Information-acquiring von Neumann architecture of a computer: A theoretical design

Abstract

We design the information-acquiring von Neumann architecture of a computer in a fine-grained or coarse-grained model of the registers (quickly accessible memories) in the central processing unit, where information is carried by classical bits. This architecture enables both a Hamiltonian evolution process converting a given input pure state to another output pure state of the system under consideration (functionality) and a physical process to acquire information. The latter process is identified with the projection hypothesis (state reduction) in projective quantum measurement in the ensemble interpretation of quantum mechanics. As a novelty of this work, we treat projective quantum measurement as a classical measurement (reduction of ignorance) in the coarse-grained model. The main objective is to examine the present author's previously proposed state-reduction mechanism in the architecture within quantum electrodynamics in the presence of the orbital superselection rule. As a result, the electric potential incorporated into the architecture serves as a binary switch for the state reduction. As a consequence of this architecture, information-acquiring artificial intelligence (physically grounded induction machines or agents) can be established.