Bayesian optimization for material discovery processes with noise

Abstract

An augmented Bayesian optimization approach is presented for materials discovery with noisy and unreliable measurements. A challenging non-Gaussian, non-sub-Gaussian noise process is used as a case study for the discovery...

Data science accelerates energy device development

“…Process optimization using ML techniques has the advantages of rapidly assessing and screening a large number of parameters and iteratively converging to optimal decisions. 48,49 A popular approach used in the scientific community research is active learning, a subfield of AI, and have been demonstrated in virtual materials discovery, [50][51][52][53] synthesis, 54,55 and electrochemical performance testings. 49 In this section, we review two case studies, where active learning techniques were applied to optimize device fabrication and device operation, respectively.…”

Computational and data-driven modelling of solid polymer electrolytes

“…Moreover, Diwale et al presented an augmented BO method to overcome the noise issues in either experiments or simulations. 102…”

“…In most materials design scenarios, the acquisition function is of less importance, where most studies employed the expected improvement and/or upper (lower) condence bound. [100][101][102][103] However, the evaluation of the objective, i.e. the mapping X / y, is of key interest in most cases applying BO for materials design.…”

“…Moreover, Diwale et al presented an augmented BO method to overcome the noise issues in either experiments or simulations. 102 Besides applications in so nanomaterials, BO has also been extensively applied in energy storage materials, 100 microstructures of nanomechanical resonators, 108 and alloy design using multi-delity approaches. 109,110 From the optimization process perspective, Nakayama et al surveyed the use of acquisition functions and initial values in the BO materials synthesis as a simplied 1D case.…”

Optimizing telescoped heterogeneous catalysis with noise-resilient multi-objective Bayesian optimization

“…However, the tedious process of eliminating internal and external mass transfer and determining residence time limits the efficiency and accuracy of kinetic modeling. In addition, experimental results of heterogeneous catalysis often come with noise introduced by temperature, barometric pressure, and other environmental factors, which may lead to lower search efficiency and convergence properties in the optimization 19 . Our previously developed MOBO platform 20 used NEHVI, a MOBO acquisition function for hypervolume maximization in both noisy and noise-free environments 21 , but the comparison of different acquisition functions in terms of their ability to handle noise has not been systematically studied.…”