Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Relative entropy of orthogonal mixtures

Open
QAlgorithms.Papers.Independent.Support50.P44.relative_entropy_orthogonal_mixture

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

prove2quantum

Let ρ,ρ′,σ,σ′\rho,\rho',\sigma,\sigma'ρ,ρ′,σ,σ′ be density operators on a complex separable Hilbert space. Assume each of ρ,σ\rho,\sigmaρ,σ is orthogonal to each of ρ′,σ′\rho',\sigma'ρ′,σ′. For p∈[0,1]p\in[0,1]p∈[0,1], let DDD denote Umegaki relative entropy in bits, allowing +∞+\infty+∞. Then

D(pρ+(1−p)ρ′∥pσ+(1−p)σ′)=pD(ρ∥σ)+(1−p)D(ρ′∥σ′).D(p\rho+(1-p)\rho'\|p\sigma+(1-p)\sigma')=pD(\rho\|\sigma)+(1-p)D(\rho'\|\sigma').D(pρ+(1−p)ρ′∥pσ+(1−p)σ′)=pD(ρ∥σ)+(1−p)D(ρ′∥σ′).

Here 0⋅∞=00\cdot\infty=00⋅∞=0. This identity gives exact affinity of relative entropy across matched orthogonal blocks.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P44
Formal statement
noncomputable section

namespace QAlgorithms.Papers.Independent.Support50.P44

open QAlgorithms.Papers.Independent.Dependency50.P13 (IsTraceClass compression)

open QAlgorithms.Papers.Independent.Dependency50.P21 (marginalAB)

open QAlgorithms.Papers.Independent.Renew46.P37 (classicalKL)

open Filter

open scoped ENNReal ComplexOrder Matrix

attribute [local instance] Classical.propDecidable

attribute [local instance] Classical.decEq

variable {H : Type*} [NormedAddCommGroup H] [InnerProductSpace ℂ H] [CompleteSpace H]

/-- PDF p.24, Lemma B.2, (147): orthogonal-block affinity of extended Umegaki
entropy. ENNReal multiplication makes 0 times infinity zero at p = 0 or 1. -/
theorem relative_entropy_orthogonal_mixture {I : Type*} [Countable I]
    (b : HilbertBasis I ℂ H) (rho rho2 sigma sigma2 : H →L[ℂ] H)
    (hrho : IsHilbertDensity b rho) (hrho2 : IsHilbertDensity b rho2)
    (hsigma : IsHilbertDensity b sigma) (hsigma2 : IsHilbertDensity b sigma2)
    (horth : OrthogonalBlocks rho rho2 sigma sigma2)
    (p : ℝ) (hp0 : 0 ≤ p) (hp1 : p ≤ 1) :
    hilbertRelativeEntropy (hilbertMix p rho rho2) (hilbertMix p sigma sigma2) =
      ENNReal.ofReal p * hilbertRelativeEntropy rho sigma +
        ENNReal.ofReal (1-p) * hilbertRelativeEntropy rho2 sigma2 := by sorry

end QAlgorithms.Papers.Independent.Support50.P44

end
Source
Universal recovery map for approximate Markov chains, arXiv:1504.07251v3, PDF p. 24, Lemma B.2
Read-back

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

Let HHH be a complex Hilbert space, let III be countable, and let b=(bi)i∈Ib=(b_i)_{i\in I}b=(bi​)i∈I​ be a Hilbert basis. Let ρ,ρ2,σ,σ2\rho,\rho_2,\sigma,\sigma_2ρ,ρ2​,σ,σ2​ be positive bounded complex linear operators with ∑i⟨bi,Tbi⟩=1\sum_i\langle b_i,Tb_i\rangle=1∑i​⟨bi​,Tbi​⟩=1 for each of these four operators. Assume each is trace class: it has a representation Tx=∑r⟨vr,x⟩urTx=\sum_r\langle v_r,x\rangle u_rTx=∑r​⟨vr​,x⟩ur​ with ∑r∥ur∥∥vr∥<∞\sum_r\|u_r\|\|v_r\|<\infty∑r​∥ur​∥∥vr​∥<∞. Suppose ρρ2=ρσ2=σρ2=σσ2=0\rho\rho_2=\rho\sigma_2=\sigma\rho_2=\sigma\sigma_2=0ρρ2​=ρσ2​=σρ2​=σσ2​=0. For bounded operators S,TS,TS,T, define DH(S,T)D_H(S,T)DH​(S,T) to be the supremum, in [0,∞][0,\infty][0,∞], over every natural number nnn and every orthonormal family (v0,…,vn−1)(v_0,\ldots,v_{n-1})(v0​,…,vn−1​) in HHH, of the following quantity. Form the matrices Xij=⟨vi,Svj⟩X_{ij}=\langle v_i,Sv_j\rangleXij​=⟨vi​,Svj​⟩ and Yij=⟨vi,Tvj⟩Y_{ij}=\langle v_i,Tv_j\rangleYij​=⟨vi​,Tvj​⟩; if ker⁡Y⊆ker⁡X\ker Y\subseteq\ker XkerY⊆kerX, use

max⁡ ⁣{0, Re⁡Tr⁡ ⁣(f(X)−Xg(Y)−X+Y)ln⁡2},f(x)=xln⁡x,g(x)=ln⁡x,\max\!\left\{0,\, \frac{\operatorname{Re}\operatorname{Tr}\!\left(f(X)-Xg(Y)-X+Y\right)}{\ln2} \right\}, \qquad f(x)=x\ln x,\quad g(x)=\ln x,max{0,ln2ReTr(f(X)−Xg(Y)−X+Y)​},f(x)=xlnx,g(x)=lnx,

with f(X),g(Y)f(X),g(Y)f(X),g(Y) evaluated by real continuous functional calculus; otherwise use +∞+\infty+∞. The empty orthonormal family n=0n=0n=0 is included. Then, for every p∈[0,1]p\in[0,1]p∈[0,1],

DH ⁣(pρ+(1−p)ρ2, pσ+(1−p)σ2)=pDH(ρ,σ)+(1−p)DH(ρ2,σ2).D_H\!\left(p\rho+(1-p)\rho_2,\ p\sigma+(1-p)\sigma_2\right) =pD_H(\rho,\sigma)+(1-p)D_H(\rho_2,\sigma_2).DH​(pρ+(1−p)ρ2​, pσ+(1−p)σ2​)=pDH​(ρ,σ)+(1−p)DH​(ρ2​,σ2​).

The equality is in [0,∞][0,\infty][0,∞], with the convention 0⋅(+∞)=00\cdot(+\infty)=00⋅(+∞)=0, so it includes both endpoints.

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