Abstract
Neural network simulations are often limited because of the time required for both the learning and the evaluation phase of the simulation. Our parallel digital LNcuro circuit drastically reduces these times by updating synaptic coefficients related to one neuron in parallel. Contributions of ‘input’ neurons to one output neuron are also computed in parallel.
LNeuro-chips can easily be associated using Transputer microprocessors as controllers. Boards communicate through the reconfigurable links provided by a SuperNode architecture. This allows to simulate large size-networks or structured network architectures like Multi-Layer Perceptrons.
We report demonstrations of a parallel system built with several LNcuro-chips, which include a local learning rule (on a ‘real-time’ application), and the famous Backpropagation algorithm.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
M. DURANTON, J. GOBERT and J.A. SIRAT, Digital VLSI Module for Neural Networks. Proc. of ‘nEuro’, Paris, June 6–8 1988
M. DURANTON, J.A. SIRAT, Learning on VLSI: a General Purpose Digital NeurochipIJCNN conference on Neural Networks, Washington DC, June 18–25, 1989
M. DURANTON, N. MAUDUIT A General Purpose Digital Architecture for Neural Network SimulationFirst IEE conf. on Artificial Neural Networks. Londres, October 16–18 1989.
M. Du RANTON, J.A. SIRAT Learning on VLSI: a General Purpose Digital NeurochipMini et Micros, N°323, 29 Mai 1989
J.A. SIRAT, J.R. VIALA, C. REMUS Image Compression with Competing Multilayer PerceptronsFirst IEE conf. on Artificial Neural Networks. Londres, October 16–18 1989.
T.D. SANGER Optimal Unsupervised Learning in Single Layer Feed Forward Neural NetworkNeural Networks, vol.2, N°6, pp 459–473, 1989
G.W. COTTRELL, P.W. MUNRO, D. Zipser Image compression by Error Backpropagation: A Demonstration of Extentional Programming Advances in Cognitive Science, vol. 2, supp1.1, 399.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 1990 Springer Science+Business Media Dordrecht
About this chapter
Cite this chapter
Theeten, J.B., Duranton, M., Mauduit, N., Sirat, J.A. (1990). The LNeuro-Chip: A Digital VLSI with on-Chip Learning Mechanism. In: International Neural Network Conference. Springer, Dordrecht. https://doi.org/10.1007/978-94-009-0643-3_11
Download citation
DOI: https://doi.org/10.1007/978-94-009-0643-3_11
Publisher Name: Springer, Dordrecht
Print ISBN: 978-0-7923-0831-7
Online ISBN: 978-94-009-0643-3
eBook Packages: Springer Book Archive