Bayesian model calibration for block copolymer self-assembly: Likelihood-free inference and expected information gain computation via measure transport

Ricardo Baptista, Lianghao Cao, Joshua Chen, Omar Ghattas, Fengyi Li, Youssef M. Marzouk, J. Tinsley Oden

Abstract: Highlights•Bayesian inference of model parameters from experimental characterizations of block copolymer structures.•Integrated likelihood functions and high-dimensional image data lead to challenges for model calibration.•Consistent likelihood-free inference (LFI) is performed via a measure transport approach using only model evaluations.•Quantified the utilities of summary statistics and experimental designs via an efficient expected information gain estimator.