Search

Radial basis function (RBF) networks of Gaussian activation functions have been widely used in many applications due to its simplicity, robustness, good approximation and generalization ability, etc.. However, the training of such a RBF network is still a rather difficult task in the general case and the main crucial problem is how to select the number and locations of the hidden units appropriately. In this paper, we utilize a new kind of Bayesian Ying-Yang (BYY) automated model selection (AMS) learning algorithm to select the appropriate number and initial locations of the hidden units or Gaussians automatically for an input data set. It is demonstrated well by the experiments that this BYY-AMS training method is quite efficient and considerably outperforms the typical existing training methods on the training of RBF networks for both clustering analysis and nonlinear time series prediction. [ABSTRACT FROM AUTHOR]

In this paper we apply harmony data smoothing learning on a weighted kernel density model to obtain a sparse density estimator. We empirically compare this method with the least squares cross-validation (LSCV) method for the classical kernel density estimator. The most remarkable result of our study is that the harmony data smoothing learning method outperforms LSCV method in most cases and the support vectors selected by harmony data smoothing learning method are located in the regions of local highest density of the sample. [ABSTRACT FROM AUTHOR]

This paper attempts to evaluate machine learning based approaches to text categorization including NTC without decomposing it into binary classification problems, and presents another learning scheme of NTC. In previous research on text categorization, state of the art approaches have been evaluated in text categorization, decomposing it into binary classification problems. With such decomposition, it becomes complicated and expensive to implement text categorization systems, using machine learning algorithms. Another learning scheme of NTC mentioned in this paper is unconditional learning where weights of words stored in its learning layer are updated whenever each training example is presented, while its previous learning scheme is mistake driven learning, where weights of words are updated only when a training example is misclassified. This research will find advantages and disadvantages of both learning schemes by comparing them with each other [ABSTRACT FROM AUTHOR]

We usually identify the categories in image databases using some clustering algorithms based on the visual features extracted from images. Due to the well-known gap between the semantic features (e.g., categories) and the visual features, the results of unsupervised image categorization may be quite disappointing. Of course, it can be improved by adding some extra semantic information. Pairwise constraints between some images are easy to provide, even when we have little prior knowledge about the image categories in a database. A semi-supervised learning algorithm is then proposed for unsupervised image categorization based on Gaussian mixture model through incorporating such semantic information into the entropy-regularized likelihood (ERL) learning, which can automatically detect the number of image categories in the database. The experiments further show that this algorithm can lead to some promising results when applied to image categorization. [ABSTRACT FROM AUTHOR]

In the image registration process, there always exists rotation transformation. The ordinary methods such as NCC (Normalized Cross Correlation Algorithm), SD (Square Difference Algorithm), SSDA (Sequential Similarity Detection Algorithm), are not suitable for rotated image registration. In this paper, a method based on circular template, intensity distribution and SD is proposed for rotation image registration. Through the CPs (Control Points) got by the proposed method, transformation model and least square method, the rotation parameters are obtained. Experimental results verify its effectiveness. Compared with the existing feature-based approaches, it is easier to obtain CPs and needs no salient objects. [ABSTRACT FROM AUTHOR]

Automatic communication signal (e.g., FM signal) classification and identification focus on finding the fine feature contained in the almost approximate noisy communication signal comprehensively identifying the the same or different version of transmitters in modern electronic warfare. Direct use of HOS becomes unavailable for on-line application because of its huge computation time and memory space especially in the case of high frequency FM signal. This paper presents a novel view to improve the HOS analysis efficiency by sub-sampling while preserving the noise-contaminated fine feature and eliminating the random Gaussian noise. FM signal-related feature bispectra are also introduced to translate the 2-D feature matching pattern to a 1-D one applicable for an optimal adaptive k-means iterative RBF classifier. Computer simulations show that this novel feature bispectra outperform AIB and SB in terms of computation time and recognition rate for on-line steady FM signal recognition. [ABSTRACT FROM AUTHOR]

In this paper, we present two novel memetic algorithms (MAs) for gene selection. Both are synergies of Genetic Algorithm (wrapper methods) and local search methods (filter methods) under a memetic framework. In particular, the first MA is a Wrapper-Filter Feature Selection Algorithm (WFFSA) fine-tunes the population of genetic algorithm (GA) solutions by adding or deleting features based on univariate feature filter ranking method. The second MA approach, Markov Blanket-Embedded Genetic Algorithm (MBEGA), fine-tunes the population of solutions by adding relevant features, removing redundant and/or irrelevant features using Markov blanket. Our empirical studies on synthetic and real world microarray dataset suggest that both memetic approaches select more suitable gene subset than the basic GA and at the same time outperforms GA in terms of classification predictions. While the classification accuracies between WFFSA and MBEGA are not significantly statistically different on most of the datasets considered, MBEGA is observed to converge to more compact gene subsets than WFFSA. [ABSTRACT FROM AUTHOR]

This paper presents zero-crossing-based feature extraction for the speech recognition using neck-microphones. One of the solutions in noise-robust speech recognition is using neck-microphones which are not affected by the environmental noises. However, neck-microphones distort the original voice signals significantly since they only capture the vibrations of vocal tracts. In this context, we consider a new method of enhancing speech features of neck-microphone signals using zero-crossings. Furthermore, for the improvement of zero-crossing features, we consider to use the statistics of two adjacent zero-crossing intervals, that is, the statistics of two samples referred to as the second order statistics. Through the simulation for speech recognition using the neck-microphone voice command system, we have shown that the suggested method provides the better performance than other approaches using conventional speech features. [ABSTRACT FROM AUTHOR]

In this paper, we propose a probabilistic approach to feature selection for multi-class text categorization. Specifically, we regard document class and occurrence of each feature as events, calculate the probability of occurrence of each feature by the theorem on the total probability and utilize the values as a ranking criterion. Experiments on Reuters-2000 collection show that the proposed method can yield better performance than information gain and χ-square, which are two well-known feature selection methods. [ABSTRACT FROM AUTHOR]

Radar high-resolution range profile (HRRP) has received intensive attention from the radar automatic recognition (RATR) community. Since the initial phase of a complex HRRP is strongly sensitive to target position variation, which is referred to as the initial phase sensitivity, only the amplitude information in the complex HRRP, what is called the real HRRP, is used for RATR. This paper proposes a novel feature extraction method for the complex HRRP. The extracted complex feature vector contains the difference phase information between range cells but no initial phase information in the complex HRRP. The recognition algorithms, frame-template-database establishment methods and preprocessing methods used in the real HRRP-based RATR can also be applied to the proposed complex feature vector-based RATR. The recognition experiments based on measured data show that the proposed complex feature vector can obtain better recognition performance than the real HRRP if only the cell interval parameters are proper. [ABSTRACT FROM AUTHOR]

This paper presents a computational intelligence approach for predicting missing data in the presence of concept drift using an ensemble of multi-layered feed forward neural networks. An algorithm that detects concept drift by measuring heteroskedasticity is proposed. Six instances prior to the occurrence of missing data are used to approximate the missing values. The algorithm is applied to simulated time series data sets resembling non-stationary data from a sensor. Results show that the prediction of missing data in non-stationary time series data is possible but is still a challenge. For one test, up to 78% of the data could be predicted within 10% tolerance range of accuracy. [ABSTRACT FROM AUTHOR]

Network anomaly detection has been a hot topic in the past years. However, high false alarm rate, difficulties in obtaining exact clean data for the modeling of normal patterns and the deterioration of detection rate because of "unclean" training set always make it not as good as we expect. Therefore, we propose a novel data mining method for network anomaly detection in this paper. Experimental results on the well-known KDD Cup 1999 dataset demonstrate it can effectively detect anomalies with high true positives, low false positives as well as with high confidence than the state-of-the-art anomaly detection methods. Furthermore, even provided with not purely "clean" data (unclean data), the proposed method is still robust and effective. [ABSTRACT FROM AUTHOR]

With the vast amount of mobile user information available today, mining knowledge of mobile users is getting more and more important for a mobile commerce (M-commerce) system. Vector space model (VSM) is one of the most popular methods to achieve the above goal. Unfortunately, it can not identify the latent information in the user feature space, which decreases the quality of personalized services. In this paper, we present a concept-lattice based kernel method for mining the hidden user knowledge. The main idea is to employ concept lattice for constructing item proximity matrix, and then embed it into a kernel function, which transforms the original user feature space into a user concept space, and at last, perform personalized services in the user concept space. The experimental results demonstrate that our method is more encouraging than VSM. [ABSTRACT FROM AUTHOR]

Coronary heart disease (CHD) is a serious disease causing more and more morbidity and mortality. Combining western medicine and Traditional Chinese Medicine (TCM) to heal CHD becomes especially necessary for medical society today. Since western medicine faces some problems, like high cost and more side effects. TCM can be a complementary alternative to overcome these defects. Identification of what syndrome a CHD patient caught has been a challenging issue for medical society because the core of TCM is syndrome. In this paper, we carry out a large-scale clinical epidemiology to collect data with 1069 cases, each of which must be a CHD instance but may be diagnosed as different syndromes. We take blood stasis syndrome (frequency is 69%) as an example, employ four distinct kinds of data mining algorithms: Bayesian model; Neural Network; Support vector machine and Decision trees to classify the data and compare their performance. The results indicated that neural network is the best identifier with 88.6% accuracy on the holdout samples. The next is support vector machine with 82.5% accuracy, a slight higher than Bayesian model with 82.0% counterpart. The decision tree performs the worst, only 80.4%. We conclude that in identifying syndromes in CHD, neural network can provide a best insight to clinical application. [ABSTRACT FROM AUTHOR]

Data Mining is an analytic process designed to explore data (usually large amounts of data - typically business or market related) in search of consistent patterns and/or systematic relationships between variables, and then to validate the findings by applying the detected patterns to new subsets of data. One of the most commonly used techniques in data mining, Artificial Neural Networks provide non-linear predictive models that learn through training and resemble biological neural networks in structure. This paper deals with a new adaptive neural network model: a feed-forward higher order neural network with a new activation function called neuron-adaptive activation function. Experiments with function approximation and stock market movement analysis have been conducted to justify the new adaptive neural network model. Experimental results have revealed that the new adaptive neural network model presents several advantages over traditional neuron-fixed feed-forward networks such as much reduced network size, faster learning, and more promising financial analysis. [ABSTRACT FROM AUTHOR]

A method of moving horizon state estimation (MHSE) including a recurrent neural network as the dynamic model is used as an estimator of the filling level of the mill for a semiautogenous ore grinding process. The results are compared to those of a simple neural network acting as an estimator. They show the advantages of the Neural-MHSE, especially concerning robustness under large perturbations of the state variables (index of agreement > 0.9), which would favor its application to industrial scale processes. [ABSTRACT FROM AUTHOR]

A new weighted wavelet neural network is presented, And the approximation capability of such weighted wavelet neural network is also studied based on the traits of Lebesgue partition, the operator theory and the topology structure of the relatively compact set in Hilbert space. The simulation results indicate that the weighted wavelet neural network is a uniformed approximator, which can approximates the nonlinear function in compact set by arbitrary precision. [ABSTRACT FROM AUTHOR]

The problem of fault diagnosis of electrical machine has been an ongoing research in power systems. Many machine learning tools have been applied to this problem using static machine learning structures such as neural network, support vector machine that are unable to accommodate new information as it becomes available into their existing models. This paper presents a new method to bushing fault condition monitoring using fuzzy ARTMAP(FAM). FAM is introduced for bushing condition monitoring because it has the ability to incrementally learn information as it becomes available. An ensemble of classifiers is used to improve the classification accuracy of the systems. The testing results show that FAM ensemble gave an accuracy of 98.5%. Furthermore, the results show that fuzzy ARTMAP can update its knowledge in an incremental fashion without forgetting previously learned information. [ABSTRACT FROM AUTHOR]

This paper presents an informative gene set selection approach to tumor diagnosis based on the Distance Sensitive Rival Penalized Competitive Learning (DSRPCL) algorithm and redundancy analysis. Since the DSRPCL algorithm can allocate an appropriate number of clusters for an input dataset automatically, we can utilize it to classify the genes (expressed by the gene expression levels of all the samples) into certain basic clusters. Then, we apply the post-filtering algorithm to each basic gene cluster to get the typical and independent informative genes. In this way we can obtain a compact set of informative genes. To test the effectiveness of the selected informative gene set, we utilize the support vector machine (SVM) to construct a tumor diagnosis system based on the express profiles of its genes. It is shown by the experiments that the proposed method can achieve a higher diagnosis accuracy with a smaller number of informative genes and less computational complexity in comparison with the previous ones. [ABSTRACT FROM AUTHOR]

Code reuse in software reuse has several limitations such as difficulties of understanding and retrieval of the reuse code written by other developers. To overcome these problems, it should be possible to reuse the analysis/design information than source code itself. In this paper, I present analogical matching techniques for the reuse of object models and design patterns. We have suggested the design patterns as reusable components and the representation techniques to store them. The contents of the paper are as follows. 1) Analogical matching functions to retrieve analogous design patterns from reusable libraries. 2) The representation of reusable components to be stored in the library in order to support the analogical matching. [ABSTRACT FROM AUTHOR]

Compared to normal learning algorithms, for example backpropagation, the optimal bounded ellipsoid (OBE) algorithm has some better properties, such as faster convergence, since it has a similar structure as Kalman filter. OBE has some advantages over Kalman filter training, the noise is not required to be Guassian. In this paper OBE algorithm is applied traing the weights of recurrent neural networks for nonlinear system identification. Both hidden layers and output layers can be updated. From a dynamic systems point of view, such training can be useful for all neural network applications requiring real-time updating of the weights. A simple simulation gives the effectiveness of the suggested algorithm. [ABSTRACT FROM AUTHOR]

Multiple-Instance Learning is increasingly becoming one of the most promiscuous research areas in machine learning. In this paper, a new algorithm named NRBF-MI is proposed for Multi-Instance Learning based on normalized radial basis function network. This algorithm defined Compact Neighborhood of bags on which a new method is designed for training the network structure of NRBF-MI. The behavior of kernel function radius and its influence is analyzed. Furthermore a new kernel function is also defined for dealing with the labeled bags. Experimental results show that the NRBF-MI is a high efficient algorithm for Multi-Instance Learning. [ABSTRACT FROM AUTHOR]

In this paper, a novel elliptical basis function neural networks model (EBFNN) based on a hybrid learning algorithm (HLA) is proposed. Firstly, a geometry analytic algorithm is applied to construct the hyper-ellipsoid units of hidden layer of the EBFNN, i.e., initial the structure of the EBFNN. Then, the hybrid learning algorithm (HLA) is further applied to adjust the centers and the shape parameters. The experimental results demonstrated the proposed hybrid learning algorithm for the EBFNN model is feasible and efficient, and the EBFNN is not only parsimonious but also has better generalization performance than the RBFNN. [ABSTRACT FROM AUTHOR]

We extend the statistical neurodynamics to study transient dynamics of sequence processing neural networks with finite dilution, and the theoretical results are supported by extensive numerical simulations. It is found that the order parameter equations are completely equivalent to those of the Generating Functional Method, which means that crosstalk noise follows normal distribution even in the case of failure in retrieval process. In order to verify the gaussian assumption of crosstalk noise, we numerically obtain the cumulants of crosstalk noise, and third- and fourth-order cumulants are found to be indeed zero even in non-retrieval case. [ABSTRACT FROM AUTHOR]

This paper presents a new method of regularization in regression problems using a Besov norm (or semi-norm) acting as a regularization operator. This norm is more general smoothness measure to approximation spaces compared to other norms such as Sobolev and RKHS norms which are usually used in the conventional regularization methods. In our work, we also suggest a new candidate of the regularization parameter, that is, the trade-off between the data fit and the smoothness of the estimation function. Through the simulation for function approximation, we have shown that the suggested regularization method is effective and the estimated values of regularization parameters are close to the optimal values associated with the minimum expected risks. [ABSTRACT FROM AUTHOR]

According to analysis of both neuro-electrophysiological experimental data and the biophysical properties of neurons, in early research paper we proposed a new biophysical model that reflects the property of energy coding in neuronal activity. On the based of the above research work, in this paper the proposed biophysical model can reproduce the membrane potentials and the depolarizing membrane current by means of neuro-electrophysiological experimental data. Combination with our previous research results, the proposed biophysical model is demonstrated again to be more effective compared with known biophysical models of neurons. [ABSTRACT FROM AUTHOR]

This paper presents an efficient online sequential learning algorithm for extreme learning machine, which can learn data one by one. In this algorithm, the parameters of hidden nodes (the input weights and biases of additive nodes or the centers and impact factors of RBF nodes) are randomly selected and the output weights are analytically determined based on the sequentially arriving data. In the online sequence, the algorithm updates the output-layer weights with a Givens QR decomposition based on the orthogonalized least square algorithm. Simulations on benchmark problems demonstrate that the algorithm produces much better generalization performance than another online sequential extreme learning machine algorithm, or sometimes it has good performance than primitive extreme learning machine algorithm. [ABSTRACT FROM AUTHOR]

A robust tracking controller with bound estimation based on neural network is proposed to deal with the unknown factors of nonholonomic mobile robot, such as model uncertainties and external disturbances. The neural network is to approximate the uncertainties terms and the interconnection weights of the neural network can be tuned online. And the robust controller is designed to compensate for the approximation error. Moreover, an adaptive estimation algorithm is employed to estimate the bound of the approximation error. The stability of the proposed controller is proven by Lyapunov function. The proposed neural network-based robust tracking controller can overcome the uncertainties and the disturbances. The simulation results demonstrate that the proposed method has good robustness. [ABSTRACT FROM AUTHOR]

In the human mind, high-order knowledge is categorically organized, yet the nature of its internal representation system is not well understood. While it has been traditionally considered that there is a single innate representation system in our mind, recent studies suggest that the representational system is a dynamic, capable of adjusting a representation scheme to meet situational characteristics. In the present paper, we introduce a new cognitive modeling framework accounting for the flexibility in representing high-order category knowledge. Our modeling framework flexibly learns to adjust its internal knowledge representation scheme using a meta-heuristic optimization method. It also accounts for the multi-objective and the multi-notion natures of human learning, both of which are indicated as very important but often overlooked characteristics of human cognition. [ABSTRACT FROM AUTHOR]

Machine learning, a cornerstone of intelligent systems, has typically been studied in the context of specific tasks, including clustering (unsupervised learning), classification (supervised learning), and control (reinforcement learning). This paper presents a learning architecture within which a universal adaptation mechanism unifies a rich set of traditionally distinct learning paradigms, including learning by matching, learning by association, learning by instruction, and learning by reinforcement. In accordance with the notion of embodied intelligence, such a learning theory provides a computational account of how an autonomous agent may acquire the knowledge of its environment in a real-time, incremental, and continuous manner. Through a case study on a minefield navigation domain, we illustrate the efficacy of the proposed model, the learning paradigms encompassed, and the various types of knowledge learned. [ABSTRACT FROM AUTHOR]

Investigated in this paper are the uniform approximation capabilities of sum-of-product (SOPNN) and sigma-pi-sigma (SPSNN) neural networks. It is proved that the set of functions that are generated by an SOPNN with its activation function in C(ℝ) is dense in $C(\mathbb{K})$ for any compact $\mathbb{K}\in \mathbb{R}^N$, if and only if the activation function is not a polynomial. It is also shown that if the activation function of an SPSNN is in C(ℝ), then the functions generated by the SPSNN are dense in $C(\mathbb{K})$ if and only if the activation function is not a constant. [ABSTRACT FROM AUTHOR]

Multiple-instance learning (MIL) is a variation on supervised learning, where the task is to learn a concept given positive and negative bags of instances. In this paper a novel algorithm has been introduced for multiple-instance learning. This method was inspired by both diverse density (DD) and its expectation maximization version (EM-DD). It converts MIL problem to a single-instance setting. This improved method has better accuracy and time complexity than DD and EM-DD. We apply it to drug activity prediction and image retrieval. The experiments show it has competitive accuracy values compared with other previous approaches. [ABSTRACT FROM AUTHOR]

This paper proposes a novel artificial neural oscillator consisted of two neurons with excellent control properties. The mutual connections between the neurons are just linear functions and determine the oscillation angular frequency. And each neuron has a nonlinear self-feedback connection to hold up oscillation amplitude. The dynamics of the neural oscillator was modelled with nonlinear coupling functions. And the stability, amplitude, angular frequency of the oscillator are determined independently by three parameters of the functions. Since it has simple structure and favorable control advantages, it can be used in bionic robot's locomotion control system. The first application is an artificial central pattern generator (CPG) controller for bionic robot's joint. The second is a bionic neural network for fish-robot's locomotion control. [ABSTRACT FROM AUTHOR]

Online-sequential extreme learning machine (OS-ELM) shows a good solution to online learning using extreme learning machine approach for single-hidden-layer feedforward network. However, the algorithm tends to be data-dependent, i.e. the bias values need to be adjusted depending on each particular problem. In this paper, we propose an enhancement to OS-ELM, which is referred to as robust OS-ELM (ROS-ELM). ROS-ELM has a systematic method to select the bias that allows the bias to be selected following the input weights. Hence, the proposed algorithm works well for every benchmark dataset. ROS-ELM has all the pros of OS-ELM, i.e. the capable of learning one-by-one, chunk-by-chunk with fixed or varying chunk size. Moreover, the performance of the algorithm is higher than OS-ELM and it produces a better generalization performance with benchmark datasets. [ABSTRACT FROM AUTHOR]

Discrete-time analogues of continuous-time neural networks with continuously distributed delays and periodic inputs are introduced. The discrete-time analogues are considered to be numerical discretizations of the continuous-time networks and we study their dynamical characteristics. By employing Halanay-type inequality, we obtain easily verifiable sufficient conditions ensuring that every solutions of the discrete-time analogue converge exponentially to the unique periodic solutions. It is shown that the discrete-time analogues preserve the periodicity of the continuous-time networks. [ABSTRACT FROM AUTHOR]

Extreme learning machine (ELM) is a novel learning algorithm much faster than the traditional gradient-based learning algorithms for single-hidden-layer feedforward neural networks (SLFNs). Neural network ensemble is a learning paradigm where several neural networks are jointly used to solve a problem. In our work, we investigated the performance of ELMs ensemble on regression problems. A simple ensembling approach Product Index based Excluding ensemble(PIEx) was proposed to ensemble accurate and diverse member networks. The experimental results show that the ensemble can effectively improve the performance compared with the generalization ability of single ELM and PIEx outperforms Bagging and Simple Averaging. The results also show ELM training can generate diverse neural networks even though using the same training set. [ABSTRACT FROM AUTHOR]

In this paper, the global uniform asymptotical stability is studied for neural networks with multiple time-varying delays by constructing appropriate Lyapunov-Krasovskii functional and using the linear matrix inequality (LMI) approach. The restriction on the derivative of the time-varying delay function τij(t) to be less than unit is removed by using slack matrix method. A numerical example is provided to demonstrate the effectiveness and applicability of the proposed criteria. [ABSTRACT FROM AUTHOR]

In this paper, we investigate dynamical behavior of a general class of competitive or cooperative Lotka-Volterra systems with delays. Positive solutions and global stability of nonnegative equilibrium are discussed. Sufficient condition independent of delays guaranteeing existence of globally stable equilibrium is given. A Simulation verifying theoretical results is given, too. [ABSTRACT FROM AUTHOR]

We present an efficient stream authentication scheme that improves the verification probability of Park-Chung-Cho's scheme [1] by using Rabin's Information Dispersal Algorithm. It is shown that under the same communication overhead the verification probability of the proposed scheme is higher than those of SAIDA as well as Park-Chung-Cho's scheme, and that the execution time of the proposed scheme is smaller than that of SAIDA. [ABSTRACT FROM AUTHOR]

In this paper, the difference between using the inverted neuron output and negative resistor for expressing inhibitory synapse is studied. We analyzed that the total conductance seen at the neuron input is different in these two methods. And this total conductance has been proved to effect on the system stability in this paper. Also, we proposed the method how to stabilize the input space and improve the system's performance by adjusting the input conductance between neuron input and ground. Pspice is used for circuit level simulation. [ABSTRACT FROM AUTHOR]

In this paper, we study the Cohen-Grossberg neural networks with mixed time-varying delays. By applying the Lyapunov functional method and combining with the inequality 3abc ≤ a3 + b3 + c3 (a,b,c > 0) technique, a series of new and useful criteria on the existence of equilibrium point and its global asymptotical stability are established. The results obtained in this paper extend and generalize the corresponding results existing in previous literature. [ABSTRACT FROM AUTHOR]

This paper studies a general class of neural networks with time-varying delays and the neuron activations belong to the set of discontinuous monotone increasing functions. The discontinuities in the activations are an ideal model of the situation where the gain of the neuron amplifiers is very high. Because the delay in combination with high-gain nonlinearities is a particularly harmful source of potential instability, in the paper, conditions which ensure the global convergence of the neural network are derived. [ABSTRACT FROM AUTHOR]

In finite mixture modelling, it is crucial to select the number of components for a data set. We have proposed an entropy regularized likelihood (ERL) learning principle for the finite mixtures to solve this model selection problem under regularization theory. In this paper, we further give an asymptotic analysis of the ERL learning, and find that the global minimization of the ERL function in a simulated annealing way (i.e., the regularization factor is gradually reduced to zero) leads to automatic model selection on the finite mixtures with a good parameter estimation. As compared with the EM algorithm, the ERL learning can go across the local minima of the negative likelihood and keep robust with respect to initialization. The simulation experiments then prove our theoretic analysis. [ABSTRACT FROM AUTHOR]

The global asymptotic stability of discrete-time Cohen-Grossberg neural networks (CGNNs) with or without time delays is studied in this paper. The CGNNs are transformed into discrete-time interval systems, and several sufficient conditions of asymptotic stability for these interval systems are derived by constructing some suitable Lyapunov functionals. The obtained conditions are given in the form of linear matrix inequalities that can be checked numerically and very efficiently by resorting to the MATLAB LMI Control Toolbox. [ABSTRACT FROM AUTHOR]

In this paper, a class of generalized nonautonomous cellular neural networks with time-varying delays are studied. By means of Lyapunov functional method, improved Young inequality ambn ≤ mat− n + nbtm (0 ≤ m ≤ 1, m + n = 1,t > 0) and the homeomorphism theory, several sufficient conditions are given guaranteeing the existence, uniqueness and global exponential stability of the equilibrium point. The proposed results generalize and improve previous works. An illustrative example is also given to demonstrate the effectiveness of the proposed results. [ABSTRACT FROM AUTHOR]

In this paper, the global asymptotic and exponential stability are investigated for a class of neural networks with distributed time-varying delays. By using appropriate Lyapunov-Krasovskii functional and linear matrix inequality (LMI) technique, two delay-dependent sufficient conditions in LMIs form are obtained to guarantee the global asymptotic and exponential stability of the addressed neural networks. The proposed stability criteria do not require the monotonicity of the activation functions and the differentiability of the distributed time-varying delays, which means that the results generalize and further improve those in the earlier publications. An example is given to show the effectiveness of the obtained condition. [ABSTRACT FROM AUTHOR]

The existence and global asymptotic stability of a large class of Cohen-Grossberg neural networks is discussed in this paper. Previous papers always assume that the amplification function has positive lower and upper bounds, which excludes a large class of functions. In our paper, it is only needed that the amplification function is positive. Also, the model discussed is general, the method used is direct and the conditions needed are weak. [ABSTRACT FROM AUTHOR]

The paper investigates the existence and global stability of periodic solution of non-autonomous neural networks with delay. Then the existence and uniqueness of periodic solutions of the neural networks are discussed in the paper. Moreover, criterion on stability of periodic solutions of the neural networks is obtained by using matrix function inequality, and algorithm for the criterion on the neural networks is provided. Result in the paper generalizes and improves the result in the existing references. In the end, an illustrate example is given to verify our results. [ABSTRACT FROM AUTHOR]

In this paper, the periodic solution for a class of Cohen-Grossberg neural networks with variable coefficients is discussed. By using inequality analysis technique and matrix theory, some new sufficient conditions are obtained to ensure the existence, uniqueness, global attractivity and exponential stability of the periodic solution. An example is given to show the effectiveness of the obtained results. [ABSTRACT FROM AUTHOR]

In this paper, the global asymptotic stability of Hopfield neural networks with time delays is investigated. Some novel sufficient conditions are presented for the global stability of a given delayed Hopfield neural networks by constructing Lyapunov functional and using some well-known inequalities. A linear matrix inequality (LMI) approach is developed to establish sufficient conditions for the given neural networks. An illustrative example is provided to demonstrate the effectiveness of our theoretical results. [ABSTRACT FROM AUTHOR]