Table 1 Comparison of PAL to the state-of-the-art

From: Efficient search of compositional space for hybrid organic–inorganic perovskites via Bayesian optimization

Method

Mean

STD

99.9th percentile

PAL

30.8

8.6

45

Simple BO

47.7

19.6

96

Hutter BO

51.6

21.9

104

pySMAC

61.6

34.7

160

Random

119.3

69.3

239

  1. Comparison of the performance of PAL is done against state-of-the-art optimization approaches, pySMAC, Simple BO, Hutter BO, as well as a random search of ABX3 compositions. The first column shows the average number of iterations required to find a composition with optimal mixed-halide ABX3 − S0 intermolecular binding energy. The standard deviation of this number is stated in column 2, while the third column gives the 99.9th percentile. PAL requires considerably fewer DFT calculations to find an optimal solution, on average, and shows a robust performance with its significantly lower 99.9th percentile
Back to article page