Prior Knowledge in AI -- Is it Really "Cheating"?