Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Conditional min-entropy bounds from classical branches

Open
QAlgorithms.Papers.Independent.Intakenext100.P63.classical_branches_minEntropy

by Gohan · Oct 31, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

Let AAA be a nonzero finite-dimensional space and let JJJ be finite and nonempty. Take a probability vector (qj)(q_j)(qj​) and normalized densities ρjA\rho_j^AρjA​ for every label, including zero-weight labels. Set ρAJ=∑jqjρjA⊗∣j⟩⟨j∣\rho^{AJ}=\sum_jq_j\rho_j^A\otimes|j\rangle\langle j|ρAJ=∑j​qj​ρjA​⊗∣j⟩⟨j∣. For ε≥0\varepsilon\ge0ε≥0, 0≤δ<10\le\delta<10≤δ<1, and a set T⊆JT\subseteq JT⊆J satisfying ∑j∉Tqj≤δ\sum_{j\notin T}q_j\le\delta∑j∈/T​qj​≤δ,

Hmin⁡(A∣J)ρ≥min⁡jHmin⁡(ρj),Hmin⁡ε+δ(A∣J)ρ≥min⁡j∈THmin⁡ε(ρj).H_{\min}(A|J)_\rho\ge\min_jH_{\min}(\rho_j),\qquad H_{\min}^{\varepsilon+\delta}(A|J)_\rho\ge\min_{j\in T}H_{\min}^{\varepsilon}(\rho_j).Hmin​(A∣J)ρ​≥jmin​Hmin​(ρj​),Hminε+δ​(A∣J)ρ​≥j∈Tmin​Hminε​(ρj​).

Smoothing always ranges over normalized states within half trace-norm distance of the original state, and conditional min-entropy optimizes a normalized reference density. The retained set has positive probability and is nonempty. Radii at least one remain defined. These bounds pass uncertainty on individual retained branches to a classical conditioning register.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P63
Formal statement
noncomputable section

namespace QAlgorithms.Papers.Independent.Intakenext100.P63

open QAlgorithms Filter

open QAlgorithms.Papers.Independent.Information100.P63 (Mat)

open QAlgorithms.Papers.Independent.Information100.P04 (Measurement)

open QAlgorithms.Papers.Independent.Information100.P27 (shannon labelInformation)

open QAlgorithms.Papers.Independent.Information100.P71 (channelPower dilationMap SupportIncluded)

open QAlgorithms.Papers.Independent.Bridge50.P47 (SymmetricIndex universalSymmetricDilation)

open QAlgorithms.Learning (IsDistribution)

open scoped Matrix ComplexConjugate ComplexOrder

attribute [local instance] Classical.propDecidable Classical.decEq

/-- PDF p.9, Remark 8: normalized classical branches, including zero-weight labels.
Smoothing uses half trace norm; the retained set has positive probability. -/
theorem classical_branches_minEntropy {A J : Type}
    [Fintype A] [DecidableEq A] [Nonempty A]
    [Fintype J] [DecidableEq J] [Nonempty J]
    (q : J → ℝ) (hq : IsDistribution q) (rho : J → Mat A)
    (hrho : ∀ j, IsDensityMatrix (rho j)) (eps delta : ℝ)
    (he : 0 ≤ eps) (hd0 : 0 ≤ delta) (hd1 : delta < 1)
    (T : Finset J) (hT : ∑ j ∈ Tᶜ, q j ≤ delta) :
    (sInf (Set.range (fun j => stateMinBits (rho j))) ≤ conditionalMinBits (qcBranchDensity q rho)) ∧
    (sInf {h | ∃ j ∈ T, h = traceSmoothStateMinBits eps (rho j)} ≤
      traceSmoothConditionalMinBits (eps + delta) (qcBranchDensity q rho)) := by sorry

end QAlgorithms.Papers.Independent.Intakenext100.P63

end
Source
Weak locking capacity of quantum channels can be much larger than private capacity, arXiv:1403.6361v3, PDF p. 9, Remark 8: conditional min-entropy bounds from classical branches
Read-back

What the Lean code literally says, in plain math · gpt-6.1-sol

Let A,JA,JA,J be finite nonempty types with decidable equality, let q:J→Rq:J\to\mathbb Rq:J→R be nonnegative with ∑jqj=1\sum_jq_j=1∑j​qj​=1, and let every ρj\rho_jρj​ be a positive semidefinite complex matrix on AAA with trace 111. Fix real numbers ε≥0\varepsilon\ge0ε≥0 and 0≤δ<10\le\delta<10≤δ<1, and a subset T⊆JT\subseteq JT⊆J satisfying ∑j∉Tqj≤δ\sum_{j\notin T}q_j\le\delta∑j∈/T​qj​≤δ. Define the block diagonal density matrix

X(a,j),(a′,j′)={qj(ρj)aa′,j=j′,0,j≠j′.X_{(a,j),(a',j')}= \begin{cases} q_j(\rho_j)_{aa'},&j=j',\\ 0,&j\ne j'. \end{cases}X(a,j),(a′,j′)​={qj​(ρj​)aa′​,0,​j=j′,j=j′.​

Write hmin⁡(ω)=−log⁡2∥ω∥oph_{\min}(\omega)=-\log_2\|\omega\|_{\mathrm{op}}hmin​(ω)=−log2​∥ω∥op​ and

Hmin⁡(Y)=real⁡ ⁣(sup⁡λ∈R, σ⪰0, tr⁡σ=12−λIA⊗σ−Y⪰0λ),H_{\min}(Y)=\operatorname{real}\!\left( \sup_{\substack{\lambda\in\mathbb R,\ \sigma\succeq0,\ \operatorname{tr}\sigma=1\\ 2^{-\lambda}I_A\otimes\sigma-Y\succeq0}}\lambda\right),Hmin​(Y)=real​λ∈R, σ⪰0, trσ=12−λIA​⊗σ−Y⪰0​sup​λ​,

where σ\sigmaσ acts on JJJ, the supremum is extended real, and the real conversion sends either infinity to 000. For either real-valued quantity hhh, its trace-distance smoothing is sup⁡{h(ω):ω⪰0,tr⁡ω=1,12∥Y−ω∥tr≤u}\sup\{h(\omega):\omega\succeq0,\operatorname{tr}\omega=1,\frac12\|Y-\omega\|_{\mathrm{tr}}\le u\}sup{h(ω):ω⪰0,trω=1,21​∥Y−ω∥tr​≤u}, where ∥Z∥tr=Re⁡tr⁡Z†Z\|Z\|_{\mathrm{tr}}=\operatorname{Re}\operatorname{tr}\sqrt{Z^\dagger Z}∥Z∥tr​=RetrZ†Z​; denote these smoothings by hmin⁡uh_{\min}^{u}hminu​ and Hmin⁡uH_{\min}^{u}Hminu​, respectively. Then

inf⁡j∈Jhmin⁡(ρj)≤Hmin⁡(X),inf⁡j∈Thmin⁡ε(ρj)≤Hmin⁡ε+δ(X).\inf_{j\in J}h_{\min}(\rho_j)\le H_{\min}(X), \qquad \inf_{j\in T}h_{\min}^{\varepsilon}(\rho_j) \le H_{\min}^{\varepsilon+\delta}(X).j∈Jinf​hmin​(ρj​)≤Hmin​(X),j∈Tinf​hminε​(ρj​)≤Hminε+δ​(X).

The first infimum includes branches with qj=0q_j=0qj​=0. The tail-mass hypothesis and δ<1\delta<1δ<1 force TTT to be nonempty. There is no upper bound on ε\varepsilonε or on ε+δ\varepsilon+\deltaε+δ.

Human review
  • Endorsed by Shuze Chen · Oct 31, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Gohan · Oct 31, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me