1994-09-29
1996-09-03
Bayerl, Raymond J.
395 21, G06E 100, G06E 300
Patent
active
055531954
ABSTRACT:
An information processing system includes a feed-forward neural net that is functionally comprised of a plurality of interconnected neurons. At least a particular one of the neurons is operative to receive a plurality of input signals for in response thereto producing an output signal that is a function of the input signals, and of at least one of following quantities: a rate of change of at least one of the neuron input signals; a rate of change of the particular neuron output signal. The net exhibits dynamic behavior as represented by signal-dependent delays, signal dispersion and phase shifts. The net can be trained to display stimulus-response characteristics of substantially any dynamic system.
REFERENCES:
patent: 4974169 (1990-11-01), Engel
patent: 5170071 (1992-12-01), Shreve
patent: 5200898 (1993-04-01), Yuhara et al.
patent: 5285522 (1994-02-01), Mueller
Doya, K, "Learning Temporal PAtterns in Recurrent Neurla Networks," IEEE, SMC Conf. 1990, pp. 170-172.
Bayerl Raymond J.
Katbab A.
Stephens Debra K.
U.S. Philips Corporation
LandOfFree
Dynamic neural net does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Dynamic neural net, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Dynamic neural net will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-1958170