Abstract main result (p.1), pinned to Theorem 7.1
OpenQAlgorithms.Papers.Independent.Support50.P12.theorem_7_1Let and be concrete unital algebras on nonzero finite-dimensional complex Hilbert spaces, and let be complex linear. The adjoint is for . A substochastic map has a Schwarz-contracting adjoint; stochastic additionally means trace preserving. Let . Powers are zero on kernels and traces are not normalized. Write for the unnormalized Petz Rényi entropy, , , and . Logarithms are natural and . The Petz map is . Let be a nonempty set of density operators in . Write for its closed convex hull, for its joined support, and .
If is trace preserving and two-positive, the following are equivalent: (i) one stochastic recovers every state in the closed convex hull; (ii) every operator convex preserves every pair's divergence on that hull; (iii) one reference density with support covering and one operator convex with preserve for all ; (iv) one such reference and one , , preserve all for every code state, where ; (v) all preserve every pair on the hull for every real ; (vi) one covering reference and one give preservation for all code states and all ; (vii) every hull state is recovered by for every hull reference with support exactly ; (viii) one density reference has that recovery property for all code states; (ix) the input/output code supports have shared orthogonal tensor-block decompositions with invertible right density operators and left unitaries , such that every code state is and
The same recovery can retain each given -positivity level of , or complete positivity. For merely substochastic , still holds. Finally, (x) one covering density reference gives
where , is sufficient for (i)–(viii), and necessary when is completely positive. References, generators, parameter sets and block data are chosen before all states being tested. This is the full error-correction equivalence.
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P12
noncomputable section
namespace QAlgorithms.Papers.Independent.Support50.P12
open Filter MeasureTheory Set
open QAlgorithms.Papers.Independent.Information100.P60 (Mat copies)
open QAlgorithms.Papers.Independent.Information100.P30 (MatrixMap)
open QAlgorithms.Papers.Independent.Dependency50.P28 (tr LE logNonnegativeExtended pinch)
open QAlgorithms.Papers.Independent.Dependency50.P21 (channelAdjoint spectralFunction petz relativeEntropy)
open QAlgorithms.Papers.Independent.Frontier18.P17 (hs)
open QAlgorithms.Papers.Independent.Next46.P23 (Supports)
open QAlgorithms.Papers.Independent.Support21.P03 (ExtendedConvex)
open QAlgorithms (spectralProjector)
open QAlgorithms.Learning (IsDistribution)
open QAlgorithms.Papers.Independent.Information100.P04 (absOp)
open scoped BigOperators Matrix ComplexConjugate ComplexOrder ENNReal Topology Kronecker
attribute [local instance] Classical.propDecidable Classical.decEq
open QAlgorithms.Papers.Independent.Information100.P71 (supportPower)
variable {I J K : Type} [Fintype I] [DecidableEq I] [Fintype J] [DecidableEq J]
[Fintype K] [DecidableEq K]
/-- Theorem 7.1, PDF pp.33–34 (abstract p.1): all nine equivalences, all uniform
choices, both positivity refinements, the substochastic chain, and criterion (x).
The bound is dim(H1)^2+dim(H2)^2. -/
theorem theorem_7_1 [Nonempty I] [Nonempty J] (A : Alg I) (B : Alg J) (Phi : MatrixMap I J)
(hc : AlgebraExtension A B Phi) (C : Set (Mat I)) (hn : C.Nonempty) (hC : ∀ rho ∈ C, AlgebraDensity A rho) :
let c := codeConditions A B Phi C
((AlgebraTracePreserving A Phi ∧ AlgebraKPositive A Phi 2) →
(∀ i : Fin 9, c (Fin.castSucc i) ↔ c 0) ∧
(c 0 → RecoveryRefinement A B Phi (hull C)) ∧
(c 9 → ∀ i : Fin 8, c (Fin.castLE (by omega : 8 ≤ 10) i)) ∧
(AlgebraCompletelyPositive A Phi → (c 0 ↔ c 9))) ∧
(Substochastic A B Phi → (c 0 → c 1) ∧ (c 1 → c 2) ∧ (c 2 → c 3) ∧ (c 3 → c 7)) := by sorry
end QAlgorithms.Papers.Independent.Support50.P12
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
Let be finite nonempty index types with decidable equality, let , be unital complex star subalgebras, and let be complex linear, with for every and for every , where is the Hilbert–Schmidt adjoint. Let be a nonempty set of positive semidefinite matrices in , each of real trace , and put . Use for positive semidefiniteness of , , and , with . Say that is supported by when , and that covers when this holds for every . For a real function , let be its distinct real eigenvalues if is Hermitian, and the empty set otherwise; let be the orthogonal projection onto the -eigenspace. Define
Here takes values in the extended reals, its displayed limit is the totalized limit used by the definition, , and for and otherwise. Call a generator when it is continuous on and , regarded as extended real, converges as to its totalized limit; call it operator convex here when, for every natural matrix size, positive semidefinite , and , by functional calculus. To specify the richness condition below, replace by its totalized right limit to obtain , and select by Hilbert choice a measure on admitting some representation
with , integrable , and an integrable displayed kernel for every ; the selection is arbitrary if no such measure exists. Say that is rich to order when in cardinality. A map is stochastic from one concrete algebra to another when both it and its adjoint have their entire ranges in the respective target and source algebras, its adjoint is positive on the target algebra and satisfies the Schwarz inequality with constant , and preserves the trace on the source algebra; “substochastic” omits that last requirement. A map is -positive on a concrete algebra when its entrywise amplification preserves positive semidefiniteness of block matrices whose blocks belong to that algebra; complete positivity on the algebra means this for every natural , including . Define the following ten conditions: : some stochastic , represented on the full matrix spaces, satisfies for every ; : for every generator that is operator convex in the stated sense and every ; : there exist a density matrix covering and such a generator , rich to order , for which that equality holds for every against ; : there exist a density matrix covering and , with , such that for every , where ; : for every ; : some density matrix covering and some give this equality for every and ; : whenever and
: some density matrix satisfies for every , without an additional covering requirement; : there exist finitely many blocks, possibly zero blocks, with positive integer dimensions , isometries and , orthogonal ranges within each family, positive definite trace-one matrices on the right factors, and two-sided unitaries , such that has kernel , has kernel , every can be written
with and positive semidefinite trace-one , and, for every block and every complex matrix of size , the input block belongs to and
: some density matrix covering satisfies, for every , , and natural ,
Here tensor powers have entries given by products of the component entries, the tensor map is the linear extension of its action on elementary tensor matrices, , and uses the empty product. The theorem asserts that, if preserves traces on and is -positive on , then for every ; implies that, for every natural for which is -positive on , there is a stochastic recovery map that is -positive on and recovers every state in , and complete positivity of on gives such a completely positive recovery map; implies every for ; and complete positivity of on additionally gives . Separately, if is substochastic, then , , , and .
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.