Meta-diagrams for 2-parameter persistence (Q6878298)

This article considers invariants for multiparameter persistence modules. The authors introduce a novel invariant for 2-parameter persistence modules, the meta-rank, and use Möbius inversion to define a corresponding notion of meta-diagram, which plays a similar role for the meta-rank as the persistence diagram does for the barcode. The authors show that the meta-rank and meta-diagram contain equivalent information to the rank invariant and the signed barcode. Moreover, the authors provide an algorithm to compute the meta-rank and the meta-diagram in \(\mathcal{O}(n^4)\) time where \(n\) is the number of points in one parameter at which the isomorphism type of the module changes. The authors also present a way to visualize the meta-diagram. The main requirement for 2-parameter modules for the notion of meta-rank to be applicable is a mild notion of constructibility, which boils down to the isomorphism type of the module only changing at a discrete set of points.\N\NAfter recapturing preliminaries on multiparameter persistence modules and the rank invariant in section 2, the authors introduce the meta-rank by considering a notion of rank of morphisms of persistence modules between 1-parameter slices of a given 2-parameter persistence module in section 3. At the end of section 3 the authors state a stability result between 1-parameter modules with respect to interleaving distance and their induces meta-ranks with respect to a notion of erosion distance. The proof of this stability theorem is to be found in the appendix. Section 4 describes how to obtain the meta-diagram from the meta-rank, show it's equivalence to rank decompositions via rectangles, and state a stability theorem for meta-diagrams with respect to erosion distance and 2-parameter persistence modules with respect to interleaving distance. In section 5, the authors provide a theoretical description of an algorithm to compute the meta-rank and meta-diagram from a bifiltered simplicial complex, and give an example of how to visualize the meta-diagram. Section 6 discusses potential future directions of this approach to multiparameter persistence.