Learning collective variables that preserve transition rates

We turn Legoll and Lelievre’s quantitative coarse-graining theory into an algorithm for learning collective variables that preserve transition rates in molecular systems, with a case study on butane.