Associative Memory in Sparse Phasor Neural Networks
A. J. Noest · Europhysics Letters (EPL) · 1988
A novel network type is introduced which uses unit-length 2-vectors as local variables. A sparse, asymmetric associative memory net is defined and its recall performance is solved exactly. Real systems corresponding to such “phasor” models can be (neuro-)biological networks of limit-cycle oscillators, or optical resonators with saturating gain and holographic feedback.