Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Cube arithmetic sum of one amplitude class of non-floor rows, from the source moment bound

Proved
RH.Transport.PhysicalHigh.actual_nonfloor_class_from_raw_moments

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

analytic-number-theorynumber-theoryqrh-lw56e2-b83dbbdtowards-grh

Throughout, O=Z[ζ3]\mathcal O=\mathbb Z[\zeta_3]O=Z[ζ3​], a Hecke character is a finite-order Hecke character of Q(ζ3)\mathbb Q(\zeta_3)Q(ζ3​) in OpenAI's family-003 development, and β\betaβ is the supremum of the real parts of the zeros of the associated Hecke LLL-functions (pole exception excluded, sentinel at 1/21/21/2). Objects named in typewriter font are those of OpenAI's development. The tuned constants are those of LW56e2's refinement, from the definition bundle Transport_TunedCandidate:

p=2146511250000000,b=1928211562500,ℓ=16+p,lx=1−ℓ−b2,ly=1−ℓ+b2,h=1+3ℓ+b2,ζ0=18⋅1012,p=\tfrac{214651}{1250000000},\qquad b=\tfrac{192821}{1562500},\qquad \ell=\tfrac16+p,\qquad l_x=\tfrac{1-\ell-b}2,\qquad l_y=\tfrac{1-\ell+b}2,\qquad h=\tfrac{1+3\ell+b}2,\qquad\zeta_0=\tfrac1{8\cdot10^{12}},p=1250000000214651​,b=1562500192821​,ℓ=61​+p,lx​=21−ℓ−b​,ly​=21−ℓ+b​,h=21+3ℓ+b​,ζ0​=8⋅10121​,

the tuned boundary B=1112−ℓ4=78−p4B=\tfrac{11}{12}-\tfrac\ell4=\tfrac78-\tfrac p4B=1211​−4ℓ​=87​−4p​, also written BpB_pBp​, and the signal exponent sig⁡(s)=s+lx2−1+h6=s−23−b6\operatorname{sig}(s)=s+\tfrac{l_x}2-1+\tfrac h6=s-\tfrac23-\tfrac b6sig(s)=s+2lx​​−1+6h​=s−32​−6b​. OpenAI's development has ℓ=16\ell=\tfrac16ℓ=61​ and b=18b=\tfrac18b=81​, hence lx=1748l_x=\tfrac{17}{48}lx​=4817​, ly=2348l_y=\tfrac{23}{48}ly​=4823​, boundary 78\tfrac7887​ and sig⁡(s)=s−1116\operatorname{sig}(s)=s-\tfrac{11}{16}sig(s)=s−1611​.

Setting. Let MMM be a nonzero ideal of O\mathcal OO and H≤(O/M)×H\le(\mathcal O/M)^\timesH≤(O/M)× a subgroup containing the image of the global units. Let N∈NN\in\mathbb NN∈N be the number of slots and let the real parameters e,eps,c,bW,A,R,dmin⁡,dmax⁡,rmin⁡,τ,ε,κ,cost,mesh,margin,losse,\mathrm{eps},c,b_W,A,R,d_{\min},d_{\max},r_{\min},\tau,\varepsilon,\kappa,\mathrm{cost},\mathrm{mesh},\mathrm{margin},\mathrm{loss}e,eps,c,bW​,A,R,dmin​,dmax​,rmin​,τ,ε,κ,cost,mesh,margin,loss satisfy the budget inequalities of the Lean statement, among them 0<e<110000<e<\tfrac1{1000}0<e<10001​, 0<dmin⁡≤dmax⁡0<d_{\min}\le d_{\max}0<dmin​≤dmax​, 8eR+κ≤ε<rmin⁡ mesh8eR+\kappa\le\varepsilon<r_{\min}\,\mathrm{mesh}8eR+κ≤ε<rmin​mesh, 2τ<dmin⁡ cost2\tau<d_{\min}\,\mathrm{cost}2τ<dmin​cost and τ(2+4 eps)<loss\tau(2+4\,\mathrm{eps})<\mathrm{loss}τ(2+4eps)<loss. Let SSS be a finite set of maximal ideals with the source exclusions and FirstTail (4e) S. Let ℓ1,…,ℓN\ell_1,\dots,\ell_Nℓ1​,…,ℓN​ be pairwise distinct slot lengths with dmax⁡rmin⁡≤ℓj≤dmin⁡Rd_{\max}r_{\min}\le\ell_j\le d_{\min}Rdmax​rmin​≤ℓj​≤dmin​R, and let WjW_jWj​ be smooth windows supported in (c,bW)(c,b_W)(c,bW​) with ∣Wj∣≤A|W_j|\le A∣Wj​∣≤A. Let φ≥0\varphi\ge0φ≥0 be a smooth, compactly supported, nonzero test function supported in (a0′,b0′)(a_0',b_0')(a0′​,b0′​) with a0′>0a_0'>0a0′​>0, and let count parameters for (M,H)(M,H)(M,H) at loss εm>0\varepsilon_m>0εm​>0 be given. Let total>0\mathrm{total}>0total>0 with ∑jℓj=total\sum_j\ell_j=\mathrm{total}∑j​ℓj​=total and 737dmax⁡≤total\tfrac7{37}d_{\max}\le\mathrm{total}377​dmax​≤total, let n∈Nn\in\mathbb Nn∈N, 0≤Δ≤180\le\Delta\le\tfrac180≤Δ≤81​, ν>0\nu>0ν>0, logCost>0\mathrm{logCost}>0logCost>0 and τ<heightCost\tau<\mathrm{heightCost}τ<heightCost.

Statement. There is C>0C>0C>0 such that for every Hecke character η\etaη and all sufficiently large ZZZ the following holds. Let d∈[dmin⁡,dmax⁡]d\in[d_{\min},d_{\max}]d∈[dmin​,dmax​], v≥0v\ge0v≥0, 51100<a≤1\tfrac{51}{100}<a\le110051​<a≤1, C0≥0C_0\ge0C0​≥0, and let rows be a finite set of rows u≠1u\ne1u=1, supported for the calibration of SSS, with

Z1/100≤∥u∥≤Zd−margin,Zv≤∥u∥≤2Zv.Z^{1/100}\le\lVert u\rVert\le Z^{d-\mathrm{margin}},\qquad Z^{v}\le\lVert u\rVert\le2Z^{v}.Z1/100≤∥u∥≤Zd−margin,Zv≤∥u∥≤2Zv.

Let i≤ni\le ni≤n and assume that for every row the detector maximum of its source detector family is at least aaa at height 3iZτ3iZ^{\tau}3iZτ and below a+2ea+2ea+2e at height 3(i+1)Zτ3(i+1)Z^{\tau}3(i+1)Zτ. Let ttt be a height vector with all three coordinates at most (3i+1)Zτ(3i+1)Z^{\tau}(3i+1)Zτ in absolute value, put z=1750+it2z=\tfrac{17}{50}+it_2z=5017​+it2​, and assume the source moment bound SourceMomentsAt for these rows at conductor exponent ddd, level aaa, external point zzz, parameter Δ\DeltaΔ, constant C0ZmomentCostC_0Z^{\mathrm{momentCost}}C0​ZmomentCost and height ZheightCostZ^{\mathrm{heightCost}}ZheightCost. Then for every amplitude bin of the physical row amplitude at scale ZdZ^{d}Zd, with class mean qqq and nonempty class rows′\mathrm{rows}'rows′, one has 0≤q≤2a−120\le q\le\tfrac{2a-1}20≤q≤22a−1​ and

∥cubeArithmeticSum(η;rows′;a,e,t)∥≤C C0 N(fη)2 eps ZE,\big\lVert\texttt{cubeArithmeticSum}(\eta;\mathrm{rows}';a,e,t)\big\rVert\le C\,C_0\,N(\mathfrak f_\eta)^{2\,\mathrm{eps}}\,Z^{E},​cubeArithmeticSum(η;rows′;a,e,t)​≤CC0​N(fη​)2epsZE, E=logCost+heightCost+momentCost+d⋅adaptiveRowExponent(2a−1,q,Δ,ε,εm,R,ν)+v(a−12+12e+eps(N+8)−1750)+loss−425total+q total+mesh⋅total.E=\mathrm{logCost}+\mathrm{heightCost}+\mathrm{momentCost}+d\cdot\texttt{adaptiveRowExponent}(2a-1,q,\Delta,\varepsilon,\varepsilon_m,R,\nu)+v\big(a-\tfrac12+12e+\mathrm{eps}(N+8)-\tfrac{17}{50}\big)+\mathrm{loss}-\tfrac4{25}\mathrm{total}+q\,\mathrm{total}+\mathrm{mesh}\cdot\mathrm{total}.E=logCost+heightCost+momentCost+d⋅adaptiveRowExponent(2a−1,q,Δ,ε,εm​,R,ν)+v(a−21​+12e+eps(N+8)−5017​)+loss−254​total+qtotal+mesh⋅total.

The slot pools are pool(identity ray class,S,c,bW,Zℓj)\mathrm{pool}(\text{identity ray class},S,c,b_W,Z^{\ell_j})pool(identity ray class,S,c,bW​,Zℓj​).

The rows of a dyadic block are sorted by the size of their physical amplitude. Within one class the moment bound controls the number of rows and the class mean qqq controls the slot product, which gives the exponent EEE. The total slot length is a parameter here; OpenAI's argument fixes it at 16\tfrac1661​.

Formalization Note The definitions are OpenAI's, from the definition bundles of the OpenAI mission on this platform, including SourceMomentsAt, amplitudeRows, classMean and adaptiveRowExponent. The namespace opens of the source file that are not available from the definition bundles were omitted, the local finiteness instance was given a name, and spaces were inserted after some < and ≤ signs so that the statement tokenizes under import Mathlib.

Preamble
import Mathlib
import Definitions.Def_HeckeSevenEighths
import Definitions.Def_OAIHecke78B008
import Definitions.Def_OAIHecke78B016
import Definitions.Def_OAIHecke78B019
import Definitions.Def_OAIHecke78B021
import Definitions.Def_OAIHecke78B022
import Definitions.Def_OAIHecke78B023
import Definitions.Def_OAIHecke78B024
import Definitions.Def_OAIHecke78B026
import Definitions.Def_OAIHecke78B027
import Definitions.Def_OAIHecke78B028
import Definitions.Def_Transport_Geometry
import Definitions.Def_Transport_LocalBounds
import Definitions.Def_Transport_SupportedMass
import Definitions.Def_Transport_TunedCandidate
set_option autoImplicit false
Formal statement
namespace RH.Transport.PhysicalHigh

noncomputable section

open scoped Classical BigOperators Topology ContDiff

open Filter Set

open OAI OAI.SevenEighths

open ProbeHighRowFamily

open HeckeFamily HeckeInverseAmplification ProbePhysical ProbeMellinBoundary

open ProbeRaySlots HeckeDetectorPhysicalSelection HeckeDetectorAmplitudeFirst HeckeDetectorFiberPartition

local notation "O" => HeckeFamily.O

variable (M : Ideal O) [NeZero M]

local instance instFiniteQuotientOIdeal_transport_actual_nonfloor_class_from_raw_moments : Finite (O ⧸ M) := Ring.HasFiniteQuotients.finiteQuotient (NeZero.ne M)

variable (H : Subgroup (O ⧸ M)ˣ) (hH : RayOrthogonality.globalUnits M≤ H)

theorem actual_nonfloor_class_from_raw_moments (total : ℝ) (htotal : 0 < total) (N n : ℕ) (e eps c b A R dmin dmax rmin τ ε κ cost mesh margin loss : ℝ)
    (he : 0< e) (he1 : e<1/1000) (heps : 0< eps) (hc : 0< c) (hcb : c≤ b) (hA : 0≤ A)
    (hR : 0≤ R) (hdmin : 0< dmin) (hdmax : 0≤ dmax) (hdRange : dmin≤ dmax) (hrmin : 0< rmin)
    (hτ : 0< τ) (hε : 0< ε) (hκ : 0< κ) (hcost : 0≤ cost) (hmesh : 0< mesh)
    (hbudget : 8*e*R+κ≤ ε) (hgap : ε< rmin*mesh) (hmargin : 0< margin)
    (hheight : 2*τ< dmin*cost) (hloss : τ*(2+4*eps)< loss)
    (S : Finset (Ideal O)) (hS : SourceExclusions S) (hfirst : FirstTail (4*e) S)
    (hmax : ∀P∈S,P.IsMaximal)
    (ell : Fin N→ℝ) (hell : Function.Injective ell)
    (hello : ∀j,dmax*rmin≤ ell j) (hellhi : ∀j,ell j≤ dmin*R)
    (W : Fin N→ℝ→ℂ)
    (hWs : ∀j,Function.support (W j)⊆Ioo c b) (hW : ∀j,ContDiff ℝ ∞ (W j)) (hWB : ∀j t,‖W j t‖≤ A)
    (hellsum : ∑j,ell j=total)
    (hsupply : (7/37:ℝ)*dmax≤ total) (hε1 : ε≤1/1000) (hκ1 : κ≤1)
    (hτzero : τ< dmin/2) (hτheight : 4*τ< dmin*cost)
    (hwbudget : 12*e*((22:ℝ)+2)+8*κ+2*cost≤ ε/2)
    (φ : ℝ→ℝ) (hφ : ContDiff ℝ ∞ φ) (hφc : HasCompactSupport φ)
    (hφp : tsupport φ⊆Ioi 0) (hφ0 : ∀y,0≤ φ y) (hφne : φ≠0)
    (a₀ b₀ B₀ : ℝ) (ha₀ : 0< a₀) (hab₀ : a₀≤ b₀) (hB₀ : 0< B₀)
    (hφs : Function.support φ⊆Ioo a₀ b₀) (hφB : ∀y,φ y≤ B₀)
    (εm Δ ν logCost heightCost momentCost : ℝ)
    (hεm : 0< εm) (hΔ : 0≤ Δ) (hΔ1 : Δ≤1/8) (hν : 0< ν)
    (hlog : 0< logCost) (hMomentHeight : τ< heightCost)
    (counts : ProbeFinalAssembly.CountParameters M H εm) :
    ∃C : ℝ,0< C ∧
    ∀η : Character,∀ᶠZ : ℝ in atTop,
      ∀d : ℝ,dmin≤ d → d≤ dmax → ∀(v a C0 : ℝ),0≤ v → 51/100< a → a≤1 → 0≤ C0 →
      ∀rows : Finset FreeRow,
      (∀u∈rows,u.val≠1 ∧ Z^(1/100:ℝ)≤ rowNorm u ∧
        (calibrationForSet S hmax).residueMonoid u.val≠0 ∧ rowNorm u≤ Z^(d-margin)) →
      (∀u∈rows,Z^v≤ rowNorm u ∧ rowNorm u≤2*Z^v) →
      ∀i : ℕ,i≤ n →
      (∀u∈rows,detectorMaximum (sourceDetectorFamily S hS.prime η u (rayCubeFamily M H hH u))
        (3*(i+1:ℕ)*Z^τ)< a+2*e) →
      (∀u∈rows,a≤ detectorMaximum (sourceDetectorFamily S hS.prime η u (rayCubeFamily M H hH u))
        ((3*i:ℕ)*Z^τ)) →
      let Y : Fin N→ℝ := fun j=>Z^(ell j)
      let T : Fin N→Finset ProbePhysical.PrimeIdeal := fun j=>pool (RayQuotient.identityClass M H) S c b (Y j)
      ∀t : HeightSpace,((|t.1.1|≤(3*i+1:ℕ)*Z^τ ∧ |t.2|≤(3*i+1:ℕ)*Z^τ) ∧ |t.1.2|≤(3*i+1:ℕ)*Z^τ) →
      let z : ℂ := (17/50:ℂ)+t.1.2*Complex.I
      SourceMomentsAt M H hH S hS.prime η rows ell W Z d a ε τ dmax b R mesh i z Δ
        (if 2*a-1≤5/6 then counts.cB else counts.cH) (if 2*a-1≤5/6 then counts.kB else counts.kH)
        (C0*Z^momentCost) (Z^heightCost) εm →
      let Q := physical M H (fun u : FreeRow=>u.val) W (fun _=>b) (fun j=>ell j/d) (fun _=>z) (Z^d)
      ∀bin : BinLabel (Finset.univ : Finset (Fin N)) ((2*a-1)/2) mesh,
      let rows' := amplitudeRows rows Finset.univ (Z^d) ((2*a-1)/2) mesh hmesh (fun j=>ell j/d) Q bin
      let q := classMean Finset.univ ((2*a-1)/2) mesh (fun j=>ell j/d) bin
      rows'.Nonempty → 0≤ q ∧ q≤(2*a-1)/2 ∧
      ‖cubeArithmeticSum S hS hmax η rows' T (nonfloorPoolOutside M H S N c b Y) W Y a e t‖≤
        C*C0*(η.modulus.absNorm:ℝ)^(2*eps)*
          Z^(logCost+heightCost+momentCost+
            d*adaptiveRowExponent (2*a-1) q Δ ε εm R ν+
            v*(a-1/2+12*e+eps*(N+8)-17/50)+loss-(4/25)*total+q*total+mesh*total) := by sorry

end

end RH.Transport.PhysicalHigh
Source
LW56e2, quasi-riemann-improved, commit b83dbbd914c5751acf9bc3389cde58970437278f, lean/Transport/PhysicalHigh/NonfloorClass.lean, theorem actual_nonfloor_class_from_raw_moments (L37-L89): https://github.com/LW56e2/quasi-riemann-improved/blob/b83dbbd914c5751acf9bc3389cde58970437278f/lean/Transport/PhysicalHigh/NonfloorClass.lean#L37-L89

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