• Journal of Ecology and Environment
• /
• v.47 no.4
• /
• pp.228-240
• /
• 2023

Growing complexity in ecosystem structure and functions, under impacts of climate and land-use changes, requires interdisciplinary understandings of processes and the whole-system, and accurate estimates of the changing functions. In the last three decades, observation networks for biodiversity, ecosystems, and ecosystem functions under climate change, have been developed by interested scientists, research institutions and universities. In this paper we will review (1) the development and on-going activities of those observation networks, (2) some outcomes from forest carbon cycle studies at our super-site "Takayama site" in Japan, and (3) a few ideas how we connect in-situ and satellite observations as well as fill observation gaps in the Asia-Oceania region. There have been many intensive research and networking efforts to promote investigations for ecosystem change and functions (e.g., Long-Term Ecological Research Network), measurements of greenhouse gas, heat, and water fluxes (flux network), and biodiversity from genetic to ecosystem level (Biodiversity Observation Network). Combining those in-situ field research data with modeling analysis and satellite remote sensing allows the research communities to up-scale spatially from local to global, and temporally from the past to future. These observation networks oftern use different methodologies and target different scientific disciplines. However growing needs for comprehensive observations to understand the response of biodiversity and ecosystem functions to climate and societal changes at local, national, regional, and global scales are providing opportunities and expectations to network these networks. Among the challenges to produce and share integrated knowledge on climate, ecosystem functions and biodiversity, filling scale-gaps in space and time among the phenomena is crucial. To showcase such efforts, interdisciplinary research at 'Takayama super-site' was reviewed by focusing on studies on forest carbon cycle and phenology. A key approach to respond to multidisciplinary questions is to integrate in-situ field research, ecosystem modeling, and satellite remote sensing by developing cross-scale methodologies at long-term observation field sites called "super-sites". The research approach at 'Takayama site' in Japan showcases this response to the needs of multidisciplinary questions and further development of terrestrial ecosystem research to address environmental change issues from local to national, regional and global scales.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.18 no.7
• /
• pp.907-912
• /
• 1993

In this Paper we have proposed a new method to implement the interpolation of the functions, using a neural network. The architecture of neural network is a three-layer perceptron and the training algorithm is a modified error back propagation algorithm adding neurons to hidden layer. The interpolated functions are sin(7 X), 3rd order polynomial 0.5$\times$3_2$\times$2+X+2.5 and rectangular pulse 0.99 U (X-0.2) -0.99 U(X-0.8) +0.01, where U(X) is the unit step. The root mean squred errors of the interpolated functions are 0.00258, 0.00164 and 0.00116 respectively.

• Journal of Korea Multimedia Society
• /
• v.21 no.10
• /
• pp.1142-1149
• /
• 2018

Recently, computer vision application is increasing by using CNN which is one of the deep learning algorithms. However, CNN does not provide perfect classification performance due to gradient vanishing problem. Most of CNN algorithms use an activation function called ReLU to mitigate the gradient vanishing problem. In this study, four activation functions that can replace ReLU were applied to four different structural networks. Experimental results show that ReLU has the lowest performance in accuracy, loss rate, and speed of initial learning convergence from 20 experiments. It is concluded that the optimal activation function varied from network to network but the four activation functions were higher than ReLU.

• ETRI Journal
• /
• v.44 no.2
• /
• pp.194-207
• /
• 2022

In the last decade, substantial progress has been achieved in intelligent traffic control technologies to overcome consistent difficulties of traffic congestion and its adverse effect on smart cities. Edge computing is one such advanced progress facilitating real-time data transmission among vehicles and roadside units to mitigate congestion. An edge computing-based deep reinforcement learning system is demonstrated in this study that appropriately designs a multiobjective reward function for optimizing different objectives. The system seeks to overcome the challenge of evaluating actions with a simple numerical reward. The selection of reward functions has a significant impact on agents' ability to acquire the ideal behavior for managing multiple traffic signals in a large-scale road network. To ascertain effective reward functions, the agent is trained withusing the proximal policy optimization method in several deep neural network models, including the state-of-the-art transformer network. The system is verified using both hypothetical scenarios and real-world traffic maps. The comprehensive simulation outcomes demonstrate the potency of the suggested reward functions.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.21 no.1
• /
• pp.251-268
• /
• 1996

In this paper, a higher-order feedforward neural network called ridge polynomial network (RPN) which shows good approximation capability for nonlnear continuous functions defined on compact subsets in multi-dimensional Euclidean spaces, is presented. This network provides more efficient and regular structure as compared to ordinary higher-order feedforward networks based on Gabor-Kolmogrov polynomial expansions, while maintating their fast learning property. the ridge polynomial network is a generalization of the pi-sigma network (PSN) and uses a specialform of ridge polynomials. It is shown that any multivariate polynomial can be exactly represented in this form, and thus realized by a RPN. The approximation capability of the RPNs for arbitrary continuous functions is shown by this representation theorem and the classical weierstrass polynomial approximation theorem. The RPN provides a natural mechanism for incremental function approximation based on learning algorithm of the PSN. Simulation results on several applications such as multivariate function approximation and pattern classification assert nonlinear approximation capability of the RPN.

• Knowledge Management Research
• /
• v.11 no.1
• /
• pp.51-64
• /
• 2010

The exhibition and convention industries, which are recognized as industrial hub linked with many relational industries. It is important to prevent a excessive competition for convention centers among districts and to gain some benefit fairly. Each convention center needs a lot of convention to held in it's district. In order to obtain many conventions to held in Korea for many convention center, convention bureau has to manage the processes to attract convention with many collaborators using the convention hub-network system. We should, therefore, pay attention to construction of the convention Hub-Network system to activate the convention industries in the district. We lay out the critical design for three functions of the convention Hub-Network system using UML. Finally we show the demonstrative web pages that explain critical three functions.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.12 no.6
• /
• pp.511-516
• /
• 2002

The neural networks may have problem such that the amount of calculation for the network learning goes too big according to the dimension of the dimension. To overcome this problem, the wavelet neural networks(WNN) which use the orthogonal basis function in the hidden node are proposed. One can compose wavelet functions as activation functions in the WNN by determining the scale and center of wavelet function. In this paper, when we compose the WNN using wavelet functions, we set a single scale function as a node function together. We intend that one scale function approximates the target function roughly, the other wavelet functions approximate it finely During the determination of the parameters, the wavelet functions can be determined by the global search for solutions suitable for the suggested problem using the genetic algorithm and finally, we use the back-propagation algorithm in the learning of the weights.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.13 no.7
• /
• pp.1444-1452
• /
• 2009

Probabilistic distribution functions based data generation method is very effective. Probabilistic distribution functions are defined under the assumption that daily routine contexts are mainly depended on a time-based schedule. However, daily life contexts are frequently determined by previous contexts because contexts have consistency and/or sequential flows. In order to refect previous contexts effect, a situation propagation network is proposed in this paper. As proposed situation propagation network make parameters of related probabilistic distribution functions update, generated contexts can be more realistic and natural. Through the simulation study, proposed context-data generation model generated general outworker's data about 11 daily contexts at home. Generated data are evaluated with respect to reduction of ambiguity and confliction using newly defined indexes of ambiguity and confliction of sequential contexts. In conclusion, in case of combining situation propagation network with probabilistic distribution functions, ambiguity and confliction of data can be reduced 6.45% and 4.60% respectively.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.8 no.12
• /
• pp.4604-4623
• /
• 2014

Authentication mechanisms coupled with strong encryption techniques are used for network security purposes; however, given sufficient time, well-equipped intruders are successful for compromising system security. The authentication protocols often fail when they are analysed critically. Formal approaches have emerged to analyse protocol failures. In this study, Communicating Sequential Processes (CSP) which is an abstract language designed especially for the description of communication patterns is employed. Rank functions are also used for verification and analysis which are helpful to establish that some critical information is not available to the intruder. In order to establish this, by assigning a value or rank to each critical information, it is shown that all the critical information that can be generated within the network have a particular characterizing property. This paper presents an application of rank functions approach to an authentication protocol that combines delaying the decryption process with timed authentication while keys are dynamically renewed under pseudo-secure situations. The analysis and verification of authentication properties and results are presented and discussed.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.13 no.5
• /
• pp.101-105
• /
• 2013

This paper addresses QoS functions in mobile backhaul network to accommodate the diverse traffics in cell site. The traffics assigned to the switching function in RAN system, such as Ethernet frame, IP packet, and ATM cell, are segmented, and then encapsulated to transfer then to the mobile backhaul network. ISP can converge all generation traffics, such as voice, HSPA, over all-IP RAN through standard pseudowire encapsulation. These can be enhanced with diverse QoS methods as well as comprehensive monitoring and diagnostic capabilities. Therefore in this paper, QoS functions under theses operations is simulated according to the encapsulation functions.