Fig. 4: Pareto fronts from multi-objective Bayesian optimization.
From: Refining coarse-grained molecular topologies: a Bayesian optimization approach
The plots show the evolution of the Pareto front for 10-PE and 50-PE over 60 optimization iterations (including 10 iterations for initial design). Each point represents a CG parameter set, plotted by its resulting density (\(\rho\)) and radius of gyration (\({R}_{g}\)). The color of the points indicates the iteration number. The AA target (red circle), initial Martini3 value (green triangle), and an optimal prediction from the final front (cyan circle) are highlighted.
