Theorem 8 — stabilized Dobrushin characterization
OpenQSDMixing.GU.Supplement.stabilized_dobrushin_characterizationfourier-analysismarkov-chainp2o-gran-per-chapterp2o-plan-paperp2o-v1quantum-mixing-timequantum-state-discrimination
Let . There is a constant , independent of the finite-dimensional uniform mixed-state ensemble, such that the following holds whenever distinct states have fidelity below one. If is formed from the ancillary-stabilized trace-norm contraction coefficients of the inhomogeneous channels , then , all relevant copy counts are finite, and
Moreover, the complete chain is
These inequalities state and the clause of equation (33) with constants depending only on .
Differences from the source:
- reading: and → one positive constant chosen after and before the ensemble, with every comparison written as an inequality; reason: taking the maximum of the finitely many source constants yields one ensemble-uniform factor
- encoding: → a supremum over ancillary spaces for all positive integers , Hermitian nonzero inputs, and literal zero diagonal blocks; reason: every finite-dimensional ancillary system is unitarily some
- encoding: → its direct matrix action multiplying each block by ; reason: this is exactly the composite specified by equations (162)–(164)
- encoding: asymptotic comparisons of possibly undefined reciprocal gaps or infinite minima → and finite natural witnesses are conclusions; reason: Lean's real division by zero and total infima would otherwise trivialize the statement
Preamble
import Mathlib import Definitions.Def_WildeQIT_fidelity import Definitions.Def_QSDMixing_GU_kMin import Definitions.Def_QSDMixing_GU_kWc import Definitions.Def_QSDMixing_GU_Supplement_MixedEnsembleFactorization set_option autoImplicit false
Formal statement
namespace QSDMixing.GU.Supplement
/-- **Theorem 8, pp.33--34.** Uniformly for `ε∈(0,1/2)`, worst-case discrimination
sample complexity is comparable to the reciprocal gap of the stabilized generalized Dobrushin
coefficient. The same ensemble-independent constant also records the complete chain
`t_min(4√ε) ≤ k_min(ε) ≤ t_min(ε) ≤ O_ε(k_wc)`.
Zhou, Leditzky, *A mixing time method for estimating the sample complexity of quantum state
discrimination*, arXiv:2609.40182 (held version, sha256 62fb68dd…), Theorem 8,
pp.33--34, eqs. (32)--(33). -/
theorem stabilized_dobrushin_characterization (ε : ℝ) (hε : 0 < ε) (hε' : ε < 1 / 2) :
∃ C : ℝ, 0 < C ∧
∀ (I : Type) [Fintype I] [Nonempty I] (d : ℕ)
(E : MixedEnsembleFactorization I d),
(∀ i j, i ≠ j →
WildeQIT.fidelity (mixedStateMatrix (E.state i)) (mixedStateMatrix (E.state j)) < 1) →
let κ := stabilizedDobrushinGen E
κ < 1 ∧
∃ km tm tw : ℕ,
QSDMixing.GU.kMin E.state ε = km ∧
mixedWeakMixingTime E ε = tm ∧
mixedWeakMixingTime E (4 * Real.sqrt ε) = tw ∧
∃ kw : ℕ,
QSDMixing.GU.kWc E.state ε = kw ∧
(tw : ℝ) ≤ km ∧ (km : ℝ) ≤ tm ∧ (tm : ℝ) ≤ C * kw ∧
(kw : ℝ) ≤ C * (1 / (1 - κ)) ∧ 1 / (1 - κ) ≤ C * kw := by sorry
end QSDMixing.GU.Supplement
Source
Zhou, Leditzky, A mixing time method for estimating the sample complexity of quantum state discrimination, arXiv:2609.40182 (held version, sha256 62fb68dd…), Theorem 8, pp.33–34, eqs. (32)–(33)