Induced-matching stability above a threshold: interleaved tame 0-dim filtrations admit an eps-close injection on all bars with margins
ProvedPersistClust.Count.deep_injection_matchingbauer-lesnickcounthallinduced-matchinginterleavingmatchingpersistclustpersistencestabilitysupport-tool
Let X and Y be tame 0-dimensional persistence filtrations (FiltrationLaw + IsTameModule) that are strongly eps-interleaved above alpha (StrongInterleavingAbove). Fix kappa > 0. Then the multiplicity copies of X at real bars (b, d) with margins alpha + kappa <= d and d + 2 eps + kappa <= b admit an INJECTIVE map F into the multiplicity copies of Y such that every copy is sent to a copy eps-close in both coordinates (closeE in birth and in death). This is the hard direction of algebraic stability (the matching), localized above the threshold, for bars bounded away from alpha and from the diagonal.
Preamble
import Mathlib import Definitions.Def_PersistClust_Count_Setting import Definitions.Def_PersistClust_Count_Diagram import Definitions.Def_PersistClust_Count_FiltrationLaw import Definitions.Def_PersistClust_Count_MatchingCensus import Definitions.Def_PersistClust_Count_MatchingGate open PersistClust.Count open scoped ENNReal
Formal statement
namespace PersistClust.Count
theorem deep_injection_matching {ιX ιY : Type*} {stageX : ℝ → Set ιX} {JX : ℝ → ιX → ιX → Prop}
{stageY : ℝ → Set ιY} {JY : ℝ → ιY → ιY → Prop}
(hLawX : FiltrationLaw stageX JX) (hLawY : FiltrationLaw stageY JY)
(hTameX : IsTameModule stageX JX) (hTameY : IsTameModule stageY JY) {α ε : ℝ}
(hε : 0 ≤ ε) (hIL : StrongInterleavingAbove stageX JX stageY JY hLawX hLawY α ε hε)
{κ : ℝ} (hκ : 0 < κ) :
∃ F : {x : offD (mult (rankFn stageX JX)) // ∃ b d : ℝ,
offPt x = (((b : ℝ) : EReal), ((d : ℝ) : EReal)) ∧ α + κ ≤ d ∧ d + 2 * ε + κ ≤ b} →
offD (mult (rankFn stageY JY)),
Function.Injective F ∧ ∀ x, closeE (offPt x.1).1 (offPt (F x)).1 ε ∧
closeE (offPt x.1).2 (offPt (F x)).2 ε
:= by sorrySource
Support tool of the persistence-clustering deliverable (RR-6968, Theorem 4.5 / 4.8 cluster): the generalized box lemma for overlapping eps-boxes, proved via the Bauer-Lesnick induced matching (Induced Matchings and the Algebraic Stability of Persistence Barcodes) adapted to component filtrations. It closes the census atoms of the deliverable's matching gate (the demand routing and the rank-kill leaf). Mirrors from Def_PersistClust_Count_MatchingCensus.