Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proposition 10

Open
QAlgorithms.Papers.Independent.Intakenext100.P56.squashed_product_subadditive

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 α>0\alpha>0α>0, α≠1\alpha\ne1α=1, and let σA1B1\sigma_{A_1B_1}σA1​B1​​ and τA2B2\tau_{A_2B_2}τA2​B2​​ be independent states. Regroup their tensor product as ρA1A2B1B2=σ⊗τ\rho_{A_1A_2B_1B_2}=\sigma\otimes\tauρA1​A2​B1​B2​​=σ⊗τ. Then

Eαsq(A1A2;B1B2)ρ≤Eαsq(A1;B1)σ+Eαsq(A2;B2)τ.E^{\rm sq}_\alpha(A_1A_2;B_1B_2)_\rho\le E^{\rm sq}_\alpha(A_1;B_1)_\sigma+E^{\rm sq}_\alpha(A_2;B_2)_\tau.Eαsq​(A1​A2​;B1​B2​)ρ​≤Eαsq​(A1​;B1​)σ​+Eαsq​(A2​;B2​)τ​.

The extension dimensions of the two factors are independent. This is subadditivity on product states; it does not assert equality or superadditivity.

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 p.11, Proposition 10, (4.30): SUBadditivity. -/
theorem squashed_product_subadditive (a : ℝ) (ha : RenyiOrder a)
    (rho : Op (A × B)) (sigma : Op (C × D))
    (hrho : IsDensityMatrix rho) (hsigma : IsDensityMatrix sigma) :
    renyiSquashed a (bipartiteProduct rho sigma) ≤
      renyiSquashed a rho + renyiSquashed a sigma := by sorry

end QAlgorithms.Papers.Independent.Intakenext100.P56

end
Source
Rényi squashed entanglement, discord, and relative entropy differences, arXiv:1410.1443v3, PDF p. 11, Proposition 10
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 finite register index types with decidable equality, let a>0a>0a>0, a≠1a\ne1a=1, and let ρ\rhoρ and σ\sigmaσ be positive semidefinite trace-111 matrices on A×BA\times BA×B and C×DC\times DC×D, respectively. For positive semidefinite ZZZ, let Pt(Z)P_t(Z)Pt​(Z) be spectral functional calculus assigning 000 to a zero eigenvalue and xtx^txt otherwise. For a density matrix ω\omegaω on L×(R×E)L\times(R\times E)L×(R×E), define

Ia(ω)=aa−1log⁡Re⁡tr⁡P1/a ⁣(S Tr⁡L(R0Pa(ω)R0)S),R0=embed⁡LEP(1−a)/2(ωLE),S=IR⊗P(a−1)/2(ωE),I_a(\omega)=\frac{a}{a-1}\log\operatorname{Re}\operatorname{tr} P_{1/a}\!\left(S\,\operatorname{Tr}_L(R_0P_a(\omega)R_0)S\right), \quad R_0=\operatorname{embed}_{LE}P_{(1-a)/2}(\omega_{LE}),\quad S=I_R\otimes P_{(a-1)/2}(\omega_E),Ia​(ω)=a−1a​logRetrP1/a​(STrL​(R0​Pa​(ω)R0​)S),R0​=embedLE​P(1−a)/2​(ωLE​),S=IR​⊗P(a−1)/2​(ωE​),

where marginals are partial traces and the embedding acts identically on RRR. Let Qa(τ)Q_a(\tau)Qa​(τ) be one half of the extended-real infimum of Ia(ω)I_a(\omega)Ia​(ω) over all density-matrix extensions of τ\tauτ, allowing every auxiliary dimension e+1e+1e+1, e∈Ne\in\mathbb Ne∈N. Define the matrix ρ⊠σ\rho\boxtimes\sigmaρ⊠σ on (A×C)×(B×D)(A\times C)\times(B\times D)(A×C)×(B×D) by

(ρ⊠σ)((u,c),(v,d)),((u′,c′),(v′,d′))=ρ(u,v),(u′,v′)σ(c,d),(c′,d′).(\rho\boxtimes\sigma)_{((u,c),(v,d)),((u',c'),(v',d'))} =\rho_{(u,v),(u',v')}\sigma_{(c,d),(c',d')}.(ρ⊠σ)((u,c),(v,d)),((u′,c′),(v′,d′))​=ρ(u,v),(u′,v′)​σ(c,d),(c′,d′)​.

Then Qa(ρ⊠σ)≤Qa(ρ)+Qa(σ)Q_a(\rho\boxtimes\sigma)\le Q_a(\rho)+Q_a(\sigma)Qa​(ρ⊠σ)≤Qa​(ρ)+Qa​(σ), as an inequality in the extended real numbers. Logarithms are natural, and no upper bound on aaa is assumed.

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