(599g) Multi-Fidelity Bayesian Optimization of Porous Materials for Gas Separations

Multi-fidelity Bayesian optimization (MFBO) constitutes an experiment-update-plan closed loop to cost-effectively find the optimal material, while leveraging multiple experiments that trade fidelity and cost. In this talk, we employ MFBO to efficiently find a covalent organic framework with the highest adsorptive selectivity for xenon over krypton, while leveraging both high-fidelity and low-fidelity molecular simulations.