Fig. 5 | Scientific Reports

Fig. 5

From: Performance of uncertainty-based active learning for efficient approximation of black-box functions in materials science

Fig. 5

(a) Results of \(\langle{\Delta}{R}^{2}\rangle\) in a two-dimensional space, where the variance of the objective functions \(\:\{{y}_{i}{\}}_{i=1,\dots\:,N}\) is vertical axis and the dimension of the inputs for a BBF is horizontal axis, when \(\:{N}_{\text{i}\text{n}\text{i}}\:=\:10\). (b) Results of \(\langle{\Delta}{R}^{2}\rangle\) for these properties after 100 iterations when \(\:{N}_{\text{i}\text{n}\text{i}}\:=\:10,\:100\), and \(500\) for the bandgap with the matminer descriptor. The ML model is trained by GPR. The depicted dimensions in Panel (a) are active dimension which is summarized in Table S1.

Back to article page