1986
A widely influential paper explains learning internal representations by propagating errors backward.
It helps revive multilayer neural-network research; the method has earlier roots.
Explore on Humanity Unfolds
