Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Bennett’s local-discarding example increases clipped coherent information

Proved
QAlgorithms.Papers.Independent.Support50.P39.bennettDiscarding

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

prove2quantum

Let T,A,BT,A,BT,A,B be nonzero finite dimensional registers, with Alice owning TTT and AAA and Bob owning BBB. Let τT\tau_TτT​ and σAB\sigma_{AB}σAB​ be normalized density operators. Write IB(ω)=max⁡{S(ωB)−S(ω),0}I^B(\omega)=\max\{S(\omega_B)-S(\omega),0\}IB(ω)=max{S(ωB​)−S(ω),0}, with entropies in bits. Suppose t=S(τ)>0t=S(\tau)>0t=S(τ)>0 and b=IB(σ)>0b=I^B(\sigma)>0b=IB(σ)>0. Regard ρ=τT⊗σAB\rho=\tau_T\otimes\sigma_{AB}ρ=τT​⊗σAB​ as a state on (T⊗A):B(T\otimes A):B(T⊗A):B. Then

Tr⁡Tρ=σ,IB(ρ)=max⁡{b−t,0}<b=IB(Tr⁡Tρ).\operatorname{Tr}_T\rho=\sigma,\qquad I^B(\rho)=\max\{b-t,0\}<b=I^B(\operatorname{Tr}_T\rho).TrT​ρ=σ,IB(ρ)=max{b−t,0}<b=IB(TrT​ρ).

Discarding a mixed local factor therefore strictly increases the paper's clipped coherent information, including when t≥bt\ge bt≥b. The source attributes this example to Charles Bennett; its subsequent question about two-qubit states is not a theorem.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P39
Formal statement
namespace QAlgorithms.Papers.Independent.Support50.P39

noncomputable section

open QAlgorithms

open Filter Topology

open scoped Matrix ComplexConjugate ComplexOrder

attribute [local instance] Classical.propDecidable Classical.decEq

/-- arXiv:quant-ph/0106080v3, PDF p.6, Bennett's example (attributed result).
For t=S(tau)>0 and b=I^B(sigma)>0 the initial clipped information is
max(b-t,0), and actually discarding Alice's T factor gives sigma and strictly
increases coherent information. No two-qubit conjecture is asserted. -/
theorem bennettDiscarding {T A B : Type*} [Fintype T] [Fintype A] [Fintype B]
    [DecidableEq T] [DecidableEq A] [DecidableEq B]
    [Nonempty T] [Nonempty A] [Nonempty B]
    (tau : Matrix T T ℂ) (sigma : Matrix (A × B) (A × B) ℂ)
    (htau : IsDensityMatrix tau) (hsigma : IsDensityMatrix sigma)
    (ht : 0 < vnEntropy tau) (hb : 0 < clippedCoherentInformation sigma) :
    discardFirstAlice (productFactor tau sigma) = sigma ∧
    clippedCoherentInformation (productFactor tau sigma) =
      max (clippedCoherentInformation sigma - vnEntropy tau) 0 ∧
    clippedCoherentInformation (productFactor tau sigma) <
      clippedCoherentInformation (discardFirstAlice (productFactor tau sigma)) := by sorry

end

end QAlgorithms.Papers.Independent.Support50.P39
Source
Classical capacity of a noiseless quantum channel assisted by noisy entanglement, arXiv:quant-ph/0106080v3, PDF p. 6, Section 2, example due to Bennett
Read-back

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

Let T,A,BT,A,BT,A,B be finite, nonempty types with decidable equality. Let τ\tauτ be a positive semidefinite complex matrix on TTT with trace 111, and let σ\sigmaσ be a positive semidefinite complex matrix on A×BA\times BA×B with trace 111. For any matrix XXX, write S(X)=−Re⁡tr⁡(f(X))S(X)=-\operatorname{Re}\operatorname{tr}(f(X))S(X)=−Retr(f(X)), where f(t)=tlog⁡2tf(t)=t\log_2tf(t)=tlog2​t is applied by continuous functional calculus, and, for a bipartite matrix, write XBX_BXB​ for the partial trace over its Alice subsystem and J(X)=max⁡{S(XB)−S(X),0}J(X)=\max\{S(X_B)-S(X),0\}J(X)=max{S(XB​)−S(X),0}. Assume S(τ)>0S(\tau)>0S(τ)>0 and J(σ)>0J(\sigma)>0J(σ)>0. Define the matrix ω\omegaω on (T×A)×B(T\times A)\times B(T×A)×B by

ω((t,a),b),((t′,a′),b′)=τt,t′σ(a,b),(a′,b′),\omega_{((t,a),b),((t',a'),b')}=\tau_{t,t'}\sigma_{(a,b),(a',b')},ω((t,a),b),((t′,a′),b′)​=τt,t′​σ(a,b),(a′,b′)​,

and define the result of discarding the TTT factor by

D(ω)(a,b),(a′,b′)=∑t∈Tω((t,a),b),((t,a′),b′).D(\omega)_{(a,b),(a',b')}=\sum_{t\in T}\omega_{((t,a),b),((t,a'),b')}.D(ω)(a,b),(a′,b′)​=t∈T∑​ω((t,a),b),((t,a′),b′)​.

The declaration asserts all three conclusions

D(ω)=σ,J(ω)=max⁡{J(σ)−S(τ),0},J(ω)<J(D(ω)),D(\omega)=\sigma,\qquad J(\omega)=\max\{J(\sigma)-S(\tau),0\},\qquad J(\omega)<J(D(\omega)),D(ω)=σ,J(ω)=max{J(σ)−S(τ),0},J(ω)<J(D(ω)),

where the Alice subsystem for J(ω)J(\omega)J(ω) is T×AT\times AT×A, and that for J(D(ω))J(D(\omega))J(D(ω)) is AAA.

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