Entropy-efficient exact singlet concentration
OpenQAlgorithms.Papers.Independent.Support50.P42.concentrationLet be a normalized pure state of two positive finite-dimensional systems shared by Alice and Bob. Let be its entropy of entanglement in bits. Local operations and classical communication admit protocols on every number of identical copies whose outputs are exact integer numbers of standard singlets. If is their expected singlet count, then there are constants and , depending only on the fixed input state, such that
If , all these yields are zero. The expected singlet count refers to actual local protocol leaves after standardization, rather than to a fractional logarithm of a Schmidt rank. This is the paper's asymptotic conservation of pure-state entanglement.
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P42
noncomputable section
namespace QAlgorithms.Papers.Independent.Support50.P42
open scoped BigOperators Matrix ComplexConjugate
open MeasureTheory ProbabilityTheory
attribute [local instance] Classical.propDecidable Classical.decEq
-- Reuse the exact accepted objects, without introducing same-name aliases.
open Information100.P21 (Op Vec FiniteLOCC)
open Dependency50.P42 (dimensionedPower singlets)
open Dependency50.P48 (pureFidelity)
open Gap48.P32 (pairOfState)
open Next46.P39 (rectAct)
/-- Abstract; PDF pp.4--9,15: exact singlet yield in expectation, for general
finite-dimensional pure inputs. Constants depend on the fixed input and precede
n. The log-rank yield and standardization loss are not identified with each other. -/
theorem concentration (s : Shared) (hpure : IsPureShared s) :
∃ Y : ℕ → ℝ, (∀ n, ExactYield s n (Y n)) ∧
(∃ C : ℝ, 0<C ∧ ∃ N : ℕ, ∀ n≥N,
(n : ℝ)*ent s-C*Real.logb 2 (n+1 : ℕ)≤Y n ∧ Y n≤(n : ℝ)*ent s) ∧
Filter.Tendsto (fun n : ℕ => Y n/(n : ℝ)) Filter.atTop (nhds (ent s)) ∧
(ent s=0 → ∀ n, Y n=0) := by sorry
end QAlgorithms.Papers.Independent.Support50.P42
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For every shared state , consisting of two positive finite local dimensions and a positive semidefinite, trace-one matrix on their joint basis, assume that for some unit vector . Write , where with applied by continuous functional calculus. There exists a sequence such that, for every , is an exact yield in the following specified sense: there is a shared state representing , with local dimensions under the fixed cardinality identifications, and a finite local-operation tree with output list , together with a function assigning a natural number to every shared state, such that every listed branch with represents exactly singlets and . Representation requires equality of local dimensions and equality of all matrix entries, with entries outside those dimensions extended by zero; a singlet is , and singlets means its -fold density-matrix tensor power, with local dimensions . The allowed trees are generated by stopping with weight ; applying a finite instrument at one party and recursively continuing separately on its outcomes; and replacing an output list of common local dimensions by its weighted average. Each instrument has nonnegative outcome weights summing to , leaves the other party’s dimension unchanged, and has local Kraus operators whose lifts satisfy and . No singlet-representation condition is imposed on branches of zero weight. Moreover, there exist and such that
and implies for every . The exact-yield assertion includes ; the displayed quotient at uses division by zero equal to , while the limit concerns .
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.