Deep learning for magnetic filed estimation

Hello i hope this message find all of you well. i have problem for the existing paper they estimate the flux distribution for transformer and motor how they make dataset .they mention they use Latin hypercube method for data set

Maybe this is useful?

Do you have a link to the paper? Just generally speaking, I get a little worried when people talk about solving physics problems with Deep Learning. Deep Learning is good at learning patterns, but it can only handle what it has actually seen in terms of data. The laws of physics apply in all cases, not just the ones you’ve seen before, right?

Here is the link Deep Learning for Magnetic Field Estimation | IEEE Journals & Magazine | IEEE Xplore you can check it sir