: local receptive fields, shared weights, and pooling. By local receptive fields, we mean that each neuron in the first(or any) hidden layer will be connected to a small region of the input(or previous layer's) neurons. Shared weights mean that we're going to use the same weights and bias for each of the local receptive field. This means that all the neurons in the hidden layer detect exactly the same feature, just at different locations in the input image. In addition to the convolutional layers just described, convolutional neural networks also contain pooling layers. Pooling layers are usually used immediately after convolutional layers. What the pooling layers do is to simplify the information in the output from the convolutional layer. Recent convolutional network architectures have 10 to 20 hidden layers and billions of connections between units. Training deep learning networks has taken weeks several years ago, but thanks to progress in GPU and algorithm enhancement, training time has reduced to several hours. Neural networks with time-varying behavior are known as recurrent neural networks or RNNs. A recurrent neural network is a class of artificial neural network where connections between units form a directed cycle. This creates an internal state of the network which allows it to exhibit dynamic temporal behavior. Unlike feedforward neural networks, RNNs can use their internal memory to process arbitrary sequences of inputs. Early RNN models turned out to be very difficult to train, harder even than deep feedforward networks. The reason is the unstable gradient problem such as vanishing gradient and exploding gradient. The gradient can get smaller and smaller as it is propagated back through layers. This makes learning in early layers extremely slow. The problem actually gets worse in RNNs, since gradients aren't just propagated backward through layers, they're propagated backward through time. If the network runs for a long time, that can make the gradient extremely unstable and hard to learn from. It has been possible to incorporate an idea known as long short-term memory units (LSTMs) into RNNs. LSTMs make it much easier to get good results when training RNNs, and many recent papers make use of LSTMs or related ideas.

• The Journal of the Korea Contents Association
• /
• v.9 no.12
• /
• pp.582-591
• /
• 2009

The pyramid graph is an interconnection network topology based on regular square mesh and tree structures. In this paper, we adopt a strategy of classification into two disjoint groups of edges in regular square mesh as a base sub-graph constituting of each layer in the pyramid graph. Edge set in the mesh can be divided into two disjoint sub-sets called as NPC(represents candidate edge for neighbor-parent) and SPC(represents candidate edge for shared-parent) whether the parents vertices adjacent to two end vertices of the corresponding edge have a relation of neighbor or shared in the upper layer of pyramid graph. In addition, we also introduce a notion of shrink graph to focus only on the NPC-edges by hiding SPC-edges in the original graph within the shrunk super-vertex on the resulting graph. In this paper, we analyze that the lower and upper bound on the number of NPC-edges in a Hamiltonian cycle constructed on $2^n\times2^n$ mesh is $2^{2n-2}$ and $3*(2^{2n-2}-2^{n-1})$ respectively. By expanding this result into the pyramid graph, we also prove that the maximum number of NPC-edges containable in a Hamiltonian cycle is $4^{n-1}-3*2^{n-1}$-2n+7 in the n-dimensional pyramid.

• Journal of IKEEE
• /
• v.22 no.1
• /
• pp.38-45
• /
• 2018

An integrated cryptographic processor that efficiently integrates ARIA, AES block ciphers and Whirlpool hash function into a single hardware architecture is described. Based on the algorithm characteristics of ARIA, AES, and Whirlpool, we optimized the design so that the hardware resources of the substitution layer and the diffusion layer were shared. The round block was designed to operate in a time-division manner for the round transformation and the round key expansion of the Whirlpool hash, resulting in a lightweight hardware implementation. The hardware operation of the integrated ARIA-AES-Whirlpool crypto-processor was verified by Virtex5 FPGA implementation, and it occupied 68,531 gate equivalents (GEs) with a 0.18um CMOS cell library. When operating at 80 MHz clock frequency, it was estimated that the throughputs of ARIA, AES block ciphers, and Whirlpool hash were 602~787 Mbps, 682~930 Mbps, and 512 Mbps, respectively.

• Journal of Conservation Science
• /
• v.26 no.3
• /
• pp.341-348
• /
• 2010

After excavation work of relics, historical site which has historically meaningful, are preserved using the various methods. There are three method in the relocation methods that are original relocation method, the remaining structure-peel method, and soil peel method. The original relocation method is restored after relocating in historical site such as residential site, iron foundry site, kiln site, old mound. The remaining structure-peel method are restored only the feature of exposed remaining structure using polymeric resin, when it is difficult to relocate the entire remaining structure. And soil-peel method is exhibited after peeling in case when soil layers such as grave of old mound, foundation of building site, sedimentary deposit layer, shell heap, and etc. Soil-peel method becomes important historical data of changes according to environment at that time, that is, flooding by storm, traces of fire and living features of that time such as heaps of shells discarded after eating shellfish and living wastes. In particular, in case of soil layer for preparing foundation sites of building by compacting in turn soils with different components such as clay soil, rough sand soil and etc, it becomes important data which can judge foundation technology of that time. It can be said to be an important data preservation method for utilizing these historical data as historical data as well as for the purpose of education, exhibitions and public relations which can be shared not only by experts but also by general public. In this paper, we present the reliable definition of soil-peel method in various preservation methods and explain the using polymer in this method. So, we will come up with the accurate index about this method that is used the eminent analysis method for soil layer.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.9 no.2
• /
• pp.268-276
• /
• 2005

This paper presents the motivation, design, and a preliminary evaluation of a virtual world builder, CLOVES. CLOVES is designed to support rapid construction of data-rich virtual environments and instruments for young children's science inquiry teaming. It provides a three-tiered programming Interface: a visual design environment, scripting layer, and low-level application programming interface targeting for multiple levels of programming expertise. It is also intended to be a collaborative medium among interdisciplinary domain experts such as educators, 3D modelers and software developers. A preliminary case study was conducted to evaluate the capabilities and effectiveness of CLOVES. The results showed that designers actively participated in decision making at every stage of the design process and shared knowledge among one another.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.22 no.10
• /
• pp.2282-2291
• /
• 1997

This paper describes an IC design for the HCR protocol which is designed to accomodata both real and non-real time data for high-speed multimedia services. The designed HCR IC is perfectly compatible with the conventional ATM IC's and physical layer IC's. The standard UTOPIA interface is adopted. Simulation results show that the proposed HCR IC operates very well according to the priority and the quota state of the real and non-real time data. The proposed HCR IC can be merged into the conventional ATM IC's without any modification and will be useful for multimedia service applications.

• Journal of Communications and Networks
• /
• v.12 no.2
• /
• pp.174-180
• /
• 2010

Automatic Repeat reQuest (ARQ) is a very effective technique against transmission error at the medium access control (MAC) layer. An erroneous MAC protocol data unit can be typically retransmitted within a given limit. In order to improve the IP-level performance, which directly affects the user-perceived quality-of-service (QoS), we propose a new truncated ARQ strategy, called MAC service data unit-based ARQ (S-ARQ), where the finite number of opportunities for retransmissions are shared by multiple fragments out of an IP datagram. We describe how SARQ can be implemented in a practical system, and then propose another variant of S-ARQ employing a functionality called early detection of failure. Basically, we evaluate the performance of SARQ in two different manners. First, assuming i.i.d. error process, we analyze both the probability of the delivery failure and the average delay of IP datagram. Then, we assess the performance of S-ARQ via simulation over a 2-state Markov channel.

• Journal of the Korea Society of Computer and Information
• /
• v.12 no.3
• /
• pp.19-26
• /
• 2007

In recent rears. web portal system has received much attention as a user interface for the grid environment. Grid system uses symmetric key for authenticating user identity while the traditional portal system does a password-based authentication. Regarding this, many researches are progressing to integrate portal accounts with symmetric key. Specially. researches such as GAMA and PURSE are active and those focus on easy usability for users who familiar with password-based authentication. However the protection of data and resources is a critical issue in Grid environment, because those are shared through a wide-area network. In this paper, we suggest a new authentication mechanism which unify authentication mechanisms between portal system and grid service by using symmetric key. It will improve a security level in UI layer as much as in grid service.