Originally posted by @sonhouse
You mean certain classes like sifting through large data sets looking for correlation of some characteristic Vs say calculating 3 body relativistic kind of problems?
Not sure exactly how that method would be 'applied' to 3 body relativistic kind of problems. I will have to think about that one.
As I understand it, what is called the "TWO-body problem in general relativity" is hard enough to solve (let alone some kind ot "THREE-body problem in general relativity"! ) and so far only approximate solutions have been found, no exact solution. See;
By the way, when they talk about "directions" in the context of the OP link about algorithms, they aren't usually talking about the special directions in particular but rather "direction" in a much more generic and abstract way where we consider 2 or more variables, which might have nothing whatsoever to do with the special directions but could be mass, vibration, how fat you are, how 'much' two drags interact etc etc and make a plot of data points with each variable being represented on a graph along a different graph axis and typically a multidimensional graph except if it has more than 3 axis then it is usually far to problematic to represent visually and thus such a 'graph' isn't normally represented visually but as a pure maths construct in which case it is called a "multidimensional network". See
I had once studied multidimensional networks at university but so long ago that I am afraid I think I have forgotten most of what I learned about them which I think is just terrible! If I can find the time later, I should revise this subject area because it is important; maybe also try and see how it might relate to my current research if at all.