Fig. 2
From: Inferring broken detailed balance in the absence of observable currents
Decimation of Markov processes. a, b Molecular motor model: An observer with access only to the position (vertical axis) cannot resolve the internal states (circles). a Decimation to position results in a first-order Markov process, since spatial jumps connect the same internal state. b Decimation results in a second-order semi-Markov process, where the waiting time distribution for spatial transitions depends on whether the motor previously jumped down or up. c Hidden kinetic network: An observer unable to resolve states 3, 4, and 5, treats them as a single hidden state H. The resulting decimated network is a second-order semi-Markov process on the three states 1, 2, and H, where the non-Poissonian waiting time distributions for transitions out of state H depend on the past
