Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proposition 21

Open
QAlgorithms.Papers.Independent.Intakenext100.P56.pure_discord_formula

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 ∣ψ⟩AB|\psi\rangle_{AB}∣ψ⟩AB​ be normalized and α∈(0,1)∪(1,2]\alpha\in(0,1)\cup(1,2]α∈(0,1)∪(1,2]. For a finite rank-one POVM ∑x∣φx⟩⟨φx∣=IA\sum_x|\varphi_x\rangle\langle\varphi_x|=I_A∑x​∣φx​⟩⟨φx​∣=IA​, set zx=⟨φx∣ψ⟩z_x=\langle\varphi_x|\psi\ranglezx​=⟨φx​∣ψ⟩, px=∥zx∥2p_x=\|z_x\|^2px​=∥zx​∥2, and ξx=zx/px\xi_x=z_x/\sqrt{p_x}ξx​=zx​/px​​ whenever px>0p_x>0px​>0. Then

Dα(Aˉ;B)ψ=inf⁡φαα−1log⁡∑x:px>0px⟨ξx∣ψB1−α∣ξx⟩1/α.D_\alpha(\bar A;B)_\psi=\inf_{\varphi}\frac{\alpha}{\alpha-1}\log\sum_{x:p_x>0}p_x\langle\xi_x|\psi_B^{1-\alpha}|\xi_x\rangle^{1/\alpha}.Dα​(Aˉ;B)ψ​=φinf​α−1α​logx:px​>0∑​px​⟨ξx​∣ψB1−α​∣ξx​⟩1/α.

The prefactor remains inside the infimum and no normalization is imposed on the POVM kets. This evaluates pure-state discord through the computed conditional kets.

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.20–21, Proposition 21, (5.27)–(5.28). -/
theorem pure_discord_formula (a : ℝ) (ha : DPOrder a)
    (psi : EuclideanSpace ℂ (A × B)) (hpsi : ‖psi‖ = 1) :
    renyiDiscord a (pureDensity psi) = sInf {v : EReal |
      ∃ m : ℕ, 0 < m ∧ ∃ phi : Fin m → EuclideanSpace ℂ A,
        KetPOVM phi ∧ v = (pureDiscordObjective a psi phi : EReal)} := by sorry

end QAlgorithms.Papers.Independent.Intakenext100.P56

end
Source
Rényi squashed entanglement, discord, and relative entropy differences, arXiv:1410.1443v3, PDF p. 20, Proposition 21
Read-back

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

For finite register types A,BA,BA,B, with decidable equality, every 0<a≤20<a\le20<a≤2 with a≠1a\ne1a=1, and every vector ψ∈CA×B\psi\in\mathbb C^{A\times B}ψ∈CA×B with ∥ψ∥=1\|\psi\|=1∥ψ∥=1, put ρ=∣ψ⟩⟨ψ∣\rho=|\psi\rangle\langle\psi|ρ=∣ψ⟩⟨ψ∣ and ρB=Tr⁡Aρ\rho_B=\operatorname{Tr}_A\rhoρB​=TrA​ρ. The declaration equates two extended-real infima:

Da(ρ)=inf⁡m>0, ϕx∈CA∑x<m∣ϕx⟩⟨ϕx∣=IAaa−1log⁡∑x<m{px(Re⁡⟨ξx,P1−a(ρB)ξx⟩)1/a,px>0,0,px=0,\mathcal D_a(\rho)= \inf_{\substack{m>0,\ \phi_x\in\mathbb C^A\\ \sum_{x<m}|\phi_x\rangle\langle\phi_x|=I_A}} \frac{a}{a-1}\log \sum_{x<m} \begin{cases} p_x\bigl(\operatorname{Re}\langle\xi_x,P_{1-a}(\rho_B)\xi_x\rangle\bigr)^{1/a}, &p_x>0,\\ 0,&p_x=0, \end{cases}Da​(ρ)=m>0, ϕx​∈CA∑x<m​∣ϕx​⟩⟨ϕx​∣=IA​​inf​a−1a​logx<m∑​{px​(Re⟨ξx​,P1−a​(ρB​)ξx​⟩)1/a,0,​px​>0,px​=0,​

where zx(b)=∑i∈Aϕx(i)‾ψ(i,b)z_x(b)=\sum_{i\in A}\overline{\phi_x(i)}\psi(i,b)zx​(b)=∑i∈A​ϕx​(i)​ψ(i,b), px=∥zx∥2p_x=\|z_x\|^2px​=∥zx​∥2, and, in the positive branch, ξx=zx/px\xi_x=z_x/\sqrt{p_x}ξx​=zx​/px​​. Here PtP_tPt​ is spectral functional calculus with scalar value 000 at 000 and xtx^txt elsewhere. The left-hand quantity Da(ρ)\mathcal D_a(\rho)Da​(ρ) is the infimum over all positive integers mmm and positive semidefinite families MxM_xMx​ summing to IAI_AIA​ of

aa−1log⁡Re⁡Tr⁡P1/a ⁣(S Tr⁡H(RPa(ΩM)R) S),R=P(1−a)/2((ΩM)HL)⊗IK,S=IK⊗P(a−1)/2((ΩM)L),\frac{a}{a-1}\log\operatorname{Re}\operatorname{Tr} P_{1/a}\!\left(S\,\operatorname{Tr}_H(RP_a(\Omega_M)R)\,S\right), \quad R=P_{(1-a)/2}((\Omega_M)_{HL})\otimes I_K,\quad S=I_K\otimes P_{(a-1)/2}((\Omega_M)_L),a−1a​logReTrP1/a​(STrH​(RPa​(ΩM​)R)S),R=P(1−a)/2​((ΩM​)HL​)⊗IK​,S=IK​⊗P(a−1)/2​((ΩM​)L​),

with H={0,…,m−1}×AH=\{0,\ldots,m-1\}\times AH={0,…,m−1}×A, K=BK=BK=B, L={0,…,m−1}L=\{0,\ldots,m-1\}L={0,…,m−1}, RRR acting on H,LH,LH,L, and

ΩM((x,i),(b,y);(x′,i′),(b′,y′))={∑k,l(Mx)ikρ(k,b),(l,b′)(Mx′)i′l‾,x=y, x′=y′,0,otherwise.\Omega_M((x,i),(b,y);(x',i'),(b',y'))= \begin{cases} \displaystyle\sum_{k,l}(\sqrt{M_x})_{ik}\rho_{(k,b),(l,b')} \overline{(\sqrt{M_{x'}})_{i'l}},&x=y,\ x'=y',\\ 0,&\text{otherwise}. \end{cases}ΩM​((x,i),(b,y);(x′,i′),(b′,y′))=⎩⎨⎧​k,l∑​(Mx​​)ik​ρ(k,b),(l,b′)​(Mx′​​)i′l​​,0,​x=y, x′=y′,otherwise.​

All real objectives are embedded into the extended reals. The measurement vectors need not be normalized and may be zero; zero-probability terms are explicitly 000. Logarithms are natural, with log⁡0=0\log0=0log0=0.

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