Limits of action convergent graph sequences with unbounded $(p,q)$-norms
Aranka Hru\v{s}kov\'a

TL;DR
This paper explores the limits of graph sequences under action convergence when only the $( ext{infinity,}1)$-norm is bounded, revealing non-uniqueness and non-self-adjointness of limit operators, thus showing the space of graphops is not compact.
Contribution
It constructs graph sequences with only bounded $( ext{infinity,}1)$-norms that still converge, demonstrating the non-uniqueness and non-self-adjointness of their limits.
Findings
Limit operators are not unique in this setting.
Limit operators may not be self-adjoint or positivity-preserving.
The space of graphops is not compact under action convergence.
Abstract
The recently developed notion of action convergence by Backhausz and Szegedy unifies and generalises the dense (graphon) and local-global (graphing) convergences of graph sequences. This is done through viewing graphs as operators and examining their dynamical properties. Suppose is a sequence of operators representing graphs, Cauchy with respect to the action metric. If has uniformly bounded -norms where is any pair in , then Backhausz and Szegedy prove that has a limit operator which, moreover, must be self-adjoint and positivity-preserving. In the present work, we construct a large class of graph sequences whose only uniformly bounded -norm is the -norm, but which converge nonetheless. We show that the limit operators in this case are not unique, not self-adjoint, and need…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsMathematical Dynamics and Fractals · semigroups and automata theory · Rings, Modules, and Algebras
