Cube arithmetic sum of one amplitude class of non-floor rows, from the source moment bound
ProvedRH.Transport.PhysicalHigh.actual_nonfloor_class_from_raw_momentsThroughout, , a Hecke character is a finite-order Hecke character of in OpenAI's family-003 development, and is the supremum of the real parts of the zeros of the associated Hecke -functions (pole exception excluded, sentinel at ). 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:
the tuned boundary , also written , and the signal exponent . OpenAI's development has and , hence , , boundary and .
Setting. Let be a nonzero ideal of and a subgroup containing the image of the global units. Let be the number of slots and let the real parameters satisfy the budget inequalities of the Lean statement, among them , , , and . Let be a finite set of maximal ideals with the source exclusions and FirstTail (4e) S. Let be pairwise distinct slot lengths with , and let be smooth windows supported in with . Let be a smooth, compactly supported, nonzero test function supported in with , and let count parameters for at loss be given. Let with and , let , , , and .
Statement. There is such that for every Hecke character and all sufficiently large the following holds. Let , , , , and let rows be a finite set of rows , supported for the calibration of , with
Let and assume that for every row the detector maximum of its source detector family is at least at height and below at height . Let be a height vector with all three coordinates at most in absolute value, put , and assume the source moment bound SourceMomentsAt for these rows at conductor exponent , level , external point , parameter , constant and height . Then for every amplitude bin of the physical row amplitude at scale , with class mean and nonempty class , one has and
The slot pools are .
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 controls the slot product, which gives the exponent . The total slot length is a parameter here; OpenAI's argument fixes it at .
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.
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
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