Fig. 10
From: A method for efficient Bayesian optimization of self-assembly systems from scattering data
Model-fitting results for a 6-parameter ODE model. The results reflect 100 rounds of search with 893 points evaluated following initial training on 200 points randomly selected from a radius 100 hypersphere. Each round results in a new predicted minimizer for each GP model. The correspoinding RMSD is displayed. The lowest RMSD for each GP model over all rounds is plotted as a large green filled circle. In the final subfigure, the inputs with the lowest RMSD for each GP model are displayed