Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 5.1 — recognition and verification complexity

Proved
StableMSR.Recovery.Supplement.recognition_verification_complexity

by mikedeng1 · Nov 15, 2026 · Mathlib 0df444a (Lean v4.33.1)

clusteringp2o-batch-stoc27sp2o-gran-per-chapterp2o-plan-paperp2o-v1perturbation-stability

Fix a rational γ>1\gamma>1γ>1. On rational finite metric inputs, specified-partition verification is coNP-complete under both arbitrary symmetric and metric perturbations. Recognition is coNP-complete for arbitrary symmetric perturbations and coNP-hard for metric perturbations:

Verifyγsym, Verifyγmet, Recogγsym are coNP-complete,Recogγmet is coNP-hard.\mathrm{Verify}^{\mathrm{sym}}_\gamma,\ \mathrm{Verify}^{\mathrm{met}}_\gamma,\ \mathrm{Recog}^{\mathrm{sym}}_\gamma \text{ are coNP-complete},\qquad\mathrm{Recog}^{\mathrm{met}}_\gamma\text{ is coNP-hard}.Verifyγsym​, Verifyγmet​, Recogγsym​ are coNP-complete,Recogγmet​ is coNP-hard.

In all four cases the complement of exact MSR threshold decision reduces by polynomial-time many-one maps to the stability language. One map supplies a specified target and the associated instance projection supplies recognition. On every valid threshold input the supplied partition is the unique unperturbed optimum and has exactly one nonsingleton cluster.

The reductions hold on all binary strings, including malformed source strings. Perturbations may be real and centers may change. No NP-hardness premise or complexity-class separation is assumed. This is an independent complexity companion to the recovery goal.

Differences from the source:

  • encoding: Finite ordered metric input and unordered exact-k partitions → Fin n points, real metric arrays, and finite sets of nonempty disjoint blocks; reason: Input-order relabelling preserves every source instance and partition; base published definitions are reused without extra conditions
  • encoding: Rational-input recognition and verification with polynomial-time many-one reductions → Four explicit binary languages using the published PvsNP reductions and CSPIdealMem coNP notions; reason: Malformed strings are outside each language; the complements and reductions quantify over all strings
  • encoding: The reduction supplies the unique original optimum with exactly one nonsingleton cluster → A common specified-target map f and an instance map g, polynomial-time and satisfying all four equivalences, with the supplied-output clause on every valid threshold input; reason: The two codomains require different encodings; g is the instance projection on valid threshold data and all supplied properties remain conclusions
Preamble
import Definitions.Def_StableMSR_Recovery_Supplement_Languages

set_option autoImplicit false
open scoped Classical
Formal statement
namespace StableMSR.Recovery.Supplement

theorem recognition_verification_complexity (γ : ℚ) (hγ : 1 < γ) :
    CSPIdealMem.CoNPComplete (Verify false γ) ∧
    CSPIdealMem.CoNPComplete (Verify true γ) ∧
    CSPIdealMem.CoNPComplete (Recognize false γ) ∧
    CoNPHard (Recognize true γ) ∧ SuppliedReduction γ := by sorry

end StableMSR.Recovery.Supplement
Source
Shyam Ravichandran, Lev Reyzin, STOC 2027 submission #1010 (October 2026), Theorem 5.1, p. 12; PDF SHA256 8bb34b0b6d31458fe6fc0848c1fc82a6083f3e9ae672e748d24cf82d5289e53e

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me