ARRAYLIKE LEARNING SYSTEMS
Manissa J. Dobrée Wilson, Igor Aleksander · Journal of Cybernetics · 1976
We shall present 20 results which relate state-transition behavior to the functions of arraylike networks. Arrays have been chosen (as opposed to, say, randomly connected networks) since they have an easily representable and sufficiently varied set of feedback loops. Our concern is also restricted to arrays in which all the elements perform the same function. This provides a position-invariant global action with respect to state patterns.