Single-letter lower bound on the strong-converse exponent
OpenQAlgorithms.Papers.Independent.Intakenext100.P15.singleletter_success_boundAll Hilbert spaces are finite-dimensional and nonzero. A density operator is positive semidefinite with trace one; singular states are included. All logarithms are base two. For real , denotes the sandwiched Rényi divergence and the traditional Petz divergence, with value on support failure and negative spectral powers taken on the support.
Write , where is unnormalized, and . Its reference marginal is and its channel-input marginal is . Define over all normalized densities. The finite attained extrema are retained as conclusions in the minimax statement.
An -use parallel entanglement-assisted classical code has , an integer of uniform messages, an arbitrary finite message-independent shared bipartite density , message-dependent local CPTP encoders , and a joint POVM on . Its rate is , and its average success probability is .
For every real rate lower bound ,
The supremum covers the entire order interval . Exponentiation uses extended endpoints and . This finite-block inequality supplies a lower bound on the decay exponent of every code; it does not identify an optimal exponent.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P15
noncomputable section
open scoped Matrix ComplexOrder ComplexConjugate Topology
open Filter
namespace QAlgorithms.Papers.Independent.Intakenext100.P15
attribute [local instance] Classical.propDecidable
attribute [local instance] Classical.decEq
-- These are pinned accepted definitions, not other chunks from the current run.
open QAlgorithms.Papers.Independent.Bridge50.P43 (NormalizedDensity payoff)
open QAlgorithms.Papers.Independent.Dependency50.P09 (Divergence logExt sandwichedD petzD)
open QAlgorithms.Papers.Independent.Information100.P34 (choi)
open QAlgorithms.Papers.Independent.Information100.P63 (schatten)
open QAlgorithms.Papers.Independent.Information100.P71
(supportPower SupportIncluded stateRelative channelPower diamond)
/-- PDF pp.6,13, (2.20),(6.1). An upper bound on parallelEASuccess of each code, rather than
an assertion identifying the optimal strong-converse exponent. -/
theorem singleletter_success_bound
{A B : Type} [Fintype A] [Fintype B] [DecidableEq A] [DecidableEq B]
[Nonempty A] [Nonempty B] (N : Superop A B) (hN : IsChannel N)
(n : ℕ) (hn : 1 ≤ n) (C : ParallelEACode A B n) (R : ℝ) (hR : R ≤ parallelEARate C) :
ENNReal.ofReal (parallelEASuccess N C) ≤ expBits
(-((n : ℝ) : EReal) *
⨆ alpha : {a : ℝ // 1 < a}, payoff alpha.1 R (channelInfo (sandwichedD alpha.1) N)) := by sorry
end QAlgorithms.Papers.Independent.Intakenext100.P15
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For nonempty finite types with decidable equality, let for a finite family with , and let be natural. A given code consists of a positive message count , positive auxiliary dimensions , a positive semidefinite trace-one matrix on , a finite-Kraus channel for each message , and positive semidefinite matrices on summing to the identity. Each encoding channel’s Kraus completeness sum is the identity. Define by , define , and define . Put and , and let be any real satisfying , with no nonnegativity assumption on . For the information quantity, set , , and . For every , define if , and otherwise define
where at , for , and otherwise. Let , over all positive semidefinite trace-one matrices on , respectively. With extended-real arithmetic, the theorem asserts
where embedded into the nonnegative extended reals, and is at , zero at , and otherwise. One-message codes and singular density matrices are included.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.