Table 4 Comparative results of GAN and NSGAN methods in generation performance

From: NSGAN: a non-dominant sorting optimisation-based generative adversarial design framework for alloy discovery

Model

Data generated

Average elongation

Average UTS

Samples within pre-defined criterion (%)

Average novelty

GAN

1,000,000

13.15%

317.4 MPa

116 (0.01)

0.0096

GA+ML

4000

31.32%

302.73 MPa

563 (14.07)

0.0199

NSGAN

4000

18.97%

453.91 MPa

163 (4.08)

0.0145