# Guido Sanguinetti: Machine learning for continuous time Markov chains

Continuous Time Markov Chains (CTMCs) play a key role in many areas of the physical and computational sciences. My own interest in CTMCs sparks from their use as models of biochemical reaction systems. Frequently, in biology, we can also obtain time-resolved observations of (some parts) of the system; however, the system is frequently not fully characterised, both in terms of its parametrisation and in terms of the structure of the underlying reaction network. Here, I will give a rather tutorial introduction to how ideas from machine learning could be used to perform statistical inference for CTMC systems. I'm aiming to explain both the fundamental ideas and some key derivations, but I may be overoptimistic and would welcome an interactive approach.