Would it be possible to do Bayesian inference over $(\phi, q)$ rather than just $\phi$?
Potentially accepting some prior over $q$?
This could easily be extended in the case of sequential Monte Carlo Bayesian inference, I wonder how it could be done in terms of exact (Fourier) or moment-matching (von Mises) - perhaps could make an independence approximation $p(\phi, q) = p(\phi) p(q)$.
The nice thing is that this could be done without additional quantum overhead (i.e. same number of shots per update).
Would it be possible to do Bayesian inference over$(\phi, q)$ rather than just $\phi$ ?
Potentially accepting some prior over$q$ ?
This could easily be extended in the case of sequential Monte Carlo Bayesian inference, I wonder how it could be done in terms of exact (Fourier) or moment-matching (von Mises) - perhaps could make an independence approximation$p(\phi, q) = p(\phi) p(q)$ .
The nice thing is that this could be done without additional quantum overhead (i.e. same number of shots per
update).