Reliable Memory from Unreliable Components

We return to the antiphon structure considered in earlier papers, and present exact results for its memory capacity. However, for this structure it is supposed that one set of nodes (the α-nodes) behaves reliably and can realise a perfect cut-off. We consider the effect of relaxing this unrealistic assumption.

[1]  P. Whittle,et al.  A stochastic model of an artificial neuron , 1991, Advances in Applied Probability.

[2]  Peter Whittle,et al.  The antiphon: a device for reliable memory from unreliable components. I , 1989, Proceedings of the Royal Society of London. A. Mathematical and Physical Sciences.

[3]  P. Whittle.  Neural Nets and Implicit Inference , 1991 .

[4]  J. G. Taylor.  New Developments in Neural Computing , 1989 .