Title :
Continuous-time temporal back-propagation
Author :
Day, Shawn P. ; Camporese, Daniel S.
Author_Institution :
Dept. of Electr. Eng., British Columbia Univ., Vancouver, BC, Canada
Abstract :
The back-propagation training technique for feedforward neural networks is extended to networks with delay elements in the connections. Such an extension has previously been described by E.A. Wan (1990) for discrete-time presentation of training patterns and weight updates. The authors present a continuous-time generalization of these results. Such networks can be used for temporal and spatiotemporal pattern recognition
Keywords :
delays; learning systems; neural nets; pattern recognition; back-propagation training; continuous-time temporal back-propagation; delay; discrete-time training pattern presentation; feedforward neural networks; spatiotemporal pattern recognition; temporal pattern recognition; weight updates; Adaptive filters; Delay; Feedforward neural networks; Feedforward systems; Feeds; Neural networks; Neurons; Noise cancellation; Nonhomogeneous media; Pattern recognition;
Conference_Titel :
Neural Networks, 1991., IJCNN-91-Seattle International Joint Conference on
Conference_Location :
Seattle, WA
Print_ISBN :
0-7803-0164-1
DOI :
10.1109/IJCNN.1991.155319
