Advancing Frameworks for Analyzing Complex Reaction Dynamics from Trajectory Data

One approach to understanding complex reaction dynamics is to estimate dynamical observables—such as committors, mean first passage times, and transition rates—from simulation data. This dissertation extends the applicability of these data-driven methods to more challenging systems—such as those with memory effects or multiple pathways—with a view towards enabling the estimation of richer dynamical observables. Key contributions include a robust method for estimating slow modes, the incorporation of memory effects via the Mori–Zwanzig formalism for accurate estimation of committors and related quantities, and a generalization of transition path theory that allows for direct analysis of specific reaction steps and pathways. Together, these advances enable detailed analysis of reaction kinetics and mechanism directly from simulation data.