Neural network modeling of associative memory: Beyond the Hopfield model

A number of neural network models, in which fixed-point and limit-cycle attractors of the underlying dynamics are used to store and associatively recall information, are described. In the first class of models, a hierarchical structure is used to store an exponentially large number of strongly correlated memories. The second class of models uses limit cycles to store and retrieve individual memories. A neurobiologically plausible network that generates low-amplitude periodic variations of activity, similar to the oscillations observed in electroencephalographic recordings, is also described. Results obtained from analytic and numerical studies of the properties of these networks are discussed.

Modeling of magnetostrictive materials and structures

The constitutive model for a magnetostrictive material and its effect on the structural response is presented in this article. The example of magnetostrictive material considered is the TERFENOL-D. As like the piezoelectric material, this material has two constitutive laws, one of which is the sensing law and the other is the actuation law, both of which are highly coupled and non-linear. For the purpose of analysis, the constitutive laws can be characterized as coupled or uncoupled and linear or non linear. Coupled model is studied without assuming any explicit direct relationship with magnetic field. In the linear coupled model, which is assumed to preserve the magnetic flux line continuity, the elastic modulus, the permeability and magneto-elastic constant are assumed as constant. In the nonlinear-coupled model, the nonlinearity is decoupled and solved separately for the magnetic domain and the mechanical domain using two nonlinear curves, namely the stress vs. strain curve and the magnetic flux density vs. magnetic field curve. This is performed by two different methods. In the first, the magnetic flux density is computed iteratively, while in the second, the artificial neural network is used, where in the trained network will give the necessary strain and magnetic flux density for a given magnetic field and stress level. The effect of nonlinearity is demonstrated on a simple magnetostrictive rod.

Multiple contour extraction from graylevel images using an artificial neural network

For active contour modeling (ACM), we propose a novel self-organizing map (SOM)-based approach, called the batch-SOM (BSOM), that attempts to integrate the advantages of SOM- and snake-based ACMs in order to extract the desired contours from images. We employ feature points, in the form of ail edge-map (as obtained from a standard edge-detection operation), to guide the contour (as in the case of SOM-based ACMs) along with the gradient and intensity variations in a local region to ensure that the contour does not "leak" into the object boundary in case of faulty feature points (weak or broken edges). In contrast with the snake-based ACMs, however, we do not use an explicit energy functional (based on gradient or intensity) for controlling the contour movement. We extend the BSOM to handle extraction of contours of multiple objects, by splitting a single contour into as many subcontours as the objects in the image. The BSOM and its extended version are tested on synthetic binary and gray-level images with both single and multiple objects. We also demonstrate the efficacy of the BSOM on images of objects having both convex and nonconvex boundaries. The results demonstrate the superiority of the BSOM over others. Finally, we analyze the limitations of the BSOM.

Artificial neural networks and multicriterion analysis for sustainable irrigation planning

The objective of the present paper is to select the best compromise irrigation planning strategy for the case study of Jayakwadi irrigation project, Maharashtra, India. Four-phase methodology is employed. In phase 1, separate linear programming (LP) models are formulated for the three objectives, namely. net economic benefits, agricultural production and labour employment. In phase 2, nondominated (compromise) irrigation planning strategies are generated using the constraint method of multiobjective optimisation. In phase 3, Kohonen neural networks (KNN) based classification algorithm is employed to sort nondominated irrigation planning strategies into smaller groups. In phase 4, multicriterion analysis (MCA) technique, namely, Compromise Programming is applied to rank strategies obtained from phase 3. It is concluded that the above integrated methodology is effective for modeling multiobjective irrigation planning problems and the present approach can be extended to situations where number of irrigation planning strategies are even large in number. (c) 2004 Elsevier Ltd. All rights reserved.

Voltage and Temperature Aware Statistical Leakage Analysis Framework Using Artificial Neural Networks

Artificial neural networks (ANNs) have shown great promise in modeling circuit parameters for computer aided design applications. Leakage currents, which depend on process parameters, supply voltage and temperature can be modeled accurately with ANNs. However, the complex nature of the ANN model, with the standard sigmoidal activation functions, does not allow analytical expressions for its mean and variance. We propose the use of a new activation function that allows us to derive an analytical expression for the mean and a semi-analytical expression for the variance of the ANN-based leakage model. To the best of our knowledge this is the first result in this direction. Our neural network model also includes the voltage and temperature as input parameters, thereby enabling voltage and temperature aware statistical leakage analysis (SLA). All existing SLA frameworks are closely tied to the exponential polynomial leakage model and hence fail to work with sophisticated ANN models. In this paper, we also set up an SLA framework that can efficiently work with these ANN models. Results show that the cumulative distribution function of leakage current of ISCAS'85 circuits can be predicted accurately with the error in mean and standard deviation, compared to Monte Carlo-based simulations, being less than 1% and 2% respectively across a range of voltage and temperature values.

A Long Memory Process Based Parametric Modeling and Recognition of PD Signal

We address the problem of recognition and retrieval of relatively weak industrial signal such as Partial Discharges (PD) buried in excessive noise. The major bottleneck being the recognition and suppression of stochastic pulsive interference (PI) which has similar time-frequency characteristics as PD pulse. Therefore conventional frequency based DSP techniques are not useful in retrieving PD pulses. We employ statistical signal modeling based on combination of long-memory process and probabilistic principal component analysis (PPCA). An parametric analysis of the signal is exercised for extracting the features of desired pules. We incorporate a wavelet based bootstrap method for obtaining the noise training vectors from observed data. The procedure adopted in this work is completely different from the research work reported in the literature, which is generally based on deserved signal frequency and noise frequency.

A Comparison of Machine Learning Techniques for Modeling River Flow Time Series: The Case of Upper Cauvery River Basin

Models of river flow time series are essential in efficient management of a river basin. It helps policy makers in developing efficient water utilization strategies to maximize the utility of scarce water resource. Time series analysis has been used extensively for modeling river flow data. The use of machine learning techniques such as support-vector regression and neural network models is gaining increasing popularity. In this paper we compare the performance of these techniques by applying it to a long-term time-series data of the inflows into the Krishnaraja Sagar reservoir (KRS) from three tributaries of the river Cauvery. In this study flow data over a period of 30 years from three different observation points established in upper Cauvery river sub-basin is analyzed to estimate their contribution to KRS. Specifically, ANN model uses a multi-layer feed forward network trained with a back-propagation algorithm and support vector regression with epsilon intensive-loss function is used. Auto-regressive moving average models are also applied to the same data. The performance of different techniques is compared using performance metrics such as root mean squared error (RMSE), correlation, normalized root mean squared error (NRMSE) and Nash-Sutcliffe Efficiency (NSE).

Modeling of Channel Potential and Subthreshold Slope of Symmetric Double-Gate Transistor

A new physically based classical continuous potential distribution model, particularly considering the channel center, is proposed for a short-channel undoped body symmetrical double-gate transistor. It involves a novel technique for solving the 2-D nonlinear Poisson's equation in a rectangular coordinate system, which makes the model valid from weak to strong inversion regimes and from the channel center to the surface. We demonstrated, using the proposed model, that the channel potential versus gate voltage characteristics for the devices having equal channel lengths but different thicknesses pass through a single common point (termed ``crossover point''). Based on the potential model, a new compact model for the subthreshold swing is formulated. It is shown that for the devices having very high short-channel effects (SCE), the effective subthreshold slope factor is mainly dictated by the potential close to the channel center rather than the surface. SCEs and drain-induced barrier lowering are also assessed using the proposed model and validated against a professional numerical device simulator.

Modeling of sonochemical decomposition of CCI4 in aqueous solutions

In the context of removal of organic pollutants from wastewater, sonolysis of CCl4 dissolved in water has been widely investigated. These investigations are either completely experimental or correlate data empirically. In this work, a quantitative model is developed to predict the rate of sonolysis of aqueous CCl4. The model considers the isothermal growth and partially adiabatic collapse of cavitation bubbles containing gas and vapor leading to conditions of high temperatures and pressures in them, attainment of thermodynamic equilibrium at the end of collapse, release of bubble contents into the liquid pool, and reactions in the well-mixed pool. The model successfully predicts the extent of degradation of dissolved CCl4, and the influence of various parameters such as initial concentration of CCl4, temperature, and nature of gas atmosphere above the liquid. in particular, it predicts the results of Hua and Hoffmann (Environ. Sci Technol, 1996, 30, 864-871), who found that degradation is first order with CCl4 and that Argon as well as Ar-O-3 atmospheres give the same results. The framework of the model is capable of quantitatively describing the degradation of many dissolved organics by considering all the involved species.

Implementation of Art1 and Art2 Artificial Neural Networks on Ring and Mesh Architectures

The Artificial Neural Networks (ANNs) are being used to solve a variety of problems in pattern recognition, robotic control, VLSI CAD and other areas. In most of these applications, a speedy response from the ANNs is imperative. However, ANNs comprise a large number of artificial neurons, and a massive interconnection network among them. Hence, implementation of these ANNs involves execution of computer-intensive operations. The usage of multiprocessor systems therefore becomes necessary. In this article, we have presented the implementation of ART1 and ART2 ANNs on ring and mesh architectures. The overall system design and implementation aspects are presented. The performance of the algorithm on ring, 2-dimensional mesh and n-dimensional mesh topologies is presented. The parallel algorithm presented for implementation of ART1 is not specific to any particular architecture. The parallel algorithm for ARTE is more suitable for a ring architecture.

A direct adaptive neural command controller design for an unstable helicopter

This paper presents an off-line (finite time interval) and on-line learning direct adaptive neural controller for an unstable helicopter. The neural controller is designed to track pitch rate command signal generated using the reference model. A helicopter having a soft inplane four-bladed hingeless main rotor and a four-bladed tail rotor with conventional mechanical controls is used for the simulation studies. For the simulation study, a linearized helicopter model at different straight and level flight conditions is considered. A neural network with a linear filter architecture trained using backpropagation through time is used to approximate the control law. The controller network parameters are adapted using updated rules Lyapunov synthesis. The off-line trained (for finite time interval) network provides the necessary stability and tracking performance. The on-line learning is used to adapt the network under varying flight conditions. The on-line learning ability is demonstrated through parameter uncertainties. The performance of the proposed direct adaptive neural controller (DANC) is compared with feedback error learning neural controller (FENC).

Multiscale modeling of lamellar mesophases

The mesoscale simulation of a lamellar mesophase based on a free energy functional is examined with the objective of determining the relationship between the parameters in the model and molecular parameters. Attention is restricted to a symmetric lamellar phase with equal volumes of hydrophilic and hydrophobic components. Apart from the lamellar spacing, there are two parameters in the free energy functional. One of the parameters, r, determines the sharpness of the interface, and it is shown how this parameter can be obtained from the interface profile in a molecular simulation. The other parameter, A, provides an energy scale. Analytical expressions are derived to relate these parameters to r and A to the bending and compression moduli and the permeation constant in the macroscopic equation to the Onsager coefficient in the concentration diffusion equation. The linear hydrodynamic response predicted by the theory is verified by carrying out a mesoscale simulation using the lattice-Boltzmann technique and verifying that the analytical predictions are in agreement with simulation results. A macroscale model based on the layer thickness field and the layer normal field is proposed, and the relationship between the parameters in the macroscale model from the parameters in the mesoscale free energy functional is obtained.

Molecule Modeling of Human Pentameric alpha(7) Neuronal Nicotinic Acetylcholine Receptor and Its Interaction with its Agonist and Competitive Antagonist

The nicotinic Acetylcholine Receptor (nAChR) is the major class of neurotransmitter receptors that is involved in many neurodegenerative conditions such as schizophrenia, Alzheimer's and Parkinson's diseases. The N-terminal region or Ligand Binding Domain (LBD) of nAChR is located at pre- and post-synaptic nervous system, which mediates synaptic transmission. nAChR acts as the drug target for agonist and competitive antagonist molecules that modulate signal transmission at the nerve terminals. Based on Acetylcholine Binding Protein (AChBP) from Lymnea stagnalis as the structural template, the homology modeling approach was carried out to build three dimensional model of the N-terminal region of human alpha(7)nAChR. This theoretical model is an assembly of five alpha(7) subunits with 5 fold axis symmetry, constituting a channel, with the binding picket present at the interface region of the subunits. alpha-netlrotoxin is a potent nAChR competitive antagonist that readily blocks the channel resulting in paralysis. The molecular interaction of alpha-Bungarotoxin, a long chain alpha-neurotoxin from (Bungarus multicinctus) and human alpha(7)nAChR seas studied. Agonists such as acetylcholine, nicotine, which are used in it diverse array of biological activities, such as enhancements of cognitive performances, were also docked with the theoretical model of human alpha(7)nAChR. These docked complexes were analyzed further for identifying the crucial residues involved i interaction. These results provide the details of interaction of agonists and competitive antagonists with three dimensional model of the N-terminal region of human alpha(7)nAChR and thereby point to the design of novel lead compounds.

Denoising neural data with state-space smoothing: Method and application

Neural data are inevitably contaminated by noise. When such noisy data are subjected to statistical analysis, misleading conclusions can be reached. Here we attempt to address this problem by applying a state-space smoothing method, based on the combined use of the Kalman filter theory and the Expectation–Maximization algorithm, to denoise two datasets of local field potentials recorded from monkeys performing a visuomotor task. For the first dataset, it was found that the analysis of the high gamma band (60–90 Hz) neural activity in the prefrontal cortex is highly susceptible to the effect of noise, and denoising leads to markedly improved results that were physiologically interpretable. For the second dataset, Granger causality between primary motor and primary somatosensory cortices was not consistent across two monkeys and the effect of noise was suspected. After denoising, the discrepancy between the two subjects was significantly reduced.

Analysis Of The Energy Quantization Effects On Single Electron Inverter Performance Through Noise Margin Modeling

Possible integration of Single Electron Transistor (SET) with CMOS technology is making the study of semiconductor SET more important than the metallic SET and consequently, the study of energy quantization effects on semiconductor SET devices and circuits is gaining significance. In this paper, for the first time, the effects of energy quantization on SET inverter performance are examined through analytical modeling and Monte Carlo simulations. It is observed that the primary effect of energy quantization is to change the Coulomb Blockade region and drain current of SET devices and as a result affects the noise margin, power dissipation, and the propagation delay of SET inverter. A new model for the noise margin of SET inverter is proposed which includes the energy quantization effects. Using the noise margin as a metric, the robustness of SET inverter is studied against the effects of energy quantization. It is shown that SET inverter designed with CT : CG = 1/3 (where CT and CG are tunnel junction and gate capacitances respectively) offers maximum robustness against energy quantization.