• Korean Management Science Review
• /
• v.18 no.2
• /
• pp.25-38
• /
• 2001

The well-known back-propagation algorithm for multi-layered neural network has successfully been applied to pattern c1assification problems with remarkable flexibility. Recently. the multi-layered neural network is used as a powerful data mining tool. Nevertheless, in many cases with complex boundary of classification, the successful learning is not guaranteed and the problems of long learning time and local minimum attraction restrict the field application. In this paper, an Improved learning procedure of multi-layered neural network is proposed. The procedure is based on the generalized delta rule but it is particular in the point that the architecture of network is not fixed but enlarged during learning. That is, the number of hidden nodes or hidden layers are increased to help finding the classification boundary and such procedure is controlled by entropy evaluation. The learning speed and the pattern classification performance are analyzed and compared with the back-propagation algorithm.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2009.10a
• /
• pp.705-708
• /
• 2009

This research describes a neural network-based phoneme segmenter for recognizing spontaneous speech. The input of the phoneme segmenter for spontaneous speech is 16th order mel-scaled FFT, normalized frame energy, ratio of energy among 0~3[KHz] band and more than 3[KHz] band. All the features are differences of two consecutive 10 [msec] frame. The main body of the segmenter is single-hidden layer MLP(Multi-Layer Perceptron) with 72 inputs, 20 hidden nodes, and one output node. The segmentation accuracy is 78% with 7.8% insertion.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.43 no.10 s.352
• /
• pp.79-89
• /
• 2006

TCP is the most widely used transport protocol in Internet applications that guarantees a reliable data transfer. But, in the wireless multi-hop networks, TCP performance is degraded because it is designed for wired networks. The main reasons of TCP performance degradation are contention for wireless medium at the MAC layer, hidden terminal problem, exposed terminal problem, packet losses in the link layer, unfairness problem, reordering problem caused by path disconnection, bandwidth waste caused by exponential backoff of retransmission timer due to node's mobility and so on. Specially, in the mobile ad-hoc networks, discrepancy between a station's transmission range and interference range produces hidden terminal problem that decreases TCP performance greatly by limiting simultaneous transmission at a time. In this paper, we propose a new MAC algorithm for mobile ad-hoc networks to solve the problem that a node can not transmit and just increase CW by hidden terminal. In the IEEE 802.11 MAC DCF, a node increases CW exponentially when it fails to transmit, but the proposed algorithm, changes CW adaptively according to the reason of failure so we get a TCP performance enhancement. We show by ns-2 simulation that the proposed algorithm enhances the TCP performance by fairly distributing the transmission opportunity to the failed nodes by hidden terminal problems.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.6 no.9
• /
• pp.2285-2301
• /
• 2012

Peer-to-Peer (P2P) networks are becoming more and more popular in video content delivery services, such as Video on Demand (VoD). Scalability feature of P2P allows a higher number of simultaneous users at a given server load and bandwidth to use stream service. However, the quality of service (QoS) in these networks is difficult to be guaranteed because of the free-riding problem that nodes download the recourses while never uploading recourses, which degrades the performance of P2P VoD networks. In this paper, a distributed trust model is designed to reduce node's free-riding phenomenon in P2P VoD networks. In this model, the P2P network is abstracted to be a super node hierarchical structure to monitor the reputation of nodes. In order to calculate the reputation of nodes, the Hidden Markov Model (HMM) is introduced in this paper. Besides, a distinction algorithm is proposed to distinguish the free-riders and malicious nodes. The free-riders are the nodes which have a low frequency to free-ride. And the malicious nodes have a high frequency to free-ride. The distinction algorithm takes different measures to response to the request of these two kinds of free-riders. The simulation results demonstrate that this proposed trust model can improve QoS effectively in P2P VoD networks.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.7
• /
• pp.1749-1758
• /
• 1997

This paper suggests the development of dynamic forecasting model for short-term power demand based on Radial Basis Function Network and Pal's GLVQ algorithm. Radial Basis Function methods are often compared with the backpropagation training, feed-forward network, which is the most widely used neural network paradigm. The Radial Basis Function Network is a single hidden layer feed-forward neural network. Each node of the hidden layer has a parameter vector called center. This center is determined by clustering algorithm. Theatments of classical approached to clustering methods include theories by Hartigan(K-means algorithm), Kohonen(Self Organized Feature Maps %3A SOFM and Learning Vector Quantization %3A LVQ model), Carpenter and Grossberg(ART-2 model). In this model, the first approach organizes the load pattern into two clusters by Pal's GLVQ clustering algorithm. The reason of using GLVQ algorithm in this model is that GLVQ algorithm can classify the patterns better than other algorithms. And the second approach forecasts hourly load patterns by radial basis function network which has been constructed two hidden nodes. These nodes are determined from the cluster centers of the GLVQ in first step. This model was applied to forecast the hourly loads on Mar. $4^{th},\;Jun.\;4^{th},\;Jul.\;4^{th},\;Sep.\;4^{th},\;Nov.\;4^{th},$ 1995, after having trained the data for the days from Mar. $1^{th}\;to\;3^{th},\;from\;Jun.\;1^{th}\;to\;3^{th},\;from\;Jul.\;1^{th}\;to\;3^{th},\;from\;Sep.\;1^{th}\;to\;3^{th},\;and\;from\;Nov.\;1^{th}\;to\;3^{th},$ 1995, respectively. In the experiments, the average absolute errors of one-hour ahead forecasts on utility actual data are shown to be 1.3795%.

• Korean Management Science Review
• /
• v.15 no.2
• /
• pp.229-237
• /
• 1998

This paper presents a new algorithm for the K Shortest Paths Problem which is developed with a Double Shortest Arborescence and an inward arc breaking method. A Double Shortest Arborescence is made from merging a forward shortest arborescence and a backward one with Dijkstra algorithm. and shows us information about each shorter path to traverse each arc. Then K shorter paths are selected in ascending order of the length of each short path to traverse each arc, and some paths of the K shorter paths need to be replaced with some hidden shorter paths in order to get the optimal paths. And if the cross nodes which have more than 2 inward arcs are found at least three times in K shorter path, the first inward arc of the shorter than the Kth shorter path, the exposed path replaces the Kth shorter path. This procedure is repeated until cross nodes are not found in K shorter paths, and then the K shortest paths problem is solved exactly. This algorithm are computed with complexity o($n^3$) and especially O($n^2$) in the case K=3.

• Proceedings of the KIEE Conference
• /
• 1992.07a
• /
• pp.484-487
• /
• 1992

To resolve the knowledge acquisition bottleneck in the expert systems, the connectionist expert systems have been proposed, which facilitate learning capability of neural networks. This paper is to modify Gallant's connectionist expert network so that it can be applied to more general problems : 1) The hidden nodes are added between the input nodes and an output node, so that the back propagation learning algorithm is used instead of perception based Pocket algorithm. 2) Inference engine is thus modified by modeling that a node may have uncertainties due to unknown inputs.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.28B no.9
• /
• pp.739-746
• /
• 1991

This paper develops a modified multilayer perceptron architecture which speeds up learning as well as the net's mapping accuracy. In Phase I, a cluster partitioning algorithm like the Kohonen's self-organizing feature map or the leader clustering algorithm is used as the front end that determines the cluster to which the input data belongs. In Phase II, this cluster selects a subset of the hidden layer nodes that combines the input and outputs nodes into a subnet of the full scale backpropagation network. The proposed net has been applied to two mapping problems, one rather smooth and the other highly nonlinear. Namely, the inverse kinematic problem for a 3-link robot manipulator and the 5-bit parity mapping have been chosen as examples. The results demonstrate the proposed net's superior accuracy and convergence properties over the original backpropagation network or its existing improvement techniques.

• Phonetics and Speech Sciences
• /
• v.9 no.2
• /
• pp.95-101
• /
• 2017

This paper proposes a new method to train Deep Neural Network (DNN)-based acoustic models for speech recognition of native and foreign speakers. The proposed method consists of determining multi-set state clusters with various acoustic properties, training a DNN-based acoustic model, and recognizing speech based on the model. In the proposed method, hidden nodes of DNN are shared, but output nodes are separated to accommodate different acoustic properties for native and foreign speech. In an English speech recognition task for speakers of Korean and English respectively, the proposed method is shown to slightly improve recognition accuracy compared to the conventional multi-condition training method.

• Proceedings of the IEEK Conference
• /
• 2001.06c
• /
• pp.15-168
• /
• 2001

In this paper, we propose growing algorithm of wavelet neural network. It is growing algorithm that adds hidden nodes using wavelet frame which approximately supports orthogonality in wavelet neural network based on wavelet theory. The result of this processing can be reduced global error and progresses performance efficiency of wavelet neural network. We apply the proposed algorithm to approximation problem and evaluate effectiveness of proposed algorithm.