Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Joint quasi-convexity on positive operators

Open
QAlgorithms.Papers.Independent.Support50.P19.jointQuasiConvexity

by Goku · Oct 23, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

Let 1<α≤21<\alpha\le21<α≤2, let XXX be a finite nonempty index set, let px≥0p_x\ge0px​≥0 with ∑xpx=1\sum_xp_x=1∑x​px​=1, and let Ax,Bx≥0A_x,B_x\ge0Ax​,Bx​≥0 act on one nonzero finite complex space. Write A=∑xpxAxA=\sum_xp_xA_xA=∑x​px​Ax​ and B=∑xpxBxB=\sum_xp_xB_xB=∑x​px​Bx​. Then

D~α(A∥B)≤max⁡x∈XD~α(Ax∥Bx).\widetilde D_\alpha(A\Vert B)\le\max_{x\in X}\widetilde D_\alpha(A_x\Vert B_x).Dα​(A∥B)≤x∈Xmax​Dα​(Ax​∥Bx​).

The maximum is taken in the extended real line, so infinite values and zero probabilities are allowed. The operators are not assumed normalized. This controls a mixture by its worst constituent pair.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P19
Formal statement
noncomputable section

open scoped Matrix ComplexOrder ComplexConjugate ENNReal

open Filter Topology

namespace QAlgorithms.Papers.Independent.Support50.P19

attribute [local instance] Classical.propDecidable Classical.decEq

open QAlgorithms.Papers.Independent.Dependency50.P35
  (tr Dominates spectralPower elog elogNN rawTrace cq)

open QAlgorithms.Papers.Independent.Dependency50.P09 (sandwichedD petzD relativeD)

open QAlgorithms.Papers.Independent.Information100.P56 (sandwichedTrace)

open QAlgorithms.Papers.Independent.Information100.P27 (average)

open QAlgorithms.Papers.Independent.Information100.P67
  (IsIsometry receiverChannel complementaryChannel)

open QAlgorithms.Papers.Independent.Claimed50.P48 (FiniteKrausCP)

open QAlgorithms.Papers.Independent.Dependency50.P23
  (UACode uaCodeSuccess ordinaryClassicalCapacity IsEB IsHadamard)

open QAlgorithms.Papers.Independent.Information100.P08 (Ensemble)

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

open QAlgorithms.Papers.Independent.Information100.P62 (tensorChannel)

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

open QAlgorithms.Papers.Independent.Claimed50.P26
  (CovariantIrreducible minOutput UnassistedExponentialConverse)

variable {I J K L : Type} [Fintype I] [Fintype J] [Fintype K] [Fintype L]

variable [DecidableEq I] [DecidableEq J] [DecidableEq K] [DecidableEq L]

/-- PDF p.15, Corollary 9: finite nonempty family of arbitrary positive operators. -/
theorem jointQuasiConvexity [Nonempty I] (a : ℝ) (ha : 1 < a) (ha2 : a ≤ 2)
    (m : ℕ) (hm : 0 < m) (p : Fin m → ℝ) (hp : ∀ x, 0 ≤ p x) (hs : ∑ x, p x = 1)
    (A B : Fin m → Matrix I I ℂ) (hA : ∀ x, (A x).PosSemidef) (hB : ∀ x, (B x).PosSemidef) :
    sandwichedD a (average p A) (average p B) ≤ ⨆ x : Fin m, sandwichedD a (A x) (B x) := by sorry

end QAlgorithms.Papers.Independent.Support50.P19

end
Source
Strong converse for the classical capacity of entanglement-breaking and Hadamard channels via a sandwiched Renyi relative entropy, arXiv:1306.1586v4, PDF p. 15, corollary9
Read-back

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

Let III be a finite nonempty index set with decidable equality, let 1<a≤21<a\le21<a≤2, let m>0m>0m>0 be a natural number, and let p0,…,pm−1p_0,\ldots,p_{m-1}p0​,…,pm−1​ be nonnegative real weights summing to one. For every pair of families Ax,BxA_x,B_xAx​,Bx​ of positive semidefinite complex matrices indexed by III, with no trace normalization or nonzero requirements,

Da ⁣(∑x=0m−1pxAx ∥ ∑x=0m−1pxBx)≤sup⁡0≤x<mDa(Ax∥Bx).D_a\!\left(\sum_{x=0}^{m-1}p_xA_x\,\middle\|\,\sum_{x=0}^{m-1}p_xB_x\right) \le \sup_{0\le x<m}D_a(A_x\|B_x).Da​(x=0∑m−1​px​Ax​​x=0∑m−1​px​Bx​)≤0≤x<msup​Da​(Ax​∥Bx​).

The supremum includes every index, including indices with px=0p_x=0px​=0. For positive semidefinite X,YX,YX,Y, the extended-real quantity Da(X∥Y)D_a(X\|Y)Da​(X∥Y) is +∞+\infty+∞ if ker⁡Y⊈ker⁡X\ker Y\not\subseteq\ker XkerY⊆kerX, and otherwise is

1a−1L ⁣(Re⁡tr⁡[(Y[(1−a)/(2a)]XY[(1−a)/(2a)])[a]]),\frac{1}{a-1}\mathcal L\!\left(\operatorname{Re}\operatorname{tr} \left[(Y^{[(1-a)/(2a)]}XY^{[(1-a)/(2a)]})^{[a]}\right]\right),a−11​L(Retr[(Y[(1−a)/(2a)]XY[(1−a)/(2a)])[a]]),

where spectral powers send zero eigenvalues to zero and L(t)=log⁡2t\mathcal L(t)=\log_2tL(t)=log2​t for t>0t>0t>0 and −∞-\infty−∞ otherwise. Thus zero first matrices, which give −∞-\infty−∞, and support failures, which give +∞+\infty+∞, are included.

Human review
  • Endorsed by Shuze Chen · Oct 24, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Goku · Oct 24, 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