Markov Jump

Markov jump processes (MJPs) model systems transitioning between discrete states over continuous time, finding broad application in diverse fields from physics to biology. Current research focuses on improving inference methods for MJPs, employing techniques like tensor networks for efficient sampling, neural networks for zero-shot inference, and Markov Chain Monte Carlo for Bayesian parameter estimation. These advancements address the inherent challenges of inferring MJP parameters from noisy or sparse data, enabling more accurate modeling of complex dynamical systems and facilitating improved control strategies in various applications.