Rapid design of top-performing metal-organic frameworks with qualitative representations of building blocks

Abstract Data-driven materials design often encounters challenges where systems possess qualitative (categorical) information. Specifically, representing Metal-organic frameworks (MOFs) through different building blocks poses a challenge for designers to incorporate information into optimization, and leads combinatorial challenge, with large number of MOFs that could be explored. In this work, we integrated Latent Variable Gaussian Process (LVGP) Multi-Objective Batch-Bayesian Optimization (MOBBO) identify top-performing adaptively, autonomously, efficiently. We showcased our method (i) requires no specific physical descriptors only uses construct the global optimization representations, (ii) is application property independent, (iii) provides an interpretable model justification. By searching ~1% space, LVGP-MOBBO identified all on Pareto front 97% 50 designs CO 2 working capacity /N selectivity properties.