The paper shows that patch's position update is randomly selected within a 2D L∞ ball, but looking at the code indicates that it seems to be randomly selected within a square domain.
I'm not sure to understand the question. I think the square domain can be seen as an $\ell_\infty$-ball around the corner of the patch. Or do you mean something else?
The paper shows that patch's position update is randomly selected within a 2D L∞ ball, but looking at the code indicates that it seems to be randomly selected within a square domain.
Looking forward to your reply