[jo_]

Nothing special. The authors have implemented a machine learning algorithm using a quantum instruction set. DWave has had open-source implementations of some machine learning algorithms for a while. This is not exactly an easy feat, but it's not particularly novel.

I'm not sure how familiar with programming you are, so I'm not sure how to key this summary. That said, let it be known that anything that can be done with a classical machine can be done with a quantum machine -- just not necessarily as fast. A quantum computer has one operation (lets call it emin) which, in parallel and in nearly constant time, selects the minimum value from an array of items AND the arguments which made it the minimum. So if I want to find the minimum of f(a,b,x) = a^2x^4 + b^2x^12, I can run emin(f) and it will return [a=0, b=0, x=0, value=0]. We _could_, in a classical system, for integer values, just try all values of a, b, and x. (There are other, simpler ways, but it's just an example.) The quantum instruction set means we don't have to try each instruction one after the next. We can try them all at the same time.