• The Journal of Korean Association of Computer Education
• /
• v.12 no.2
• /
• pp.77-86
• /
• 2009

Large-scale applications, that needs large-capacity R&D resources and realtime data transmission, have demanded more stable and high-performance network environment than current Internet environments. Recently, global R&D networks have focuses on utilizing Lambda networking technologies and resource reservation systems to be satisfied with various applications' requirements. In this paper, we modify the existing DRS (Dynamic Right-Sizing) model to reflect various advantages in terms of the stability and high-capacity of Lambda network. In addition, we suggest the design methodology of high-performance Lambda network, which can integrate NRPS (Network Resource Provisioning System) into our modified DRS model.

• The Transactions of the Korea Information Processing Society
• /
• v.3 no.3
• /
• pp.478-492
• /
• 1996

For parallel implementations of functional programs without concurrent primitives, the λ-calculus encodings have been introduced. A functional program may be trans for med into a process network using process calculiby the λ-calculus encoding and there sult of a program can be obtained by a deal of communication actions in it's process network. But the λ-calculus encodings cause too many communication actions even in constant expressions. This paper shows the encoding for a combinator program without concurrency primitives which can combine the graph reduction and the process-net reduction using computable processes,'chores'. A 'chore' may have graph reduction functions for primitive operations of constants for which local graph reduction may be possible, and be encoded from a 'G-reducible' subexpression which is obtained by an annotation and trans for mati-on for a combinator program, assuring that it does not include any combinator application. Also, we show that a process network with chores raises less commu-nication actions than one without chores.

• The Journal of the Korea Contents Association
• /
• v.6 no.3
• /
• pp.28-37
• /
• 2006

In the last decades, inter-networking technologies advanced more rapidly than any other field. Today, the Internet is one of the most important infrastructure to society as becoming an indispensible tool of people and companies. During mid-1990's, developed countries recognized the advanced network as a basic infrastructure for the future science and technology development. They developed national research and education networks for the development of future science and network technology. In this paper, we made a comprehensive review of global research and education network developments. We also made analysis of Korea's activities on advanced network comparing with those of developed nations, then suggested approaches to improve Korean advanced networks.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.41 no.12
• /
• pp.41-49
• /
• 2004

Increase of internet users and new types of applied traffics, have led to demand for more bandwidth for each application. Hence, the amount of internet traffic has risen sharply and it has demanded to use limited resources, such as wavelength and bandwidth, more effectively. These kind of needs can be satisfied with OXC(Optical cross-connects) based on GMPLS that carry out IP packet switching and wavelength switching at the same time and Provide very wide bandwidth. In RSVP-TE signaling of GMPLS studied by IETF. every lambda router in core network should be able to convert wavelength. So, lots of wavelength converters and needed and building and managing cost is high. Another problem is that optimized traffic is limited. In this paper We suggest strengthened GMPLS RSVP-TE signaling algorithm for a better lightpath setup. When setup signaling is blocked suggested algorithm does not send PathErr message to Edge Router, but looks for nearest lambda router which can convert wavelength and carry out setup signaling from that node. Such algorithm can reduce the chance of blocked lightpath setup signaling and provide effective arrangement of lambda router in core network by calculating proper number of wavelength converter.

• KIPS Transactions on Software and Data Engineering
• /
• v.9 no.6
• /
• pp.177-186
• /
• 2020

Recent advances in deep learning technology have improved image recognition performance in the field of computer vision, and serverless computing is emerging as the next generation cloud computing technology for event-based cloud application development and services. Attempts to use deep learning and serverless computing technology to increase the number of real-world image recognition services are increasing. Therefore, this paper describes how to develop an efficient deep learning based image recognition service system using serverless computing technology. The proposed system suggests a method that can serve large neural network model to users at low cost by using AWS Lambda Server based on serverless computing. We also show that we can effectively build a serverless computing system that uses a large neural network model by addressing the shortcomings of AWS Lambda Server, cold start time and capacity limitation. Through experiments, we confirmed that the proposed system, using AWS Lambda Serverless Computing technology, is efficient for servicing large neural network models by solving processing time and capacity limitations as well as cost reduction.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.25 no.1
• /
• pp.27-35
• /
• 2021

This paper proposes a method to mitigate the Catastrophic Forgetting phenomenon in which artificial neural networks forget information on previous data. This method adjusts the Regularization strength by measuring the relationship between previous data and present data. MNIST and EMNIST data were used for performance evaluation and experimented in three scenarios. The experiment results showed a 0.1~3% improvement in the accuracy of the previous task for the same domain data and a 10~13% improvement in the accuracy of the previous task for different domain data. When continuously learning data with various domains, the accuracy of all previous tasks achieved more than 50% and the average accuracy improved by about 7%. This result shows that neural network learning can be properly performed in a CL environment in which data of different domains are successively entered by the method of this paper.

• Journal of Korea Port Economic Association
• /
• v.31 no.3
• /
• pp.155-166
• /
• 2015

This study selected container terminals of Gwangyang and Busan Ports to evaluate the influence of efficient container terminals. For the study, after data envelopment analysis (DEA) using the CCR and BCC models, the decision-making unit (DMU) system was used to define nodes; and with the use of a reference group in DEA (BCC model) and a lambda value, this study created a social network and analyzed the influences of efficient DMUs through a centrality analysis of eigenvectors. The results are presented as follows: First, as a result of the DEA, CCR efficiencies in PNC, HJNC, and HPNT container terminals of Busan Port were 1 and BCC efficiencies at Singamman Terminal, Wooam Terminal, PNC, HJNC, HPNT, and BNCT container terminals of Busan Port were 1. Second, as a result of undertaking social network analysis (SNA), according to an eigenvector centrality analysis, HJNC Terminal was referred to the most (influence score of 0.515), which indicates that it is the most influential as a container terminal. The influence of PNC Terminal was 0.512, while that of Wooam Terminal was 0.379. CJ Korea Express in Gwangyang Port was ranked fourth in influence, but its influence score of 0.256 indicates that it was the most influential of the container terminals at Gwangyang Port.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.1
• /
• pp.15-23
• /
• 2022

Among the learning methods for Continuous Learning environments, "Learning without Forgetting" has fixed regularization strengths, which can lead to poor performance in environments where various data are received. We suggest a way to set weights variable by identifying the features of the data we want to learn. We applied weights adaptively using correlation and complexity. Scenarios with various data are used for evaluation and experiments showed accuracy increases by up to 5% in the new task and up to 11% in the previous task. In addition, it was found that the adaptive weight value obtained by the algorithm proposed in this paper, approached the optimal weight value calculated manually by repeated experiments for each experimental scenario. The correlation coefficient value is 0.739, and overall average task accuracy increased. It can be seen that the method of this paper sets an appropriate lambda value every time a new task is learned, and derives the optimal result value in various scenarios.