That is, if e.g. we say that V(a,b,c) means it's legit to have a V with pattern a at top, pattern b at lower left, pattern C at lower right and likewise for U(a,b,c) (but note, I think the U needs to be thought of the other way up -- the "top" is the bit by the "nipple" on the curvy bit) then ...
... then e.g., the pattern at lower left on P2 says: there are some X and Y such that V(empty+beige, X, Y) and U(X,blue,lilac) and U(Y,blue+yellow,blue+yellow+red).
Where of course my use of symbols like + isn't meant to beg any questions, I'm just gesturing towards the actual arrangements of dots on the image.
Yes, the Us are telling you how to split up a given bunch of dots.
(maybe the dots should be thought of as prime factors, indeed)
I hope the Us don't just represent multiplication, because in that case they convey no information at all.
The one at lower right of P2 is a bit strange; it's oddly spaced compared to the others. I wonder whether it means V(empty,black+X,red) where also V(X,Y,Z) and V(Y,...,...) and U(Z,...,...). That will be impossible to make sense of without comparing against the actual picture :-).
Actually, I notice that I'm really confused about this, because some of these things I'm claiming are composite relations have a bunch of dots in the "shared" place and some don't. Maaaaybe the latter specifically have "no dots" in that place, but that doesn't seem likely. The only other way I see to fit this with my theory is that sometimes the aliens have kindly indicated what X and Y are, and sometimes not. That's a bit unsatisfactory.
So maybe I'm just entirely wrong.