Someone has represented defense in depth as slices of Swiss cheese.
Why not a plate of pasta? Is your security strategy like spaghetti… far more adaptive than a hard cheese?
Ironically, this paper is about “machine learning”, which suggests to me anything capable of minimal learning would breeze right through Swiss cheese holes like an obstacle course. Bad model.
I know I say “food for thought” too often on this blog, yet here it really seems the most appropriate phrase ever.