Fault tolerant characteristics of the linear array architecture for WSI implementation of neural nets

A novel digital architecture supporting implementation of feedforward, multilayered artificial neural networks is presented. Based on a switched bus array philosophy, particular care is taken to minimize area requirements while maximizing throughput and parallelism. The architecture is well suited to support, with minimal additional changes, fault/defect tolerance with respect to faults located in the PEs (processing elements), switches, or buses.<<ETX>>

[1]  John J. Hopfield,et al.  Simple 'neural' optimization networks: An A/D converter, signal decision circuit, and a linear programming circuit , 1986 .

[2]  L. E. Atlas,et al.  A study of regular architectures for digital implementation of neural networks , 1989, IEEE International Symposium on Circuits and Systems,.

[3]  Mariagiovanna Sami,et al.  A configurable array architecture for WSI implementation of neural nets , 1990, Ninth Annual International Phoenix Conference on Computers and Communications. 1990 Conference Proceedings.

[4]  F. Blayo,et al.  A reconfigurable WSI neural network , 1989, [1989] Proceedings International Conference on Wafer Scale Integration.

[5]  Mariagiovanna Sami,et al.  Multistage Interleaved Architectures for Implementation of Neural Networks , 1989 .

[6]  S. Y. Kung,et al.  Digital VLSI architectures for neural networks , 1989, IEEE International Symposium on Circuits and Systems,.