Mike Evans of University of Toronto developed the “relative belief” (RB) paradigm to perform hypothesis assessment: plausible hypothesis is such that its posterior > prior. This presents a minimalistic and most complete way of extracting information from data concerning the plausibility of any hypothesis about the system without ad hoc rules and assumptions. The maximum-likelihood estimation method is, by definition the most plausible estimation rule compatible with this Bayesian paradigm. We applied this paradigm for the first time to quantum-optical applications in the context of plausible-dimension certification.
♠ More direct than hypothesis testing—no ad hoc significance-level construct or use of p values. It has numerous applications:
- State tomography for certifying the dimension of a CV system, here a time-frequency state (first-order HG mode) generated by University of Paderborn,
- Photon-source integrity validation (assessing whether the source is truly single-photon, for instance), etc.