Ram N. Yadav

Time series prediction with single multiplicative neuron model

Single neuron models are typical functional replica of the biological neuron that are derived using their individual and group responses in networks. In recent past, a lot of work in this area has produced advanced neuron models for both analog and binary data patterns. Popular among these are the higher-order neurons, fuzzy neurons and other polynomial neurons. In this paper, we propose a new neuron model based on a polynomial architecture. Instead of considering all the higher-order terms, a simple aggregation function is used. The aggregation function is considered as a product of linear functions in different dimensions of the space. The functional mapping capability of the proposed neuron model is demonstrated through some well known time series prediction problems and is compared with the standard multilayer neural network.

Channel Equalization Using Neural Networks: A Review

Equalization refers to any signal processing technique used at the receiver to combat intersymbol interference in dispersive channels. This paper reviews the applications of artificial neural networks (ANNs) in modeling nonlinear phenomenon of channel equalization. The literature associated with different feedforward neural network (NN) based equalizers like multilayer perceptron, functional-link ANN, radial basis function, and its variants are reviewed. Feedback-based NN architectures like recurrent NN equalizers are described. Training algorithms are compared in terms of convergence time and computational complexity for nonlinear channel models. Finally, some limitation of current research activities and further research direction is provided.

Prediction of subsidence due to coal mining in Raniganj coalfield, West Bengal, India

Abstract Raniganj coalfield is in a major coal-producing area and suffers subsidence problems due to underground mining. The occurrence of a thick coal seam at shallow depth is the main reason for the subsidence and as a result collapses have occurred in many coal mines in recent years. We attempt to predict the subsidence occurring in Indian coalfields in general and, in particular, in the Raniganj coalfield using a visco-elastic model. In this model, we have considered the coal seam as a beam resting between visco-elastic media which deform due to transverse shear after mining. The governing equations have been solved for two cases, coal mines with (a) rigid (hard), and (b) flexible (soft) overburdens. The ground subsidence has been computed for two coal mines. The computed subsidence profiles of the Ratibati and Shivadanga coal mines of the Raniganj coalfield are compared and the results show a reasonable match between the predicted and observed subsidence in the case of Ratibati coal mine. In the case of Shivadanga coal mine the predicted subsidence profile is found to be larger than the observed profile. However, the bottom of the observed troughs is found to match the predicted troughs. Detailed analyses have been carried out to investigate the effect of mine parameters on the subsidence.

Efficient face recognition using wavelet-based generalized neural network

This paper presents an efficient face recognition method where enhanced local Gabor binary pattern histogram sequence has been used for efficient face feature extraction and generalized neural network with wavelet as activation function is being used for classification. In this method the face is first decomposed into multiresolution Gabor wavelets the magnitude responses of which are applied to enhanced local binary patterns. The efficiency has been significantly improved by combining two efficient local appearance descriptors named Gabor wavelet and enhanced local binary pattern with generalized neural network. Generalized neural network is a proven technique for pattern recognition and is insensitive to small changes in input data. The proposed method is robust-to-slight variation of imaging conditions and pose variations. Performance comparison with other existing techniques shows that the proposed technique provides better results in terms of false acceptance rate, false rejection rate, equal error rate and time complexity.

Classification using single neuron

Since the neuron is the basic information processing unit of the brain, the ANN have played a great role in the study of the brain. Because of the complexity and less understanding about the biological neurons, many scientists and researchers have given various architecture for it. Experimental studies in the area of neuroscience has proven that the response of a biological neuron appears random and the predicted results can be obtained in many ways. We have presented some new models of the artificial neuron that can be used to solve the various bench-mark problems in a very simple and systematic manner.

Time-series prediction with single integrate-and-fire neuron

In this paper, a learning algorithm for a single integrate-and-fire neuron (IFN) is proposed and tested for various applications in which a multilayer perceptron neural network is conventionally used. It is found that a single IFN is sufficient for the applications that require a number of neurons in different hidden layers of a conventional neural network. Several benchmark and real-life problems of classification and time-series prediction have been illustrated. It is observed that the inclusion of some more biological phenomenon in an artificial neural network can make it more powerful.

Regularized Weighted Circular Complex-Valued Extreme Learning Machine for Imbalanced Learning

Extreme learning machine (ELM) is emerged as an effective, fast, and simple solution for real-valued classification problems. Various variants of ELM were recently proposed to enhance the performance of ELM. Circular complex-valued extreme learning machine (CC-ELM), a variant of ELM, exploits the capabilities of complex-valued neuron to achieve better performance. Another variant of ELM, weighted ELM (WELM) handles the class imbalance problem by minimizing a weighted least squares error along with regularization. In this paper, a regularized weighted CC-ELM (RWCC-ELM) is proposed, which incorporates the strength of both CC-ELM and WELM. Proposed RWCC-ELM is evaluated using imbalanced data sets taken from Keel repository. RWCC-ELM outperforms CC-ELM and WELM for most of the evaluated data sets.

Learning with generalized-mean neuron model

Abstract The artificial neuron has come a long way in modeling the functional capabilities of various neuronal processes. The higher order neurons have shown improved computational power and generalization ability. However, these models are difficult to train because of a combinatorial explosion of higher order terms as the number of inputs to the neuron increases. This work presents an artificial neural network using a neuron architecture called generalized mean neuron (GMN) model. This neuron model consists of an aggregation function which is based on the generalized mean of the all the inputs applied to it. The proposed neuron model with same number of parameters as the McCulloch–Pitts model demonstrates better computational power. The performance of this model has been benchmarked on both classification and time series prediction problems.

Complex channel equalization using polynomial neuron model

The Artificial Neural Networks (ANN) has been applied to channel equalization with quite promising results. Although an ANN takes time during it’s training, it generates instant results during its implementation phase. ANN are capable of performing complex non-linear mapping between their input and output space. In this paper we propose a new complex neural equalizer based on a simple model of polynomial neuron. A well-defined training procedure based on back propagation is used. The low complexity equalizer with three input nodes, three hidden nodes and one output node shows good tracking performance at even lower values of signal to noise ratio (SNR). The equalizer is tested on 4 QAM complex signals used in satellite channels.

Neural network learning with generalized-mean based neuron model

The advances in biophysics of computations and neurocomputing models have brought the foreground importance of dendritic structure of neuron. These structures are assumed as basic computational units of the neuron, capable of realizing the various mathematical operations. The well structured higher order neurons have shown improved computational power and generalization ability. However, these models are difficult to train because of a combinatorial explosion of higher order terms as the number of inputs to the neuron increases. In this paper we present a neural network using new neuron architecture i.e., generalized mean neuron (GMN) model. This neuron model consists of an aggregation function which is based on the generalized mean of all the inputs applied to it. The resulting neuron model has the same number of parameters with improved computational power as the existing multilayer perceptron (MLP) model. The capability of this model has been tested on the classification and time series prediction problems.