“…3,[14][15][16] Learning this high dimensional function has attracted interest from a diversity of fields, and significant advancements has been made through methods that employ importance sampling and machine learning. [16][17][18][19][20][21][22][23][24][25][26][27][28] Some notable approaches have leveraged the confinement of the transition region to compute it using string methods, 16,17,29 coarse-grained the phase-space to approximate it through diffusion maps, 19,28,30,31 and parameterized neural-networks by either fitting the committor directly 18,21 or solving the variational form of the steady-state backward Kolmogorov equation 22 by combining it with importance sampling methods. [23][24][25] While the learning procedures applied previously have been successful in fitting high dimensional representations of the reaction coordinate or committors, their nonlinearity has largely resulted in a difficulty in interpreting the relative importance of physically distinct descriptors and converting those descriptors into a robust measure of the rate.…”