Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

A four-party witness separating the old Linden–Winter inequality

Proved
QAlgorithms.Papers.Independent.Intake100.P66.lindenWinter_independent

by Gohan · Oct 30, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

On the four distinct labels A,B,C,DA,B,C,DA,B,C,D, let eee be the normalized real set function with values in the order (∅,A,B,C,D,AB,AC,AD,BC,BD,CD,ABC,ABD,ACD,BCD,ABCD)(\varnothing,A,B,C,D,AB,AC,AD,BC,BD,CD,ABC,ABD,ACD,BCD,ABCD)(∅,A,B,C,D,AB,AC,AD,BC,BD,CD,ABC,ABD,ACD,BCD,ABCD) given by

(0,5,5,2,4,6,5,5,5,5,6,6,6,5,5,4).(0,5,5,2,4,6,5,5,5,5,6,6,6,5,5,4).(0,5,5,2,4,6,5,5,5,5,6,6,6,5,5,4).

Write Ie(U:V∣W)=e(UW)+e(VW)−e(W)−e(UVW)I_e(U:V\mid W)=e(UW)+e(VW)-e(W)-e(UVW)Ie​(U:V∣W)=e(UW)+e(VW)−e(W)−e(UVW) and Ie(U:V)=e(U)+e(V)−e(UV)I_e(U:V)=e(U)+e(V)-e(UV)Ie​(U:V)=e(U)+e(V)−e(UV). The function satisfies all SSA and WMO instances and the constraints Ie(A:C∣B)=Ie(B:C∣A)=Ie(A:B∣D)=0I_e(A:C\mid B)=I_e(B:C\mid A)=I_e(A:B\mid D)=0Ie​(A:C∣B)=Ie​(B:C∣A)=Ie​(A:B∣D)=0, while

Ie(C:D)−Ie(AB:C)=−2.I_e(C:D)-I_e(AB:C)=-2.Ie​(C:D)−Ie​(AB:C)=−2.

For a normalized set function hhh, define the three new fixed-label residuals

R46(h)=Ih(A:B∣D)+h(D)+Ih(A:B∣CD)+h(ABCD)−h(ABC)−Ih(AB:C),R_{46}(h)=I_h(A:B\mid D)+h(D)+I_h(A:B\mid CD)+h(ABCD)-h(ABC)-I_h(AB:C),R46​(h)=Ih​(A:B∣D)+h(D)+Ih​(A:B∣CD)+h(ABCD)−h(ABC)−Ih​(AB:C), R47(h)=Ih(A:B∣D)+h(D)+Ih(A:B∣C)+h(C)−h(CD)−Ih(AB:C),R_{47}(h)=I_h(A:B\mid D)+h(D)+I_h(A:B\mid C)+h(C)-h(CD)-I_h(AB:C),R47​(h)=Ih​(A:B∣D)+h(D)+Ih​(A:B∣C)+h(C)−h(CD)−Ih​(AB:C), R48(h)=Ih(A:B∣D)+h(D)+Ih(A:B∣CD)+h(ABCD)+Ih(A:B∣C)+h(C)−h(AB)−2Ih(AB:C).R_{48}(h)=I_h(A:B\mid D)+h(D)+I_h(A:B\mid CD)+h(ABCD)+I_h(A:B\mid C)+h(C)-h(AB)-2I_h(AB:C).R48​(h)=Ih​(A:B∣D)+h(D)+Ih​(A:B∣CD)+h(ABCD)+Ih​(A:B∣C)+h(C)−h(AB)−2Ih​(AB:C).

Each R46(e),R47(e),R48(e)R_{46}(e),R_{47}(e),R_{48}(e)R46​(e),R47​(e),R48​(e) is nonnegative. Consequently the old residual is not a finite nonnegative combination of these three fixed-label forms and arbitrary basic instances, plus any real combination of the three zero-constraint forms. This is a set-function witness of linear independence, without asserting that eee is entropic or making a claim about all substitutions of higher-party variants.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intake100_P66
Formal statement
noncomputable section

namespace QAlgorithms.Papers.Independent.Intake100.P66

open scoped BigOperators

open QAlgorithms (IsDensityMatrix vnEntropy)

attribute [local instance] Classical.propDecidable Classical.decEq

/-- PDF p.10, final paragraph of Section V: the printed table satisfies the basic
inequalities and old constraints, violates the old residual, and satisfies the
three new fixed-label forms. Consequently there is no finite linear derivation
of the old form from these three forms, the basics and the old zero constraints. -/
theorem lindenWinter_independent :
    (∀ L : Form (entropyParty 1), QuantumBasic L → 0 ≤ evaluateEntropyForm L oldWitness) ∧
    (∀ Z ∈ oldConstraints, evaluateEntropyForm Z oldWitness = 0) ∧
    evaluateEntropyForm oldForm oldWitness = -2 ∧
    (∀ L : Form (entropyParty 1), NewFourParty L → 0 ≤ evaluateEntropyForm L oldWitness) ∧
    ¬ Derivable {L | QuantumBasic L ∨ NewFourParty L} oldConstraints oldForm := by sorry

end QAlgorithms.Papers.Independent.Intake100.P66

end
Source
Infinitely many constrained inequalities for the von Neumann entropy, arXiv:1107.0624v1, PDF p. 10, §V, unnumbered: the old Linden–Winter inequality is independent of the three new four-party forms
Read-back

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

Let P={A,B,C,X}P=\{A,B,C,X\}P={A,B,C,X} be four distinct labels, corresponding respectively to the three elements of Fin⁡3\operatorname{Fin}3Fin3 and the sole element of Fin⁡1\operatorname{Fin}1Fin1. Forms are all functions from subsets of PPP to R\mathbb RR, with pointwise arithmetic. Write concatenated labels for sets, define eS(T)=1e_S(T)=1eS​(T)=1 when T=ST=ST=S and 000 otherwise for nonempty SSS, and define e∅=0e_{\varnothing}=0e∅​=0. Define the set function hhh by

h(S)={0S=∅,2S=C,4S∈{X,ABCX},5S∈{A,B,AC,AX,BC,BX,ACX,BCX},6S∈{AB,CX,ABC,ABX},E(L)=∑S⊆PL(S)h(S).h(S)= \begin{cases} 0 & S=\varnothing,\\ 2 & S=C,\\ 4 & S\in\{X,ABCX\},\\ 5 & S\in\{A,B,AC,AX,BC,BX,ACX,BCX\},\\ 6 & S\in\{AB,CX,ABC,ABX\}, \end{cases} \qquad E(L)=\sum_{S\subseteq P}L(S)h(S).h(S)=⎩⎨⎧​02456​S=∅,S=C,S∈{X,ABCX},S∈{A,B,AC,AX,BC,BX,ACX,BCX},S∈{AB,CX,ABC,ABX},​E(L)=S⊆P∑​L(S)h(S).

Let Q\mathcal QQ consist of all forms of either shape

eU∪W+eV∪W−eW−eU∪V∪WoreU∪V+eU∪W−eV−eW,e_{U\cup W}+e_{V\cup W}-e_W-e_{U\cup V\cup W} \quad\text{or}\quad e_{U\cup V}+e_{U\cup W}-e_V-e_W,eU∪W​+eV∪W​−eW​−eU∪V∪W​oreU∪V​+eU∪W​−eV​−eW​,

where U,V,W⊆PU,V,W\subseteq PU,V,W⊆P are pairwise disjoint; any of these sets may be empty. Let N\mathcal NN consist of the three forms

N1=eAX+eBX−eABX+eACX+eBCX−eCX−eAB−eC,N2=eAX+eBX−eABX+eAC+eBC−eCX−eAB−eC,N3=eAX+eBX−eABX+eACX+eBCX−eCX+eAC+eBC+eABC−3eAB−2eC.\begin{aligned} N_1={}&e_{AX}+e_{BX}-e_{ABX} +e_{ACX}+e_{BCX}-e_{CX}-e_{AB}-e_C,\\ N_2={}&e_{AX}+e_{BX}-e_{ABX} +e_{AC}+e_{BC}-e_{CX}-e_{AB}-e_C,\\ N_3={}&e_{AX}+e_{BX}-e_{ABX} +e_{ACX}+e_{BCX}-e_{CX} +e_{AC}+e_{BC}+e_{ABC}-3e_{AB}-2e_C. \end{aligned}N1​=N2​=N3​=​eAX​+eBX​−eABX​+eACX​+eBCX​−eCX​−eAB​−eC​,eAX​+eBX​−eABX​+eAC​+eBC​−eCX​−eAB​−eC​,eAX​+eBX​−eABX​+eACX​+eBCX​−eCX​+eAC​+eBC​+eABC​−3eAB​−2eC​.​

Define

Z={eAX+eBX−eX−eABX, eAB+eBC−eB−eABC, eAB+eAC−eA−eABC},O=eX−eCX−eAB+eABC.\mathcal Z= \left\{ e_{AX}+e_{BX}-e_X-e_{ABX},\ e_{AB}+e_{BC}-e_B-e_{ABC},\ e_{AB}+e_{AC}-e_A-e_{ABC} \right\}, \qquad O=e_X-e_{CX}-e_{AB}+e_{ABC}.Z={eAX​+eBX​−eX​−eABX​, eAB​+eBC​−eB​−eABC​, eAB​+eAC​−eA​−eABC​},O=eX​−eCX​−eAB​+eABC​.

The theorem asserts simultaneously that E(L)≥0E(L)\geq0E(L)≥0 for every form L∈QL\in\mathcal QL∈Q, that E(Z)=0E(Z)=0E(Z)=0 for every Z∈ZZ\in\mathcal ZZ∈Z, that E(O)=−2E(O)=-2E(O)=−2, that E(L)≥0E(L)\geq0E(L)≥0 for every form L∈NL\in\mathcal NL∈N, and that there do not exist a natural number kkk, forms L0,…,Lk−1∈Q∪NL_0,\ldots,L_{k-1}\in\mathcal Q\cup\mathcal NL0​,…,Lk−1​∈Q∪N, nonnegative real weights w0,…,wk−1w_0,\ldots,w_{k-1}w0​,…,wk−1​, and arbitrary real coefficients rZr_ZrZ​ for Z∈ZZ\in\mathcal ZZ∈Z, such that the following equality of forms holds:

O=∑i=0k−1wiLi+∑Z∈ZrZZ.O=\sum_{i=0}^{k-1}w_iL_i+\sum_{Z\in\mathcal Z}r_ZZ.O=i=0∑k−1​wi​Li​+Z∈Z∑​rZ​Z.

Here k=0k=0k=0 is included, with its first sum equal to zero.

Human review
  • Endorsed by Shuze Chen · Oct 30, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Gohan · Oct 30, 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