Analog hardware implementation of the random neural network model

Abdelbaki, H.; Gelenbe, Erol; El‐Khamy, Said E.

doi:10.1109/ijcnn.2000.860772

Cited by 21 publications

(23 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Since all spikes need to be accounted for, for any we will have (1) where is the probability that neuron fires a spike of class , which is then directed outside the network rather to some other neuron. In the future, we will deal with the "network weights," which are defined as and we designate the corresponding weight matrices by and .…”

Section: Mathematical Modelmentioning

confidence: 99%

“…The RNN is easy to simulate, since each neuron is represented by a counter, which counts inhibitory spikes downwards ( ) and excitatory spikes upwards ( ), and can, therefore, lead to a simple hardware implementation [1]. Furthermore, the RNN model represents neuron potential as an integer, rather than as a binary variable, which leads to a more detailed state representation.…”

Section: Introductionmentioning

confidence: 99%

“…Furthermore, the RNN model represents neuron potential as an integer, rather than as a binary variable, which leads to a more detailed state representation. For instance, in the RNN the "average potential" of individual neurons (which takes values in 0 ) may also be used as the significant output variable of a neuron, rather than just the binary neuron state (on or off), or the value of neuron state represented by a real number in the interval [0,1].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Learning in the multiple class random neural network

Gelenbe

Hussain

2002

IEEE Trans. Neural Netw.

Self Cite

115

View full text Add to dashboard Cite

Spiked recurrent neural networks with "multiple classes" of signals have been recently introduced by Gelenbe and Fourneau (1999), as an extension of the recurrent spiked random neural network introduced by Gelenbe (1989). These new networks can represent interconnected neurons, which simultaneously process multiple streams of data such as the color information of images, or networks which simultaneously process streams of data from multiple sensors. This paper introduces a learning algorithm which applies both to recurrent and feedforward multiple signal class random neural networks (MCRNNs). It is based on gradient descent optimization of a cost function. The algorithm exploits the analytical properties of the MCRNN and requires the solution of a system of nC linear and nC nonlinear equations (where C is the number of signal classes and n is the number of neurons) each time the network learns a new input-output pair. Thus, the algorithm is of O([nC]/sup 3/) complexity for the recurrent case, and O([nC]/sup 2/) for a feedforward MCRNN. Finally, we apply this learning algorithm to color texture modeling (learning), based on learning the weights of a recurrent network directly from the color texture image. The same trained recurrent network is then used to generate a synthetic texture that imitates the original. This approach is illustrated with various synthetic and natural textures.

show abstract

Section: Mathematical Modelmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Learning in the multiple class random neural network

Gelenbe

Hussain

2002

IEEE Trans. Neural Netw.

Self Cite

115

View full text Add to dashboard Cite

show abstract

“…Associated with each vehicle, the sensors produce two intensity images, as shown in Figure 2. Following our work in (24) [23] in the two intensity images. [5] Maximum Height: calculating the vehicle maximum height from the two intensity images is a little harder since some computer vision issues must be addressed.…”

Section: Feature Extractionmentioning

confidence: 99%

“…The RNN is easy to simulate, since each neuron is represented by a counter, which counts inhibitory spikes downwards (-1) and excitatory spikes upwards (+1), and can, therefore, lead to a simple hardware implementation [23]. Furthermore, the RNN model represents neuron potential as an integer, rather than as a binary variable, which leads to a more detailed state representation.…”

mentioning

confidence: 99%

Laser intensity vehicle classification system based on random neural network

Hussain

Moussa

2005

Proceedings of the 43rd Annual Southeast Regional Conference - Volume 1

View full text Add to dashboard Cite

This paper presents a Laser Intensity Vehicle Classification System (LIVCS) based upon imagery obtained from range sensors (called LIVCS). Current systems that utilize loop detectors, video cameras, and range sensors have deficiencies. The loop detectors have high failure rates due to pavement failures and poor maintenance. Video based systems and range sensors do not perform well in deteriorated atmospheric conditions (such as rain and fog). The developed generations of image based range sensors offer the promise of sensors that are less sensitive to deteriorated environmental conditions. LIVCS system extracts features of laser intensity images, produced by laser sensory units. These features are used to train a random neural network (RNN). The LIVCS system recalls its trained RNN for classification of vehicles. This technique outperforms loop detectors, video cameras, and range data techniques in deteriorated environmental conditions.

show abstract

A survey on cognitive packet networks: Taxonomy, state-of-the-art, recurrent neural networks, and QoS metrics

Ray

2022

Journal of King Saud University - Computer and Information Scie

View full text Add to dashboard Cite

Analog hardware implementation of the random neural network model

Cited by 21 publications

References 8 publications

Learning in the multiple class random neural network

Learning in the multiple class random neural network

Laser intensity vehicle classification system based on random neural network

A survey on cognitive packet networks: Taxonomy, state-of-the-art, recurrent neural networks, and QoS metrics

Contact Info

Product

Resources

About