Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

§5.2.3: local isometries, sub-unital maps, and curvature

Open
QAlgorithms.Papers.Independent.Intakenext100.P07.conditional_local_properties

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

prove2quantum

All quantum systems below are finite-dimensional complex Hilbert spaces of positive dimension, and logarithms are binary. A normalized state is positive semidefinite with trace one; a subnormalized state has trace at most one. Marginals are partial traces. Tildes denote sandwiched Rényi quantities and bars denote Petz quantities. Their trace functionals are Q~α=Tr⁡(σ(1−α)/(2α)ρσ(1−α)/(2α))α\widetilde Q_\alpha=\operatorname{Tr}(\sigma^{(1-\alpha)/(2\alpha)}\rho\sigma^{(1-\alpha)/(2\alpha)})^\alphaQ​α​=Tr(σ(1−α)/(2α)ρσ(1−α)/(2α))α and Q‾α=Tr⁡(ρασ1−α)\overline Q_\alpha=\operatorname{Tr}(\rho^\alpha\sigma^{1-\alpha})Q​α​=Tr(ρασ1−α); the divergences include division by Tr⁡ρ\operatorname{Tr}\rhoTrρ inside the logarithm. Matrix powers vanish on the kernel, including power zero. The notation ρ≪σ\rho\ll\sigmaρ≪σ means support inclusion. The upward conditional entropy optimizes the negative divergence over normalized reference states on the conditioning system; the downward entropy uses the actual marginal. The notation H(A∣B)=H(AB)−H(B)H(A|B)=H(AB)-H(B)H(A∣B)=H(AB)−H(B) denotes conditional von Neumann entropy.

All four entropies in their DPI ranges are invariant under arbitrary local isometries; sub-unital CPTP operations on A cannot decrease entropy; TNI maps preserving this input trace have the analogous properties. Petz entropies concave for 0≤α≤1 and quasi-concave for 1≤α≤2; sandwiched concave for 1/2≤α≤1, quasi-concave for α>1.

Hα(A′∣B′)(U⊗V)ρ(U⊗V)†=Hα(A∣B)ρ;Hα has the stated concavity and quasi-concavity ranges.H_\alpha(A^\prime|B^\prime)_{(U\otimes V)\rho(U\otimes V)^\dagger}=H_\alpha(A|B)_\rho;\qquad H_\alpha\text{ has the stated concavity and quasi-concavity ranges}.Hα​(A′∣B′)(U⊗V)ρ(U⊗V)†​=Hα​(A∣B)ρ​;Hα​ has the stated concavity and quasi-concavity ranges.

This result records the source’s operational or structural relation, with the stated endpoint and attainment clauses retained.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P07
Formal statement
noncomputable section

open Filter

open scoped BigOperators Matrix ComplexConjugate ComplexOrder Matrix.Norms.L2Operator Topology

namespace QAlgorithms.Papers.Independent.Intakenext100.P07

attribute [local instance] Classical.propDecidable Classical.decEq

variable {I J A B C D X Y Z : Type*}

variable [Fintype I] [DecidableEq I] [Fintype J] [DecidableEq J]

variable [Fintype A] [DecidableEq A] [Fintype B] [DecidableEq B]

variable [Fintype C] [DecidableEq C] [Fintype D] [DecidableEq D]

variable [Fintype X] [DecidableEq X] [Fintype Y] [DecidableEq Y]

variable [Fintype Z] [DecidableEq Z]

/-- NODE paper:1504.00233#5.localproperties
§5.2.3: local isometries, sub-unital maps, and curvature, arXiv:1504.00233v5, PDF p.81. -/
theorem conditional_local_properties [Nonempty A] [Nonempty B] [Nonempty C] [Nonempty D]
    (r : paperMat (A × B)) (hr : IsDensityMatrix r) (sand up : Bool) (a : ENNReal) :
    (∀ U : Matrix C A ℂ, ∀ V : Matrix D B ℂ, Uᴴ*U = 1 → Vᴴ*V = 1 →
      paperConditional sand up a (paperLocalEmbedding U V r) = paperConditional sand up a r) ∧
    (DPIOrder sand a → ∀ E : paperMat A →ₗ[ℂ] paperMat C, ∀ F : paperMat B →ₗ[ℂ] paperMat D,
      CPTNI E → CPTNI F → paperBelow (E 1) 1 → (paperTensorAction E F r).trace = r.trace →
      paperConditional sand up a r ≤ paperConditional sand up a (paperTensorAction E F r)) ∧
    (DPIOrder sand a → ∀ s : paperMat (A × B), IsDensityMatrix s → ∀ l : ℝ, l ∈ Set.Icc 0 1 →
      (a ≤ 1 → (l:EReal)*paperConditional sand up a r+((1-l:ℝ):EReal)*paperConditional sand up a s ≤ paperConditional sand up a (paperMix l r s)) ∧
      (1 ≤ a → min (paperConditional sand up a r) (paperConditional sand up a s) ≤ paperConditional sand up a (paperMix l r s))) := by sorry

end QAlgorithms.Papers.Independent.Intakenext100.P07

end
Source
Quantum Information Processing with Finite Resources -- Mathematical Foundations, arXiv:1504.00233v5, PDF p. 81, §5.2.3: local isometries, sub-unital maps, and curvature
Read-back

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

Let A,B,C,DA,B,C,DA,B,C,D be nonempty finite types with decidable equality, let rrr be a positive semidefinite matrix of complex trace 111 on A×BA\times BA×B, fix Boolean parameters b,ub,ub,u, and fix an order a∈[0,∞]a\in[0,\infty]a∈[0,∞]. The conditional quantity Hab,u(X)H_a^{b,u}(X)Hab,u​(X) is defined as follows. Set X[t]=(x↦1x≠0xt)(X)X^{[t]}=(x\mapsto\mathbf1_{x\ne0}x^t)(X)X[t]=(x↦1x=0​xt)(X) by functional calculus, set L(X)=(x↦log⁡2x)(X)L(X)=(x\mapsto\log_2x)(X)L(X)=(x↦log2​x)(X) with the total real logarithm log⁡20=0\log_2 0=0log2​0=0, and let L(0)=−∞\mathcal L(0)=-\inftyL(0)=−∞, L(z)=log⁡2z\mathcal L(z)=\log_2zL(z)=log2​z for z≠0z\ne0z=0. Put

Qtfalse(X,Y)=Re⁡tr⁡(X[t]Y[1−t]),Qttrue(X,Y)=Re⁡tr⁡[(Y[(1−t)/(2t)]XY[(1−t)/(2t)])[t]].Q_t^{\mathrm{false}}(X,Y)=\operatorname{Re}\operatorname{tr}(X^{[t]}Y^{[1-t]}),\qquad Q_t^{\mathrm{true}}(X,Y)= \operatorname{Re}\operatorname{tr} \left[\left(Y^{[(1-t)/(2t)]}XY^{[(1-t)/(2t)]}\right)^{[t]}\right].Qtfalse​(X,Y)=Retr(X[t]Y[1−t]),Qttrue​(X,Y)=Retr[(Y[(1−t)/(2t)]XY[(1−t)/(2t)])[t]].

For t>0t>0t>0, the extended-real divergence Rtb(X∥Y)R_t^b(X\|Y)Rtb​(X∥Y) at t=1t=1t=1 is +∞+\infty+∞ unless ker⁡Y⊆ker⁡X\ker Y\subseteq\ker XkerY⊆kerX, and otherwise is

Re⁡tr⁡(X(L(X)−L(Y)))Re⁡tr⁡X.\frac{\operatorname{Re}\operatorname{tr}(X(L(X)-L(Y)))} {\operatorname{Re}\operatorname{tr}X}.RetrXRetr(X(L(X)−L(Y)))​.

For t≠1t\ne1t=1, it is

1t−1L ⁣(Qtb(X,Y)Re⁡tr⁡X)\frac1{t-1}\mathcal L\!\left( \frac{Q_t^b(X,Y)}{\operatorname{Re}\operatorname{tr}X}\right)t−11​L(RetrXQtb​(X,Y)​)

when (t<1 and Qtb(X,Y)>0)(t<1\ \text{and }Q_t^b(X,Y)>0)(t<1 and Qtb​(X,Y)>0) or ker⁡Y⊆ker⁡X\ker Y\subseteq\ker XkerY⊆kerX, and is +∞+\infty+∞ otherwise. Define R0b=inf⁡t>0RtbR_0^b=\inf_{t>0}R_t^bR0b​=inft>0​Rtb​; define R∞false=sup⁡t>0RtfalseR_\infty^{\mathrm{false}}=\sup_{t>0}R_t^{\mathrm{false}}R∞false​=supt>0​Rtfalse​; and define

R∞true(X∥Y)=inf⁡{ℓ∈R:X⪯2ℓY},R_\infty^{\mathrm{true}}(X\|Y)= \inf\{\ell\in\mathbb R:X\preceq2^\ell Y\},R∞true​(X∥Y)=inf{ℓ∈R:X⪯2ℓY},

where X⪯YX\preceq YX⪯Y means Y−X⪰0Y-X\succeq0Y−X⪰0, and real candidates are embedded into the extended reals. Positive finite orders use the preceding real-order definition. For a bipartite XXX on U×VU\times VU×V, with (XV)vv′=∑zX(z,v),(z,v′)(X_V)_{vv'}=\sum_zX_{(z,v),(z,v')}(XV​)vv′​=∑z​X(z,v),(z,v′)​, set

Hab,false(X)=−Rab(X∥1U⊗XV),Hab,true(X)=sup⁡s⪰0tr⁡s=1[−Rab(X∥1U⊗s)].H_a^{b,\mathrm{false}}(X)=-R_a^b(X\|1_U\otimes X_V),\qquad H_a^{b,\mathrm{true}}(X)= \sup_{\substack{s\succeq0\\\operatorname{tr}s=1}} [-R_a^b(X\|1_U\otimes s)].Hab,false​(X)=−Rab​(X∥1U​⊗XV​),Hab,true​(X)=s⪰0trs=1​sup​[−Rab​(X∥1U​⊗s)].

Then every pair of matrices U:C×A→CU:C\times A\to\mathbb CU:C×A→C, V:D×B→CV:D\times B\to\mathbb CV:D×B→C with U†U=1AU^\dagger U=1_AU†U=1A​ and V†V=1BV^\dagger V=1_BV†V=1B​ satisfies

Hab,u((U⊗V)r(U⊗V)†)=Hab,u(r).H_a^{b,u}((U\otimes V)r(U\otimes V)^\dagger)=H_a^{b,u}(r).Hab,u​((U⊗V)r(U⊗V)†)=Hab,u​(r).

Also, provided a≥12a\ge\tfrac12a≥21​ when b=trueb=\mathrm{true}b=true, or a≤2a\le2a≤2 when b=falseb=\mathrm{false}b=false, every pair of complex-linear maps E:A→CE:A\to CE:A→C, F:B→DF:B\to DF:B→D that are completely positive and trace nonincreasing on positive semidefinite matrices, and satisfy E(1A)⪯1CE(1_A)\preceq1_CE(1A​)⪯1C​ and tr⁡((E⊗F)(r))=tr⁡r\operatorname{tr}((E\otimes F)(r))=\operatorname{tr}rtr((E⊗F)(r))=trr, obeys

Hab,u(r)≤Hab,u((E⊗F)(r)).H_a^{b,u}(r)\le H_a^{b,u}((E\otimes F)(r)).Hab,u​(r)≤Hab,u​((E⊗F)(r)).

Here complete positivity means positivity of every amplification by Fin⁡(k)\operatorname{Fin}(k)Fin(k), including k=0k=0k=0, and the tensor action is explicitly

((E⊗F)(r))(c,d),(c′,d′)=∑i,i′∈A∑j,j′∈Br(i,j),(i′,j′) E(eii′)cc′ F(ejj′)dd′.((E\otimes F)(r))_{(c,d),(c',d')}= \sum_{i,i'\in A}\sum_{j,j'\in B} r_{(i,j),(i',j')}\,E(e_{ii'})_{cc'}\,F(e_{jj'})_{dd'}.((E⊗F)(r))(c,d),(c′,d′)​=i,i′∈A∑​j,j′∈B∑​r(i,j),(i′,j′)​E(eii′​)cc′​F(ejj′​)dd′​.

Under the same restriction on aaa, every density matrix sss on A×BA\times BA×B and every real 0≤λ≤10\le\lambda\le10≤λ≤1 satisfy both implications

a≤1 ⟹ λHab,u(r)+(1−λ)Hab,u(s)≤Hab,u(λr+(1−λ)s),a\le1\ \Longrightarrow\ \lambda H_a^{b,u}(r)+(1-\lambda)H_a^{b,u}(s) \le H_a^{b,u}(\lambda r+(1-\lambda)s),a≤1 ⟹ λHab,u​(r)+(1−λ)Hab,u​(s)≤Hab,u​(λr+(1−λ)s), 1≤a ⟹ min⁡{Hab,u(r),Hab,u(s)}≤Hab,u(λr+(1−λ)s).1\le a\ \Longrightarrow\ \min\{H_a^{b,u}(r),H_a^{b,u}(s)\} \le H_a^{b,u}(\lambda r+(1-\lambda)s).1≤a ⟹ min{Hab,u​(r),Hab,u​(s)}≤Hab,u​(λr+(1−λ)s).

These are extended-real inequalities, including the endpoint orders 0,+∞0,+\infty0,+∞ when permitted and mixture weights 0,10,10,1; no identities UU†=1CUU^\dagger=1_CUU†=1C​ or VV†=1DVV^\dagger=1_DVV†=1D​ are required.

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