• The KIPS Transactions:PartB
• /
• v.13B no.5 s.108
• /
• pp.533-540
• /
• 2006

The semantic web environment promise semantic search of heterogeneous data from distributed web page. Semantic search would resuit in an overwhelming number of results for users is increased, therefore elevating the need for appropriate personalized ranking schemes. Culture Finder helps semantic web agents obtain personalized culture information. It extracts meta data for each web page(culture news, culture performance, culture exhibition), perform semantic search and compute result ranking point to base user profile. In order to work efficient, Culture Finder uses five major technique: Machine learning technique for generating user profile from user search behavior and meta data repository, an efficient semantic search system for semantic web agent, query analysis for representing query and query result, personalized ranking method to provide suitable search result to user, upper ontology for generating meta data. In this paper, we also present the structure used in the Culture Finder to support personalized search service.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.5 no.4
• /
• pp.327-332
• /
• 2005

We investigatea new fuzzy-neural networks-Hybrid Fuzzy set based polynomial Neural Networks (HFSPNN). These networks consist of genetically optimized multi-layer with two kinds of heterogeneous neurons thatare fuzzy set based polynomial neurons (FSPNs) and polynomial neurons (PNs). We have developed a comprehensive design methodology to determine the optimal structure of networks dynamically. The augmented genetically optimized HFSPNN (namely gHFSPNN) results in a structurally optimized structure and comes with a higher level of flexibility in comparison to the one we encounter in the conventional HFPNN. The GA-based design procedure being applied at each layer of gHFSPNN leads to the selection leads to the selection of preferred nodes (FSPNs or PNs) available within the HFSPNN. In the sequel, the structural optimization is realized via GAs, whereas the ensuing detailed parametric optimization is carried out in the setting of a standard least square method-based learning. The performance of the gHFSPNN is quantified through experimentation where we use a number of modeling benchmarks synthetic and experimental data already experimented with in fuzzy or neurofuzzy modeling.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.15 no.4
• /
• pp.197-204
• /
• 2020

The ImageNet dataset is a large scale dataset and contains various natural scene images. In this paper, we propose a convolutional neural network (CNN)-based weighted ensemble technique for the ImageNet classification task. First, in order to fuse several models, our technique uses weights for each model, unlike the existing average-based ensemble technique. Then we propose an algorithm that automatically finds the coefficients used in later ensemble process. Our algorithm sequentially selects the model with the best performance of the validation set, and then obtains a weight that improves performance when combined with existing selected models. We applied the proposed algorithm to a total of 13 heterogeneous models, and as a result, 5 models were selected. These selected models were combined with weights, and we achieved 3.297% Top-5 error rate on the ImageNet test dataset.

• Journal of Korea Multimedia Society
• /
• v.4 no.4
• /
• pp.297-305
• /
• 2001

Recently advances in high-speed networks and multimedia computer technologies allow new types of multimedia applications to manipulate large volumes of multimedia data. However, in the real time and/or the heterogeneous data transmissions, there are many difficulties such as network transmission delay, the implementation difficulties, and so on. To solve these problems, in this paper, we extend the method of the multimedia service design which is proposed by OMG. To do this, we suggest an efficient real time audio/video stream framework, called Smart Explorer, based un CORBA and JMF Java Media API. And we separate the transmission path of control data from that of media data and use RTP/RTCP protocol for efficient real time audio/video transmission. Also we show the appropriate implementation of the audio/video stream system based on our suggested framework Smart Explorer. In the future, we expect our audio/video stream system to be applied to the real time communication software such as broadcasting, distance learning, and video conferencing.

• Journal of KIISE
• /
• v.41 no.12
• /
• pp.1110-1116
• /
• 2014

Nowadays, social networking sites (SNSs), such as Twitter, LinkedIn, and Tumblr, are coming into the forefront, due to the growth in the number of users. While users voluntarily provide their information in SNSs, privacy leakages resulting from the use of SNSs is becoming a problem owing to the evolution of large data processing techniques and the raising awareness of privacy. In order to solve this problem, the studies on protecting privacy on SNSs, based on graph and machine learning, have been conducted. However, examples of privacy leakages resulting from the advent of a new SNS are consistently being uncovered. In this paper, we propose a technique enabling a user to detect privacy leakages beforehand in the case where the service provider or third-party application developer threatens the SNS user's privacy maliciously.

• Journal of KIISE:Computing Practices and Letters
• /
• v.16 no.10
• /
• pp.980-984
• /
• 2010

Exploring microRNA (miRNA) and mRNA regulatory interactions may give new insights into diverse biological phenomena. Recently, miRNAs have been discovered as important regulators that play a major role in various cellular processes. Therefore, it is essential to identify functional interactions between miRNAs and mRNAs for understanding the context- dependent activities of miRNAs in complex biological systems. While elucidating complex miRNA-mRNA interactions has been studied with experimental and computational approaches, it is still difficult to infer miRNA-mRNA regulatory modules. Here we present a novel method, termed layered hypernetworks (LHNs), for identifying functional miRNA-mRNA interactions from heterogeneous expression data. In experiments, we apply the LHN model to miRNA and mRNA expression profiles on multiple cancers. The proposed method identifies cancer-specific miRNA-mRNA interactions. We show the biological significance of the discovered miRNA- mRNA interactions.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.40 no.1
• /
• pp.5-19
• /
• 2015

Literatures on open innovation have two major limitations. First, either on a firm level or on an industry level did they analyze the open innovation issues. The results of a firm's innovation can be diffused through the whole network and the firm can learn back from the network knowledge. Prior literatures did not consider the feedback loop among firms and network in which the firms are involved. Second, most open innovation research had a static perspective on firm's innovation performance. Since the diffusion, spill-over and learning among network members are involved over time, the open innovation is intrinsically dynamic. From the dynamic perspective, we can appreciate the fundamental attributes of the open innovation network which involves diverse firms, research institutes, and universities. In order to overcome the limitations, we suggest a dynamic model for open innovation network. We build an agent-based model which consists of heterogeneous firms. The firms are connected through a scale-free network which is formed by preferential attachment. Through the diverse scenario of simulation, we collect massive data on the firm level and analyze them both on firm and industry level. From the analysis, we found that, on industry level, the overall performance of open innovation increases as the internal research capability, absorptive capacity, and learning curve coefficient increase. Noticeably, as the deprecation rate of knowledge increases, the variability of knowledge increases. From the firm level analysis, we found that the industry-level variables had a significant effect on the firm's innovation performance lasting through all the time, whereas the firm-level variables had only on the early phase of innovation.

• Database Research
• /
• v.34 no.3
• /
• pp.124-136
• /
• 2018

There have been active research activities to use neural networks to analyze OCT images and make medical decisions. One requirement for these approaches to be promising solutions is that the trained network must be generalized to new devices without a substantial loss of performance. In this paper, we use a deep convolutional neural network to distinguish AMD from normal patients. The network was trained using a data set generated from an OCT device. We observed a significant performance degradation when it was applied to a new data set obtained from a different OCT device. To overcome this performance degradation, we propose an image normalization method which performs segmentation of OCT images to identify the retina area and aligns images so that the retina region lies horizontally in the image. We experimentally evaluated the performance of the proposed method. The experiment confirmed a significant performance improvement of our approach.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.11
• /
• pp.4290-4309
• /
• 2020

In recent years, advancements in machine learning capabilities have allowed it to see widespread adoption for tasks such as object detection, image classification, and anomaly detection. However, despite their promise, a limitation lies in the fact that a network's performance quality is based on the data which it receives. A well-trained network will still have poor performance if the subsequent data supplied to it contains artifacts, out of focus regions, or other visual distortions. Under normal circumstances, images of the same scene captured from differing points of focus, angles, or modalities must be separately analysed by the network, despite possibly containing overlapping information such as in the case of images of the same scene captured from different angles, or irrelevant information such as images captured from infrared sensors which can capture thermal information well but not topographical details. This factor can potentially add significantly to the computational time and resources required to utilize the network without providing any additional benefit. In this study, we plan to explore using image fusion techniques to assemble multiple images of the same scene into a single image that retains the most salient key features of the individual source images while discarding overlapping or irrelevant data that does not provide any benefit to the network. Utilizing this image fusion step before inputting a dataset into the network, the number of images would be significantly reduced with the potential to improve the classification performance accuracy by enhancing images while discarding irrelevant and overlapping regions.

• Journal of Internet Computing and Services
• /
• v.25 no.4
• /
• pp.23-37
• /
• 2024

Digital twin is an M&S (Modeling and Simulation) technology designed to solve or optimize problems in the real world by replicating physical objects in the real world as virtual objects in the digital world and predicting phenomena that may occur in the future through simulation. Digital twins have been elaborately designed and utilized based on data collected to achieve specific purposes in large-scale environments such as cities and industrial facilities. In order to apply this digital twin technology to real life and expand it into user-customized service technology, practical but sensitive issues such as personal information protection and personalization of simulations must be resolved. To solve this problem, this paper proposes a federated learning-based accelerated client training method (FACTS) for personalized digital twins. The basic approach is to use a cluster-driven federated learning training procedure to protect personal information while simultaneously selecting a training model similar to the user and training it adaptively. As a result of experiments under various statistically heterogeneous conditions, FACTS was found to be superior to the existing FL method in terms of training speed and resource efficiency.