> The team also used machine learning to identify patterns in the signals. > he results showed a clear difference between the signals before and after attack.

> The team also used machine learning to identify patterns in the signals. > he results showed a clear difference between the signals before and after attack. If they didn't use a rigorous separation of train, test and validation sets, I'm skeptical that this might have been overtraining.