Patent
1994-09-20
1995-10-03
Downs, Robert W.
G06F 1518
Patent
active
054558930
ABSTRACT:
The invention relates to a norms computation procedure applied within a neuronal network structure or a computer. It permits determination of the norm of the new synaptic coefficients without necessitating their explicit computation. To this end, the structure comprises a processing unit 10 which determines output potentials y.sub.k and variation .DELTA..sub.k, a storage unit 13 for the old and new norms and a computational unit 15 for the new norms. The latter are thus rapidly determined without needing to call the synaptic coefficients in memory. The procedure can be utilized with certain algorithms which necessitate a computation of norms.
REFERENCES:
patent: 4912651 (1990-03-01), Wood et al.
Hush et al., "Improving the Learning Rate of Back-Propagation with the Gradient Reuse Algorithm", IEEE International Conf. on Neural Networks, Jul. 24-27, 1988 pp. I-991-996.
Kung et al., "A Unifying Algorithm/Architecture for Neural Networks", Proc. Int'l Conf. on Acoustics, Speech and Signal Processing, v. 4, 1989, pp. 2565-2568.
Parker, "A Comparison of Algorithms for Neuron-Like Cells", AIP Conference Proc. 151 Neural Networks for Computing, 1986 pp. 327-332.
DE Rummelhart et al "Learning Internal Representations by Error Propagation", Parallel Distributed Processing vol. I (Formulations MIT 1986).
Downs Robert W.
Schreiber David
U.S. Philips Corporation
LandOfFree
Processing method and neuronal network structure applying the me does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Processing method and neuronal network structure applying the me, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Processing method and neuronal network structure applying the me will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-1082659