§4.2.4 (4.40)–(4.43): max-divergence pinching attainment
OpenQAlgorithms.Papers.Independent.Intakenext100.P07.max_divergence_pinchingAll quantum systems below are finite-dimensional complex Hilbert spaces of positive dimension, and logarithms are binary. A normalized state is positive semidefinite with trace one; a subnormalized state has trace at most one. Marginals are partial traces. Tildes denote sandwiched Rényi quantities and bars denote Petz quantities. Their trace functionals are and ; the divergences include division by inside the logarithm. Matrix powers vanish on the kernel, including power zero. The notation means support inclusion.
D∞(Pσρ||σ)≤D∞(ρ||σ)≤D∞(Pσρ||σ)+log|specσ|. After tensor powers and division by n the correction vanishes because |spec(σ^⊗n)|≤(n+1)^(d−1); obtain the limit representation.
This result records the source’s operational or structural relation, with the stated endpoint and attainment clauses retained.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P07
noncomputable section
open Filter
open scoped BigOperators Matrix ComplexConjugate ComplexOrder Matrix.Norms.L2Operator Topology
namespace QAlgorithms.Papers.Independent.Intakenext100.P07
attribute [local instance] Classical.propDecidable Classical.decEq
variable {I J A B C D X Y Z : Type*}
variable [Fintype I] [DecidableEq I] [Fintype J] [DecidableEq J]
variable [Fintype A] [DecidableEq A] [Fintype B] [DecidableEq B]
variable [Fintype C] [DecidableEq C] [Fintype D] [DecidableEq D]
variable [Fintype X] [DecidableEq X] [Fintype Y] [DecidableEq Y]
variable [Fintype Z] [DecidableEq Z]
/-- NODE paper:1504.00233#4.maxpinching
§4.2.4 (4.40)–(4.43): max-divergence pinching attainment, arXiv:1504.00233v5, PDF p.60,61. -/
theorem max_divergence_pinching [Nonempty I] (r s : paperMat I) (hr : r.PosSemidef)
(hne : r ≠ 0) (hs : s.PosSemidef) :
maxD (paperPinch s r) s ≤ maxD r s ∧ maxD r s ≤ maxD (paperPinch s r) s+(Real.logb 2 (spectralValues s).card:EReal) ∧
(∀ n : ℕ, 0 < n → (spectralValues (paperTensorPower s n)).card ≤ (n+1)^(Fintype.card I-1)) ∧
Tendsto (fun n : ℕ => Real.logb 2 (spectralValues (paperTensorPower s n)).card/(n:ℝ)) atTop (nhds 0) ∧
Tendsto (fun n : ℕ => ((1/(n:ℝ):ℝ):EReal)*maxD (paperTensorPower r n) (paperTensorPower s n)) atTop (nhds (maxD r s)) ∧
Tendsto (fun n : ℕ => ((1/(n:ℝ):ℝ):EReal)*maxD (paperPinch (paperTensorPower s n) (paperTensorPower r n)) (paperTensorPower s n))
atTop (nhds (maxD r s)) := by sorry
end QAlgorithms.Papers.Independent.Intakenext100.P07
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
Let be a finite, nonempty index type with decidable equality, and let be complex matrices indexed by , where is positive semidefinite and nonzero and is positive semidefinite. Define the extended-real quantity
where means that is positive semidefinite and the infimum is taken in the extended real line, so an empty feasible set has infimum . Let be the set of distinct eigenvalues of a Hermitian matrix , including zero if present, and let , where is functional calculus of the indicator of applied to . For , define , indexed by functions , by . Then all of the following hold:
The last two limits are in the extended real line. Neither trace normalization nor support inclusion is assumed, and may be zero. The sequences are defined also at : the tensor power is the one-dimensional matrix with entry , and real division by zero is assigned value zero; the eigenvalue-count bound itself is asserted only for .
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.