News for hackers. Stuff that matters.
Re:Machine learning works spectacularly well, but mathematic... by shasta on Sunday December 06, 2015 @08:57PM (#10687657)
I was just tossing out an example of a higher dimensional function and counting the local minima. The reasoning that critical points tend to be saddle points, not minima, seems irrelevant if critical points are increasing exponentially for a typical higher dimension function.

There are no comments on this comment yet.