Hierarchical Network Model Research Articles

A distributed networking system for multimedia Internet access service using ATM over ADSL

AbstractThis paper proposes a distributed networking system architecture for Internet‐access service provision using ATM over xDSL technology. We describe the hierarchical network model in deploying ADSL services across the ATM access networks, which can easily accommodate the explosive growth of ADSL subscribers in the future. In addition, this paper describes the distributed networking system and its capability to provide a systemic network management using the principal networking concepts of service ordering, addressing, routing, adaptation and switching. All of the networking system components with CORBA objects in favor of the distribution and location transparency are defined and described using the CORBA interface description language (IDL) for commonality. Lastly, we present its implementation and operation in Korea Telecom. Copyright © 2004 John Wiley & Sons, Ltd.

Nonlinear principal component analysis to preserve the order of principal components

Principal component analysis (PCA) is an effective method of linear dimensional reduction. Because of its simplicity in theory and implementation, it is often used for analyses in various disciplines. However, because of its linearity, PCA is not always suitable, and has redundancy in expressing data. To overcome this problem, some nonlinear PCA methods have been proposed. However, most of these methods have drawbacks, such that the number of principal components must be predetermined, and also the order of the generated principal components is not explicitly given. In this paper, we propose a nonlinear PCA algorithm that nonlinearly transforms data into principal components, and at the same time, preserving the order of the principal components, and we also propose a hierarchical neural network model to perform the algorithm. Moreover, our method does not need to know the number of principal components in advance. The effectiveness of the proposed model will be shown through experiments.

Nonlinear principal component analysis to preserve the order of principal components

Principal component analysis (PCA) is an effective method of linear dimensional reduction. Because of its simplicity in theory and implementation, it is often used for analyses in various disciplines. However, because of its linearity, PCA is not always suitable, and has redundancy in expressing data. To overcome this problem, some nonlinear PCA methods have been proposed. However, most of these methods have drawbacks, such that the number of principal components must be predetermined, and also the order of the generated principal components is not explicitly given. In this paper, we propose a nonlinear PCA algorithm that nonlinearly transforms data into principal components, and at the same time, preserving the order of the principal components, and we also propose a hierarchical neural network model to perform the algorithm. Moreover, our method does not need to know the number of principal components in advance. The effectiveness of the proposed model will be shown through experiments.

Analytic network process (ANP) approach for product mix planning in semiconductor fabricator

This paper proposes an application of the analytic network process (ANP) for the selection of product mix for efficient manufacturing in a semiconductor fabricator. In order to evaluate different product mixes, a hierarchical network model based on various factors and the interactions of factors is presented. By incorporating experts’ opinion, a priority index can be calculated for each product mix studied, and a performance ranking of product mixes can be generated. The results provide guidance to a fab regarding strategies for accepting orders to maximize the manufacturing efficiency in considering the aspects of product, equipment efficiency and finance. The model can be easily understood and followed by administrators to determine the most efficient product mix for a fab.

Performance Analysis for Hierarchical Multirate Loss Networks

The reduced load approximation technique has been extensively applied to flat networks, but the feasibility of applying it to hierarchical network model has seldom been described. Hierarchical routing is essential for large networks such as the Internet inter/intra-domain routing hierarchy and the Private Network to Node Interface (PNNI) standard. Therefore, this paper proposes an efficient and accurate analytical model for evaluating the performance of hierarchical networks with multiple classes of traffic. A performance analysis model with considering multiple classes of traffic, the complexity of analytical and explosion of computation will be extremely increased, and hence, result in inaccurate analytical. The issue of multiple classes of traffic has to be addressed in performance analysis model. In this paper, we first study the reduced load approximation model for loss networks, and then propose a novel performance evaluation model for large networks with multirate hierarchical routing. The hierarchical evaluation model is based on decomposing a hierarchical route into several analytic hierarchical segments. Once the blockings of these hierarchical segments are accurately determined, the blocking of the hierarchical path can be estimated accurately from these segments blocking. Numerical results indicate that the proposed hierarchical reduced load approximation yields quite accurate blocking probabilities as compared to that of simulation results. Furthermore, the accuracy of the proposed hierarchical reduced load approximation heuristic is independent of the blocking or the offered traffic load. Finally, we also draw some remarks on the convergence of the reduced load based approximation analysis model.

Hierarchical Bayesian Neural Network for Gene Expression Temporal Patterns

There are several important issues to be addressed for gene expression temporal patterns' analysis: first, the correlation structure of multidimensional temporal data; second, the numerous sources of variations with existing high level noise; and last, gene expression mostly involves heterogeneous multiple dynamic patterns. We propose a Hierarchical Bayesian Neural Network model to account for the input correlations of time course gene array data. The variations in absolute gene expression levels and the noise can be estimated with the hierarchical Bayesian setting. The network parameters and the hyperparameters were simultaneously optimized with Monte Carlo Markov Chain simulation. Results show that the proposed model and algorithm can well capture the dynamic feature of gene expression temporal patterns despite the high noise levels, the highly correlated inputs, the overwhelming interactions, and other complex features typically present in microarray data. We test and demonstrate the proposed models with yeast cell cycle temporal data sets. The model performance of Hierarchical Bayesian Neural Network was compared to other popular machine learning methods such as Nearest Neighbor, Support Vector Machine, and Self Organized Map.

A hierarchical self-organizing approach for learning the patterns of motion trajectories.

The understanding and description of object behaviors is a hot topic in computer vision. Trajectory analysis is one of the basic problems in behavior understanding, and the learning of trajectory patterns that can be used to detect anomalies and predict object trajectories is an interesting and important problem in trajectory analysis. In this paper, we present a hierarchical self-organizing neural network model and its application to the learning of trajectory distribution patterns for event recognition. The distribution patterns of trajectories are learnt using a hierarchical self-organizing neural network. Using the learned patterns, we consider anomaly detection as well as object behavior prediction. Compared with the existing neural network structures that are used to learn patterns of trajectories, our network structure has smaller scale and faster learning speed, and is thus more effective. Experimental results using two different sets of data demonstrate the accuracy and speed of our hierarchical self-organizing neural network in learning the distribution patterns of object trajectories.

Exact scaling properties of a hierarchical network model.

We report on the exact results for the degree K, the diameter D, the clustering coefficient C, and the betweenness centrality B of a hierarchical network model with a replication factor M. Such quantities are calculated exactly with the help of recursion relations. Using the results, we show that (i). the degree distribution follows a power law P(K) approximately K-gamma with gamma=1+ln M/ln(M-1), (ii). the diameter grows logarithmically as D approximately ln N with the number of nodes N, (iii). the clustering coefficient of each node is inversely proportional to its degree, C proportional, variant 1/K, and the average clustering coefficient is nonzero in the infinite N limit, and (iv). the betweenness centrality distribution follows a power law P(B) approximately B-2. We discuss a classification scheme of scale-free networks into the universality class with the clustering property and the betweenness centrality distribution.

Invariant object recognition in the visual system with novel views of 3D objects.

To form view-invariant representations of objects, neurons in the inferior temporal cortex may associate together different views of an object, which tend to occur close together in time under natural viewing conditions. This can be achieved in neuronal network models of this process by using an associative learning rule with a short-term temporal memory trace. It is postulated that within a view, neurons learn representations that enable them to generalize within variations of that view. When three-dimensional (3D) objects are rotated within small angles (up to, e.g., 30 degrees), their surface features undergo geometric distortion due to the change of perspective. In this article, we show how trace learning could solve the problem of in-depth rotation-invariant object recognition by developing representations of the transforms that features undergo when they are on the surfaces of 3D objects. Moreover, we show that having learned how features on 3D objects transform geometrically as the object is rotated in depth, the network can correctly recognize novel 3D variations within a generic view of an object composed of a new combination of previously learned features. These results are demonstrated in simulations of a hierarchical network model (VisNet) of the visual system that show that it can develop representations useful for the recognition of 3D objects by forming perspective-invariant representations to allow generalization within a generic view.

A neural network model for encoding and perception of vowel sounds

By simulating a hierarchical neural network model, we investigated neuronal bases for encoding and perception of vowel sounds. The lower network detects spectral peaks called formant frequencies of a vowel sound. The higher network detects the combinatory information of the first (F1) and second (F2) formant frequencies. We trained the model with five Japanese vowels spoken by different people and modified synaptic connections according to the Hebbian rule. The present model can recognize not only learned vowel sounds but also vowel sounds that model hears for the first time. We suggest that an ‘unknown’ vowel sound can be perceived if they activate portion of a cell assembly whose population activation encodes information about the category of the unknown vowel sound.

Decomposition and integration of monosyllabic information for auditory perceptual process

A hierarchical neuronal network model is proposed for monosyllabic sound perception. The first stage decomposes monosyllabic sounds into time-varying spectral elements. The second stage integrates these spectral elements into point attractors corresponding to noise burst component, frequency-modulated (FM) component, and constant frequency component of a monosyllable. The third stage integrates these three components into a single point attractor corresponding to the monosyllable. We show that the perception of a monosyllabic sound is the emergence of one point attractor corresponding to the monosyllable in network dynamics at the third stage. The FM component is necessary for encoding monosyllables into relevant point attractors but not for monosyllabic sound perception.

Invariant recognition of feature combinations in the visual system.

The operation of a hierarchical competitive network model (VisNet) of invariance learning in the visual system is investigated to determine how this class of architecture can solve problems that require the spatial binding of features. First, we show that VisNet neurons can be trained to provide transform-invariant discriminative responses to stimuli which are composed of the same basic alphabet of features, where no single stimulus contains a unique feature not shared by any other stimulus. The investigation shows that the network can discriminate stimuli consisting of sets of features which are subsets or supersets of each other. Second, a key feature-binding issue we address is how invariant representations of low-order combinations of features in the early layers of the visual system are able to uniquely specify the correct spatial arrangement of features in the overall stimulus and ensure correct stimulus identification in the output layer. We show that output layer neurons can learn new stimuli if the lower layers are trained solely through exposure to simpler feature combinations from which the new stimuli are composed. Moreover, we show that after training on the low-order feature combinations which are common to many objects, this architecture can--after training with a whole stimulus in some locations--generalise correctly to the same stimulus when it is shown in a new location. We conclude that this type of hierarchical model can solve feature-binding problems to produce correct invariant identification of whole stimuli.

A sequential feature selection-based neural network approach to dynamic voltage stability estimation

Two sequential self-organizing hierarchical neural network (SHNN) models are utilized to estimate power system voltage stability. The first SHNN determines whether the power system is dynamically stable or not. The second one is used for the dynamically stable system to estimate the voltage magnitudes at all PQ buses. Tests results at different loading conditions for two test systems, WSCC nine-bus system and New England 39-bus system, are reported.

Localization of the Hall resistivity at high magnetic fields: absence of the quantized Hall insulator

We present a real-space renormalization-group study of localization in a two-dimensional hierarchical network model that is designed to represent phase-coherent electron transport in the quantum-Hall regime. Scaling expressions for both the longitudinal and Hall resistivities are derived. In agreement with recent numerical studies, we find that the Hall resistivity is quantized in the metallic phase but diverges in the insulating phase. This suggests that the characteristics of a quantized Hall insulator can occur only in the presence of a strong dephasing mechanism.

Quantum breakdown of the quantized Hall insulator

We present an analytical scaling theory for localization in a two-dimensional hierarchical network model that is designed to represent phase-coherent electron transport in the quantum-Hall regime. Scaling expressions for both the longitudinal and Hall resistivities are derived. In agreement with recent numerical studies, we find that the Hall resistivity is quantized in the metallic phase but diverges in the insulating phase. This suggests that the characteristics of a quantized Hall insulator can occur only in the presence of a strong dephasing mechanism.

Multilevel mesh partitioning for heterogeneous communication networks

Multilevel algorithms are a successful class of optimisation techniques which address the mesh partitioning problem for distributing unstructured meshes onto parallel computers. They usually combine a graph contraction algorithm together with a local optimisation method which refines the partition at each graph level. To date these algorithms have been used almost exclusively to minimise the cut edge weight in the graph with the aim of minimising the parallel communication overhead, but recently there has been a perceived need to take into account the communications network of the parallel machine. For example the increasing use of SMP clusters (systems of multiprocessor compute nodes with very fast intra-node communications but relatively slow inter-node networks) suggest the use of hierarchical network models. Indeed this requirement is exacerbated in the early experiments with meta-computers (multiple supercomputers combined together, in extreme cases over inter-continental networks). In this paper therefore, we modify a multilevel algorithm in order to minimise a cost function based on a model of the communications network. Several network models and variants of the algorithm are tested and we establish that it is possible to successfully guide the optimisation to reflect the chosen architecture.

Invariant object recognition in the visual system with error correction and temporal difference learning

It has been proposed that invariant pattern recognition might be implemented using a learning rule that utilizes a trace of previous neural activity which, given the spatio-temporal continuity of the statistics of sensory input, is likely to be about the same object though with differing transforms in the short time scale. Recently, it has been demonstrated that a modified Hebbian rule which incorporates a trace of previous activity but no contribution from the current activity can offer substantially improved performance. In this paper we show how this rule can be related to error correction rules, and explore a number of error correction rules that can be applied to and can produce good invariant pattern recognition. An explicit relationship to temporal difference learning is then demonstrated, and from this further learning rules related to temporal difference learning are developed. This relationship to temporal difference learning allows us to begin to exploit established analyses of temporal difference learning to provide a theoretical framework for better understanding the operation and convergence properties of these learning rules, and more generally, of rules useful for learning invariant representations. The efficacy of these different rules for invariant object recognition is compared using VisNet, a hierarchical competitive network model of the operation of the visual system.

Invariant object recognition in the visual system with error correction and temporal difference learning

It has been proposed that invariant pattern recognition might be implemented using a learning rule that utilizes a trace of previous neural activity which, given the spatio-temporal continuity of the statistics of sensory input, is likely to be about the same object though with differing transforms in the short time scale. Recently, it has been demonstrated that a modified Hebbian rule which incorporates a trace of previous activity but no contribution from the current activity can offer substantially improved performance. In this paper we show how this rule can be related to error correction rules, and explore a number of error correction rules that can be applied to and can produce good invariant pattern recognition. An explicit relationship to temporal difference learning is then demonstrated, and from this further learning rules related to temporal difference learning are developed. This relationship to temporal difference learning allows us to begin to exploit established analyses of temporal difference learning to provide a theoretical framework for better understanding the operation and convergence properties of these learning rules, and more generally, of rules useful for learning invariant representations. The efficacy of these different rules for invariant object recognition is compared using VisNet, a hierarchical competitive network model of the operation of the visual system.

Hierarchical fuzzy neural network‐based serviceability evaluation

The ability to economically and effectively service products provides an avenue for extending the product’s useful life. Quantifying approaches are needed to assist in the serviceability evaluation. In this study, the serviceability characteristics of a product are categorized into three groups: disassembly, reassembly, and handling. Often, many of the serviceability characteristics of a product cannot be defined completely or properly using crisp design data. Moreover, the serviceability analysis is an imprecise science characterized by ill‐structured and subjective evaluation criteria. A formal methodology for representing and processing the design information of an artifact using a hierarchical fuzzy neural network (FNN) model is presented. First, three FNNs are used to compute the disassembly, reassembly, and handling indices. Second, the output of the three FNNs is fed into a separate FNN to compute the serviceability index. The designer can use the proposed model to rank alternate designs by computing the serviceability indices of each candidate design. The working of the proposed model is demonstrated by using two designs for a steam iron.

On multistage fuzzy neural network modeling

In the past couple of years, there has been increasing interest in the fusion of neural networks and fuzzy logic. Most of the existing fuzzy neural network (FNN) models have been proposed to implement different types of single-stage fuzzy reasoning mechanisms and inevitably they suffer from the dimensionality problem when dealing with complex real-world problems. To address the problem, FNN modeling based on multistage fuzzy reasoning (MSFR) is pursued here and two hierarchical network models, namely incremental type and aggregated type, are introduced. The new models called multistage FNN (MSFNN) model a hierarchical fuzzy rule set that allows the consequence of a rule passed to another as a fact through the intermediate variables. From the stipulated input-output data pairs, they can generate an appropriate fuzzy rule set through structure and parameter learning procedures proposed in this paper. In addition, we have particularly addressed the input selection problem of these two types of multistage network models and proposed two efficient methods for them. The effectiveness of the proposed MSFNN models in handling high-dimensional problems is demonstrated through various numerical simulations.