Year: 2026
Type: preprint
Abstract: One of the best known invariants of multi-parameter persistent homology is the rank invariant, which is subsumed by the generalized rank invariant (GRI), obtained by naturally extending its domain. Möbius inversion of the GRI not only provides a compact encoding of this invariant but is also useful for clarifying its relationship with other invariants of multi-parameter persistent homology. However, Möbius inversion of the GRI has been defined only under suitable finiteness conditions on its domain and support. This work studies the extent to which Möbius inversion can be extended beyond these finiteness conditions. Namely, we formalize the notion of Möbius invertibility, which captures when Möbius inversion exists beyond these finiteness conditions, and investigate when the GRI is Möbius invertible. A central result we establish is that Möbius invertibility of the GRI is guaranteed for finitely presentable $d$-parameter persistence modules when $d\leq 2$, but may fail when $d>2$. Establishing this failure also corrects an error in the existing literature. In response to this higher-dimensional obstruction, we show that for finitely presentable $d$-parameter persistence modules, restricting the GRI to intervals subordinate to the finite grid supporting a minimal free presentation preserves its Möbius inversion whenever the latter exists over all intervals. Since computing the GRI over all intervals in $\mathbb{R}^d$ is infeasible in practice, this result provides a principled justification for restricting the GRI to those intervals.
Source: arXiv (Cornell University)
Language: English
Subfield: Computational Theory and Mathematics
Field: Computer Science
Domain: Physical Sciences
Keywords: Möbius inversion, persistence modules
Open Access status: green
Awards: G1221402884, Collaborative Research: AF: Small: Graph Analysis: Integrating Metric and Topological Perspectives, G2090981956, RI: Medium:Collaborative Research: Through synapses to spatial learning: a topological approach, RTG: Algebraic Topology and Its Applications, G4945737875, Collaborative Research: The Topology of Functional Data on Random Metric Spaces, Graphs, and Graphons, TRIPODS: Topology, Geometry, and Data Analysis (TGDA@OSU):Discovering Structure, Shape, and Dynamics in Data, Collaborative Research: Multiparameter Topological Data Analysis, G7132792363
OpenAlex ID: W7235269015
