Theorem 1 (Variance of partial derivatives)
OpenQAlgorithms.Papers.Independent.Support50.P30.partialDerivative_mean_varianceLet and fix an orthonormal Schur decomposition of the -qubit space, independently of the data and parameters. Its sectors have block dimension and multiplicity , with the predecessor zero at . Let be a fixed layered circuit with generators from Eq. (9), let belong to Eq. (10), and fix a differentiated layer . For density matrices and arbitrary real coefficients , put and . Under a probability law on the parameters, assume the two halves , with including layer , have joint conjugation moments through order two equal to the product Haar moments over both halves and all sectors. Then
Here is an unnormalized compression in the fixed common coordinates, and . One-dimensional sectors contribute zero. This exact identity separates the circuit, measurement and dataset factors; it makes no inverse-polynomial lower-bound claim.
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P30
noncomputable section
open MeasureTheory
open scoped BigOperators ComplexConjugate Matrix.Norms.L2Operator NNReal ENNReal
namespace QAlgorithms.Papers.Independent.Support50.P30
/-- PDF p.6, Theorem 1, Eq.(16), with proof assumptions on pp.14,23-26.
One-dimensional sectors have zero commutators and are omitted explicitly. -/
theorem partialDerivative_mean_variance {n L M : ℕ} (hn : 2 ≤ n)
(S : SchurFrame n) (word : Fin L → Fin 3) (mu : Fin L)
(O : QnnOperator n) (hO : O ∈ qnnMeasurementFamily n) (hHerm : O.IsHermitian)
(rho : Fin M → QnnDensity n) (c : Fin M → ℝ)
(P : Measure (Fin L → ℝ)) [IsProbabilityMeasure P]
(hdesign : IndependentBlockTwoDesign S (fun j => qnnGenerator n (word j)) mu P) :
(∫ theta, partialLoss (fun j => qnnGenerator n (word j)) mu c rho O theta ∂P) = 0 ∧
variance P (partialLoss (fun j => qnnGenerator n (word j)) mu c rho O) =
∑ m : Sector n, if 1 < blockDim n m then
(2 * (blockDim n m : ℝ) / (((blockDim n m : ℝ) ^ 2 - 1) ^ 2)) *
centeredDelta (commonBlock S m (qnnGenerator n (word mu))) * centeredDelta (commonBlock S m O) *
centeredDelta (aggregate S m (dataOperator c rho)) else 0 := by sorry
end QAlgorithms.Papers.Independent.Support50.P30
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For any natural numbers with , let be the length- bit strings, let sectors be , and put , , and for . Let be a Schur frame with matrix , whose rows are , , , and whose columns are in : and ; every matrix commuting with all bit-coordinate permutations becomes blocks repeated over , with zero entries between different ; and becomes this pattern with block , where are the usual Pauli matrices and a subscript specifies the coordinate on which they act. For a Pauli matrix , define , where if , and otherwise, and define , , and . Choose any list , an index , and a Hermitian complex matrix on for which there exists such that , or , or . Choose any density matrices on , meaning positive semidefinite matrices of trace , and arbitrary real coefficients , for , and any probability measure on . Write
where replaces just coordinate by . Set , , and . Assume the following joint moment identities: with , , and , every choice and lists , independently for all keys , satisfies
where is normalized Haar measure, empty scalar products are , and empty matrix products are identities. Then, defining , both conclusions hold:
and
In particular, one-dimensional sectors contribute zero through the stated branch. The index makes the hypotheses unavailable for ; is allowed, with empty data and loss sums equal to zero.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.