Schr\"odinger Bridges via the Hacking of Bayesian Priors in Classical and Quantum Regimes

TL;DR

This paper demonstrates how to manipulate prior distributions to achieve desired Bayesian updates in classical and quantum systems, linking prior hacking to Schr"odinger bridge problems and establishing a new perspective on belief updating.

Contribution

It introduces the concept of prior hacking in classical and quantum Bayesian inference, providing algorithms and establishing a duality with Schr"odinger bridge problems.

Findings

Prior hacking can preserve beliefs while appearing to perform Bayesian updates.

A duality between prior hacking and Schr"odinger bridge problems is established.

Unique inference-consistent Schr"odinger bridges are identified in the quantum setting.

Abstract

Bayes' rule is widely regarded as the canonical prescription for belief updating. We show, however, that one can arbitrarily preserve pre-specified beliefs while appearing to perform Bayesian updates via "prior hacking": engineering a reference prior distribution such that, for a fixed channel and evidence, the update matches a chosen target distribution. We prove that this is generically possible in both classical and quantum settings whenever Bayesian inversions are well-defined (with the Petz recovery map as the quantum analogue to Bayes' rule), and provide constructive algorithms for doing so. We further establish a duality between prior hacking and Schr\"odinger bridge problems (a key object in statistical physics with applications in generative modelling), yielding in the quantum setting a unique, inference-consistent selection among candidate bridges. This formally establishes the Bayes-like updating that Schr\"odinger bridges are performing with respect to the process as opposed to the reference prior, both in classical and quantum settings.