Comparison of a spiking neural network and an MLP for robust identification of generator dynamics in a multimachine power system

Johnson, Cameron; Venayagamoorthy, G.K.; Mitra, Pinaki

doi:10.1016/j.neunet.2009.06.033

Cited by 16 publications

(7 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[26] So far, different types of neural network architectures and their performances have been studied. [27][28][29][30][31] This includes multilayer perceptrons (MLPs), radial basis functions (RBFs), recurrent neural networks (RNNs), and echo-state networks (ESNs). In this work, MLP neural network was used.…”

Section: Methods Artificial Neural Networkmentioning

confidence: 99%

Prediction of Chemical Inhibitors Efficiency for Reducing Deposition Thickness Using Artificial Neural Network

Lashkarbolooki

Seyfaee

Esmaeilzadeh

et al. 2014

Journal of Dispersion Science and Technology

View full text Add to dashboard Cite

Prediction of efficiency of chemical inhibitors to mitigation of deposition thickness is a key to developing crude oil transportation process. In this work, a feed-forward artificial neural network (ANN) algorithm has been applied to predict the influence of the mitigation effect of ethylene-covinyl acetate (EVA) copolymer and its combination with chloroform (C), acetone (A), P-xylene (PX), and petroleum ether (PE) on the deposition thickness in the pipeline. An optimized three-layer feed-forward ANN model using properties of the oil pipeline such as: inlet oil temperature, environmental (coolant mixture) temperature, oil Reynolds numbers; properties of injected inhibitor such as molecular weight, boiling point, and amount of injection; and time is presented. Different networks are considered and trained using 62661 data sets; the accuracy of the network is validated by 20888 testing data sets. To verify the network generalization, 29 different experiment data sets of four different set of inhibitors have been considered. It is found that the proposed ANN model is an alternative to experimentation and predicts deposition thickness without experimentation, vast information, and tedious and time-consuming calculations.

show abstract

Section: Methods Artificial Neural Networkmentioning

confidence: 99%

Prediction of Chemical Inhibitors Efficiency for Reducing Deposition Thickness Using Artificial Neural Network

Lashkarbolooki

Seyfaee

Esmaeilzadeh

et al. 2014

Journal of Dispersion Science and Technology

View full text Add to dashboard Cite

show abstract

“…The spiking neuron receiving a higher signal fires sooner than when the same neuron receives a weaker signal. The same idea is supported by the [24][25][26] and called delay coding which has been used in numerous research studies [3,6,10]. In this study the following formula is employed for encoding input variables into spike times…”

Section: Coding and Decodingmentioning

confidence: 99%

Spiking neural networks for identification and control of dynamic plants

Abiyev

Kaynak

Oniz

2012

2012 IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM)

View full text Add to dashboard Cite

In this paper a Spiking Neural Networks (SNN) -based model is developed for identification and control of dynamic plants. Spike Response Model (SRM) has been employed to design the model. The learning of the parameters of SNN is carried out using a gradient algorithm. For its use for identification and control purposes, a coding is applied to convert real numbers into spikes. The SNN structure is tested for the identification and control of the dynamic plants commonly used in the literature. It has been found that the proposed structure results in a good performance despite its smaller parameter space.

show abstract

“…Some effort has been made to take the ease of inputting data into first-generation networks and still capture the power of an SNN. The model in [7] and the application thereof in [8] utilize spiking probabilities as a complicated activation function. It is demonstrated in [8] that this is a more powerful function approximator than an MLP, but it remains merely inspired by SNNs, rather than truly operating on spike propagation and recursion.…”

Section: Introductionmentioning

confidence: 99%

“…Using the centers of each neuron/Gaussian function and their width, the magnitude of firing f(x) for each sample points of the input x is calculated using(8).…”

mentioning

confidence: 99%

A reversibility analysis of encoding methods for spiking neural networks

Johnson

Roychowdhury

Venayagamoorthy

2011

The 2011 International Joint Conference on Neural Networks

View full text Add to dashboard Cite

There is much excitement surrounding the idea of using spiking neural networks (SNNs) as the next generation of function-approximating neural networks. However, with the unique mechanism of communication (neural spikes) between neurons comes the challenge of transferring real-world data into the network to process. Many different encoding methods have been developed for SNNs, most temporal and some spatial. This paper analyzes three of them (Poisson rate encoding, Gaussian receptor fields, and a dual-neuron n-bit representation) and tests to see if the information is fully transformed into the spiking patterns. An oft-neglected consideration in encoding for SNNs is whether or not the realworld data is even truly being introduced to the network. By testing the reversibility of the encoding methods in this paper, the completeness of the information's presence in the pattern of spikes to serve as an input to an SNN is determined.

show abstract

Comparison of a spiking neural network and an MLP for robust identification of generator dynamics in a multimachine power system

Cited by 16 publications

References 7 publications

Prediction of Chemical Inhibitors Efficiency for Reducing Deposition Thickness Using Artificial Neural Network

Prediction of Chemical Inhibitors Efficiency for Reducing Deposition Thickness Using Artificial Neural Network

Spiking neural networks for identification and control of dynamic plants

A reversibility analysis of encoding methods for spiking neural networks

Contact Info

Product

Resources

About