Quantum mechanics as a calculus for estimation under epistemic restriction

Abstract

Consider a statistical model with an epistemic restriction such that, unlike in classical mechanics, the allowed distribution of positions is fundamentally restricted by the form of an underlying momentum field. Assume an agent (observer) who wishes to estimate the momentum field given information on the conjugate positions. We discuss a classically consistent, weakly unbiased, best estimation of the momentum field minimizing the mean squared error, based on which the abstract mathematical rules of quantum mechanics can be derived. The results suggest that quantum wave function is not an objective agent-independent attribute of reality, but represents the agent's best estimation of the momentum, given the positions, under epistemic restriction. Quantum uncertainty and complementarity between momentum and position find their epistemic origin from the trade-off between the mean squared errors of simultaneous estimations of momentum field and mean position, with the Gaussian wave function represents the simultaneous efficient estimations, achieving the Cram\'er-Rao bounds of the associated mean squared errors. We then argue that unitary time evolution and wave function collapse in measurement are normative rules for an agent to update her/his estimation given information on the experimental settings.