Fig. 10
From: Enhancing thin-film wafer inspection with a multi-sensor array and robot constraint maintenance
Training results from the Riemannian manifold VAE: (a) ELBO loss averaged across 10 runs, indicating convergence on a stable manifold; (b) Variance measure of the latent space. This variance takes low values in areas that the manifold has a high confidence of performance and high values in areas of high uncertainty; (c) The magnification factor J in Eq. 14 applied to the variance measure metric. The white dots indicate the training data, with a boundary around those points of high variance indicating the edge of the manifold.
