The generic constancy conjecture for causal stochastic solutions

Let \Pcin\Pc_{\mathrm{in}}^- be the space of input probability laws and \Pcstate\Pc_{\mathrm{state}}^- the space of state–input probability laws, with projection

(π\US) ⁣:\Pcstate\Pcin.(\pi_{\US^-})_* \colon \Pc_{\mathrm{state}}^- \rightarrow \Pc_{\mathrm{in}}^-.

Let #\Scc,stoch(Ξ)\#\Sc^{\mathrm{c,stoch}}(\Xi) and #\Occ,stoch(Ξ)\#\Oc^{\mathrm{c,stoch}}(\Xi) denote, respectively, the numbers of causal stochastic solutions and causal stochastic outputs associated with an input law Ξ\Xi. Assume that the state-space system measurably distinguishes reachable states when the relevant reachable-state relation is Borel measurable and admits a measurable readout of the next output.

Generic constancy conjecture. Suppose that \Pcin\Pc_{\mathrm{in}}^- and \Pcstate\Pc_{\mathrm{state}}^- are Polish and that (π\US) ⁣:\Pcstate\Pcin(\pi_{\US^-})_* \colon \Pc_{\mathrm{state}}^- \rightarrow \Pc_{\mathrm{in}}^- is proper. Then, #\Scc,stoch(Ξ)\#\Sc^{\mathrm{c,stoch}}(\Xi) is generically constant, and so is #\Occ,stoch(Ξ)\#\Oc^{\mathrm{c,stoch}}(\Xi) if the state-space system measurably distinguishes reachable states.

This is posed as an open problem because the paper does not know whether the causal stochastic FMP holds generically; the conjecture would yield generic control of bifurcations of causal stochastic solutions and outputs.

Sources & referencesView supporting material

Primary source

Juan-Pablo Ortega and Florian Rossmannek, “Stochastic dynamics learning with state-space systems”, arXiv:2508.07876 (2026).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.