Let W:[0,1]2→[−1,1]. Then
∣∣W∣∣□≤(1)∣∣W∣∣1≤(2)∣∣W∣∣2≤(3)∣∣W∣∣∞≤(4)1
And convergence Lp for p≥1 implies convergence in the cut norm
Proof
Easy to see (1) from the definition of the cut norm:
cut norm (kernels) W be a kernel in [0,1]2. Its cut norm is defined as
$\lvert \lvert W \rvert \rvert_{\square} = \sup_{S,T \subseteq [0,1]} \left|\int \int _{S\times T} W(u,v), du , dv \right|$$
Let
This is computing the L-1 norm restricted to a subset of the nodes, so ∣∣W∣∣□≤∣∣W∣∣1.
(2) and (3) are a common result from functional analysis, and (4) is because of the selected codomain for W.
Convergence in the cut norm follows immediately from this hierarchy.