Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Charged parameter preparation and sparse recursive-root initialization

Proved
NearlyOptimalPPR.Computation.recursive_preparation_exec

by Shuze Chen · Nov 19, 2026 · Mathlib 0df444a (Lean v4.33.1)

Fix an integer amplification constant B≥1B\ge1B≥1 and an input with 0<ε≤10<\varepsilon\le10<ε≤1, represented in the constant-size input layout with zero boundary counter. The recursive preparation routine has a finite execution. It computes the accessible depth ℓ\ellℓ, root threshold qqq, scale NNN, and binary rank rrr, satisfying

8≤Nq,N≤8/q+1,0<N<2r,r≤⌊log⁡2N⌋+1.8\le Nq,\qquad N\le8/q+1,\qquad 0<N<2^r,\qquad r\le\lfloor\log_2N\rfloor+1.8≤Nq,N≤8/q+1,0<N<2r,r≤⌊log2​N⌋+1.

It initializes a complete sparse representation of the source root state with queue mass at most one, amplified rank-array length, zero solve epoch, and bounded epoch marks. All work equals its instruction count, bounded by the explicit charged search, header, and sparse-initialization formula in the interface. It performs exactly one graph query and preserves existing output.

The result charges all scalar parameter searches and initializes only the seed-dependent sparse state.

Preamble
import Definitions.Def_NearlyOptimalPPR_Computation_compiler_runtime
open Finset Matrix
open scoped ENNReal
set_option autoImplicit false
noncomputable section
attribute [local instance] Classical.propDecidable
Formal statement
namespace NearlyOptimalPPR.Computation
open SparseQueue
theorem recursive_preparation_exec (m : SolveModel) {n : ℕ} (g : InputGraph n) (ω : ℕ → Bool)
    (B : ℕ) (hB : 1 ≤ B) (s : Fin n) (α ε : ℝ) (hε : 0<ε) (hε1 : ε ≤ 1)
    (st : MachineState) (hh : st.scalar.halted=false) (input : PushInputMemory s α ε st)
    (hBoundary : st.scalar.integerMemory 7=0) :
    let ell := accessibleDepth (1/ε)
    let q := rootThreshold ε ell
    ∃ (N r steps : ℕ) (out : MachineState), ∃ memory : RecursiveMachineMemory g s α ε ell (amplifiedQueueRank B r+1)
      (recursiveInit g s q ell []) out,
      MachineEval m g ω (recursivePreparationCode B) st steps out ∧
      8 ≤ (N : ℝ)*q ∧ (N : ℝ) ≤ 8/q+1 ∧ 0<N ∧ N < 2^r ∧ r ≤ Nat.log 2 N+1 ∧
      steps ≤ (12*(ell+2)^3+4*N+4*r+24)+(B+14)+
        (12*(16*(ell+3))+64+(⌈1/q⌉₊+2)*(4*ell+22)+10*(amplifiedQueueRank B r+1)+264+
          (21*(amplifiedQueueRank B r)+60*(16*(ell+3))+190)) ∧
      memory.queueMass ≤ 1 ∧ out.scalar.core.ptrs 150=0 ∧ out.scalar.core.ptrs 151=1 ∧
      out.scalar.core.ptrs 53=0 ∧ EpochInvariant (n:=n) (16*(ell+3)) out ∧
      out.work=st.work+steps ∧ out.scalar.queries=st.scalar.queries+1 ∧ out.output=st.output ∧ out.scalar.halted=false := by sorry
end NearlyOptimalPPR.Computation
Source
Li, Xu, Yang, Zhou, Nearly optimal Personalized PageRank computation, STOC 2027 submission #884, https://github.com/sigact-stoc/stoc2027/releases/download/papers-1/884.pdf, Algorithms 1 and 6–9, pp. 5,14–22; charged local-machine realization of the mission computational model.

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