Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Unnumbered: Petz difference monotonicity near one, as proved in Appendix B

Open
QAlgorithms.Papers.Independent.Intakenext100.P56.petz_difference_near_one

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

prove2quantum

All systems are finite-dimensional complex Hilbert spaces with nonempty finite bases. States are positive semidefinite operators of trace one; pure kets have norm one. All marginal states below are partial traces of the stated joint state. Logarithms are natural, and spectral powers vanish on the kernel, including at exponent zero.

Let ρ,σ\rho,\sigmaρ,σ be positive definite states and N\mathcal NN a strict CPTP map. Suppose the full computed variance V(ρ,σ,N)V(\rho,\sigma,\mathcal N)V(ρ,σ,N) of (B.14) is strictly positive. Then there exists an instance-dependent 0<η<10<\eta<10<η<1 such that

1−η<α≤β<1+η⟹Δα(ρ,σ,N)≤Δβ(ρ,σ,N).1-\eta<\alpha\le\beta<1+\eta\quad\Longrightarrow\quad\Delta_\alpha(\rho,\sigma,\mathcal N)\le\Delta_\beta(\rho,\sigma,\mathcal N).1−η<α≤β<1+η⟹Δα​(ρ,σ,N)≤Δβ​(ρ,σ,N).

The one endpoint uses the ordinary relative entropy relativeEntropyDifference. This states the Petz neighborhood monotonicity justified in Appendix B; it does not include the explicitly unhandled case V=0V=0V=0 or assert a sandwiched analogue.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P56
Formal statement
noncomputable section

open Filter Topology

open scoped Matrix ComplexConjugate ComplexOrder

namespace QAlgorithms.Papers.Independent.Intakenext100.P56

attribute [local instance] Classical.propDecidable Classical.decEq

open QAlgorithms

open QAlgorithms.Papers.Independent.Information100.P56 (IsLaw entropy average holevo)

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

open QAlgorithms.Papers.Independent.Dependency50.P21 (channelAdjoint petz extendedLog marginalAB)

open QAlgorithms.Papers.Independent.Information100.P21 (Op FiniteLOCC)

open QAlgorithms.Papers.Independent.Dependency50.P35 (marginalAC)

open QAlgorithms.Papers.Independent.Dependency50.P07 (bipartiteProduct)

open QAlgorithms.Papers.Independent.Information100.P67 (squaredFidelity)

open QAlgorithms.Papers.Independent.Advance50.P25 (renyiEntropyNats)

variable {A B C D E X : Type}

variable [Fintype A] [Fintype B] [Fintype C] [Fintype D] [Fintype E] [Fintype X]

variable [Nonempty A] [Nonempty B] [Nonempty C] [Nonempty D] [Nonempty E] [Nonempty X]

open scoped Matrix.Norms.L2Operator

/-- arXiv:1410.1443v3, PDF pp.34–37, Appendix B: PETZ local monotonicity
under the proof's explicit V>0 condition. Neighborhood depends on rho,sigma,N. -/
theorem petz_difference_near_one (rho sigma : Op A) (N : Superop A B)
    (hrho : IsDensityMatrix rho) (hprho : rho.PosDef)
    (hsigma : IsDensityMatrix sigma) (hpsigma : sigma.PosDef) (hN : StrictChannel N)
    (hV : 0 < differenceVariance rho sigma N) :
    ∃ eta : ℝ, 0 < eta ∧ eta < 1 ∧
      MonotoneOn (fun a : ℝ => differencePetz a rho sigma N) (Set.Ioo (1-eta) (1+eta)) := by sorry

end QAlgorithms.Papers.Independent.Intakenext100.P56

end
Source
Rényi squashed entanglement, discord, and relative entropy differences, arXiv:1410.1443v3, PDF p. 26, Unnumbered: Petz difference monotonicity near one, as proved in Appendix B
Read-back

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

Let A,BA,BA,B be finite index types with decidable equality, and let ρ,σ\rho,\sigmaρ,σ be positive definite complex matrices on AAA, each with trace 111. Let N(X)=∑jKjXKj†N(X)=\sum_jK_jXK_j^\daggerN(X)=∑j​Kj​XKj†​ for a finite family satisfying ∑jKj†Kj=IA\sum_jK_j^\dagger K_j=I_A∑j​Kj†​Kj​=IA​, and assume that every amplification id{0,…,r−1}⊗N\mathrm{id}_{\{0,\ldots,r-1\}}\otimes Nid{0,…,r−1}​⊗N, for every natural number rrr, sends positive definite matrices to positive definite matrices. Define

N∗(Y)ij=∑k,lN(Eij)kl‾Ykl,L=log⁡N(ρ)−log⁡N(σ),H=log⁡ρ−log⁡σ−N∗(L),d=Re⁡tr⁡ ⁣(ρ(log⁡ρ−log⁡σ))−Re⁡tr⁡ ⁣(N(ρ)L),V=Re⁡tr⁡ ⁣(ρ(H−dIA)2)+Re⁡tr⁡ ⁣(N(ρ)L2)−Re⁡tr⁡ ⁣(ρ(N∗(L))2).\begin{gathered} N^*(Y)_{ij}=\sum_{k,l}\overline{N(E_{ij})_{kl}}Y_{kl}, \qquad L=\log N(\rho)-\log N(\sigma),\\ H=\log\rho-\log\sigma-N^*(L),\qquad d=\operatorname{Re}\operatorname{tr}\!\bigl(\rho(\log\rho-\log\sigma)\bigr) -\operatorname{Re}\operatorname{tr}\!\bigl(N(\rho)L\bigr),\\ V=\operatorname{Re}\operatorname{tr}\!\left(\rho(H-dI_A)^2\right) +\operatorname{Re}\operatorname{tr}\!\left(N(\rho)L^2\right) -\operatorname{Re}\operatorname{tr}\!\left(\rho(N^*(L))^2\right). \end{gathered}N∗(Y)ij​=k,l∑​N(Eij​)kl​​Ykl​,L=logN(ρ)−logN(σ),H=logρ−logσ−N∗(L),d=Retr(ρ(logρ−logσ))−Retr(N(ρ)L),V=Retr(ρ(H−dIA​)2)+Retr(N(ρ)L2)−Retr(ρ(N∗(L))2).​

Assume V>0V>0V>0. Write Z[t]Z^{[t]}Z[t] for continuous functional calculus with scalar value 000 at zero and utu^tut otherwise, and define

Dt={d,t=1,log⁡Re⁡tr⁡(ρ[t]StN∗(TtN(ρ)[1−t]Tt)St)t−1,t≠1,St=σ[(1−t)/2],Tt=N(σ)[(t−1)/2].D_t= \begin{cases} d,&t=1,\\[2mm] \displaystyle \frac{\log\operatorname{Re}\operatorname{tr} \left(\rho^{[t]}S_tN^*(T_tN(\rho)^{[1-t]}T_t)S_t\right)}{t-1}, &t\ne1, \end{cases} \qquad S_t=\sigma^{[(1-t)/2]},\quad T_t=N(\sigma)^{[(t-1)/2]}.Dt​=⎩⎨⎧​d,t−1logRetr(ρ[t]St​N∗(Tt​N(ρ)[1−t]Tt​)St​)​,​t=1,t=1,​St​=σ[(1−t)/2],Tt​=N(σ)[(t−1)/2].

Then there exists a real η\etaη with 0<η<10<\eta<10<η<1 such that Ds≤DtD_s\le D_tDs​≤Dt​ whenever 1−η<s≤t<1+η1-\eta<s\le t<1+\eta1−η<s≤t<1+η. This is nondecreasingness on the entire open interval, including comparisons involving the defined value D1=dD_1=dD1​=d. All logarithms are natural.

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