Patent
1995-03-08
1998-01-06
Hafiz, Tariq R.
395 24, G06F 1500
Patent
active
057064008
ABSTRACT:
Any deterministic finite-state automata (DFA) can be implemented in a sparse recurrent neural network (RNN) with second-order weights and sigmoidal discriminant functions. Construction algorithms can be extended to fault-tolerant DFA implementations such that faults in an analog implementation of neurons or weights do not affect the desired network performance. The weights are replicated k times for k-1 fault tolerance. Alternatively, the independent network is replicated 2k+1 times and the majority of the outputs is used for a k fault tolerance. In a further alternative solution, a single network with k.eta. neurons uses a "n choose k"encoding algorithm for k fault tolerance.
REFERENCES:
patent: 5155801 (1992-10-01), Lincoln
Omlin and Giles, "Constructing Deterministic Finite-State Automata in Sparse Recurrent Neural Netowrks," IEEE International Conference on Neural Networks (ICNN '94), 1994, pp. 1732-1737.
May, "Fault simulation of a wafer-scale neural network," Oregon Graduate Center, Feb. 1988.
Wulff and Hertz, (1992) "Prediction with Recurrent Networks", Neural Networks for Signal Processing, Proceedings of the IEEE-SP Workshop, p. 464, Dec. 1992.
Phatak and Koren (1992). "Fault Tolerance of Feedforward Neural Nets for Classification," IEEE Neural Networks, 1992 International Conference, vol. 2., p. 386, Dec. 1992.
Giles and Miller (1992) "Learning and Extracting Finite State Automata with Second-Order Recurrent Neural Networks," Neural Computation, vol. 4, p. 393, Dec. 1992.
Sun et al. "Turning Equivalence of Neural Networks with Second Order Connection Weights," IEEE Neural Networks 1991 International Conference, vol. 2, p. 357, Dec. 1991.
Giles C. Lee
Omlin Christian Walter Peter
Chung Ji-Yong D.
Feig Philip J.
Hafiz Tariq R.
NEC Research Institute Inc.
LandOfFree
Fault-tolerant implementation of finite-state automata in recurr does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Fault-tolerant implementation of finite-state automata in recurr, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Fault-tolerant implementation of finite-state automata in recurr will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2336736