Fig. 3: Dual-GP driven exploration in a simulated scenario where a sub-space contains high noise.
From: Active oversight and quality control in standard Bayesian optimization for autonomous experiments
a amplitude of the synthetic peak model data for the Dual-GP test, here the noise quickly increases when x > 0.7, leading to resultant scalarizers diverging from the true function. b Examples of a good raw spectrum and noisy spectra. c Single GP exploration result after 50 iterations, d Dual-GP exploration result after 50 iterations. e–g Show the sGP parameter prediction of Dual-GP approaches the true values more quickly than the single GP.
