Unlike most algorithms applied to this problem, it neither represents surfaces parametrically (i.e. s(x,y,z) = f(u,v) ) nor implicitly ( f(x,y,z)=0 ) nor using a mesh of triangles but instead uses differential forms.
As a result, it makes no initial assumptions for the surface's topology ... most other minimal surface solvers require some sort of "starting point" sheet to minimize from and said starting point implies what end topology you'll get.
This algo seems to only require the boundary and work out a solution from it.
Quite surprisingly they also claim to have found a way to make convex a notoriously non-convex problem which, when you try to optimize it, typically yields local minima.
Their claim to have transformed the problem into a convex one allows them to find the global minimum. I would have intuitively bet that this would not be possible.
The math is TBH too heavy for me to digest, but if the claim holds, this is quite an impressive feat.