An interesting open problem is whether there exists a finitely dependent invariant process which cannot be expressed as an ffiid with finite expected coding radius.
Statement
An interesting open problem is whether there exists a finitely dependent invariant process which cannot be expressed as an ffiid with finite expected coding radius. Markov random fields. Another class of processes of interest are Markov ran- dom fields (MRFs), which are the higher dimensional analogue of Markov chains. These are processes X = (Xv )v\inZd with the property that the conditional dis- tribution of the process on any given finite set V given the process outside of V depends only on the values on the boundary of V . In one dimension, a finite-state Markov chain is ffiid if and only if it is ergodic (aperiodic and irreducible) [6], and a countable-state Markov chain is ffiid if and only if it is exponentially ergodic (in the sense that the return time to a fixed state has exponential tails) [1]. The situation in higher dimensions is not fully understood. For the well-known Ising model on Zd , van den Berg and Steif [2] proved that the plus state Gibbs measure is ffiid if and only if there is no phase coexistence (i.e., the plus and minus states coincide, which is known to occur precisely when the temperature is at or above the critical temperature). The picture is not as complete for general MRFs on Zd . On the one hand, it is known that if two ergodic MRFs have the same conditional distributions (this corresponds to phase coexistence in models
Record
Comments
Nothing has been published against this problem yet, and nothing has been said about it.
Sign in with an institutional address to take part in the discussion. Reading every thread stays open to everyone.
Sign inSolve with an agent
Open the statement in a chat, with the problem and the ground rules already written into the prompt.