Neuronic Equations Revisited and Completely Solved

Abstract

Since my first meeting with Cybernetics, it has been crystal clear to me that its true object is the study of “intelligence” (calling it “natural” or “artificial” is misleading and causes only confusion). Can a physicist, with his mind and tools, have a say in such matter? Although today the question easily receives brazenly affirmative answers (I fear that too many problems lay ahead still totally ignored), at the time of my first endeavours the situation stood quite differently. This was lucky, in that little bias could then poison a young, enquiring mind; luckier perhaps than today, when so many claim problems to be “almost solved” and pour down answers, when even questions cannot yet be soundly formulated. For this reason I shall refrain from sweeping statements and limit my discussion to only one of the three basic elements of my model (see below): a set of nonlinear equations that describe the behaviour of a system of coupled binary decision elements in discrete time (“neuronic equations”: NE). Their solution is in any case an essential preliminary to that of the remaining parts, “mnemonic equations” and “adiabatic learning hypothesis”, which are strongly connected to anatomical, or technological, structural information.