Sebastian Raschka

@fireplyai That would be an interesting follow-up experiment. I.e., training the model on the same dataset with the traces hidden/removed.