Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Stochastic preprocessing strictly improves the Figure 2 private-information expression

Open
QAlgorithms.Papers.Independent.Intakenext100.P30.stochastic_preprocessing_example

by Gohan · Oct 31, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

Consider the classical channel on X={0,1,2}X=\{0,1,2\}X={0,1,2} with deterministic receiver outputs Y(0)=Y(1)=0Y(0)=Y(1)=0Y(0)=Y(1)=0, Y(2)=1Y(2)=1Y(2)=1, and eavesdropper outputs Z(0)=0Z(0)=0Z(0)=0, Z(1)=1Z(1)=1Z(1)=1, with Z(2)Z(2)Z(2) a fair bit. Encode message 000 uniformly as X=0X=0X=0 or X=1X=1X=1, and message 111 as X=2X=2X=2. Both conditional encoding laws are normalized and nonnegative. For each message mmm, the receiver output is exactly mmm and Eve's output is the fair distribution, so this transmits one private bit without error. Moreover,

sup⁡PX[H(X∣Z)−H(X∣Y)]<1.\sup_{P_X}\bigl[H(X|Z)-H(X|Y)\bigr]<1.PX​sup​[H(X∣Z)−H(X∣Y)]<1.

The strict inequality concerns the maximum over the full input simplex, not merely pointwise inequality for each law. This example establishes the need to optimize stochastic preprocessing in the private-information expression.

All alphabets are finite and nonempty, and output Hilbert spaces are finite dimensional and nonzero. A classical–quantum channel assigns a positive semidefinite, trace-one matrix to each input. Encoders may use private randomness; decoders are normalized POVMs. Codes are fixed independently of the message, with no supplied shared key or entanglement. Logarithms and entropies are in bits.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P30
Formal statement
noncomputable section

open scoped BigOperators ComplexConjugate Matrix ComplexOrder

open Filter

namespace QAlgorithms.Papers.Independent.Intakenext100.P30

variable {X Y Z M U V A B C : Type*}

variable [Fintype X] [DecidableEq X] [Fintype Y] [DecidableEq Y]

variable [Fintype Z] [DecidableEq Z] [Fintype M] [DecidableEq M]

variable [Fintype U] [DecidableEq U] [Fintype V] [DecidableEq V]

variable [Fintype A] [DecidableEq A] [Fintype B] [DecidableEq B]

variable [Fintype C] [DecidableEq C]

attribute [local instance] Classical.propDecidable

/-- Figure 2 and following paragraph, PDF p.6: the concrete stochastic encoder
transmits a bit perfectly with identical fair Eve outputs, whereas the optimized
unpreprocessed private information is strictly smaller than one bit. No new
operational notion of deterministic-encoding capacity is asserted. -/
theorem stochastic_preprocessing_example :
    (∀ message : Bool,
      (∑ x : Fin 3, rrExampleEncoder message x = 1) ∧
      (∀ x, 0 ≤ rrExampleEncoder message x) ∧
      (∑ x : Fin 3, (rrExampleEncoder message x : ℂ) • rrExampleReceiver x) =
        Matrix.diagonal (fun y => if y = message then (1 : ℂ) else 0) ∧
      (∑ x : Fin 3, (rrExampleEncoder message x : ℂ) • rrExampleEve x) =
        Matrix.diagonal (fun _ => (1 / 2 : ℂ))) ∧
    rrExampleUnprocessedInformation < (1 : EReal) := by sorry

end QAlgorithms.Papers.Independent.Intakenext100.P30

end
Source
Noisy channel coding via privacy amplification and information reconciliation, arXiv:1012.4814v2, PDF p. 6, §IV Figure 2: stochastic preprocessing improves the private-information expression
Read-back

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

On the input alphabet {0,1,2}\{0,1,2\}{0,1,2} and Boolean output basis ordered as (false,true)(\mathrm{false},\mathrm{true})(false,true), define receiver matrices R0=R1=diag⁡(1,0)R_0=R_1=\operatorname{diag}(1,0)R0​=R1​=diag(1,0), R2=diag⁡(0,1)R_2=\operatorname{diag}(0,1)R2​=diag(0,1), and second-output matrices E0=diag⁡(1,0)E_0=\operatorname{diag}(1,0)E0​=diag(1,0), E1=diag⁡(0,1)E_1=\operatorname{diag}(0,1)E1​=diag(0,1), E2=12I2E_2=\tfrac12I_2E2​=21​I2​. Define encoding weights qfalse=(12,12,0)q_{\mathrm{false}}=(\tfrac12,\tfrac12,0)qfalse​=(21​,21​,0) and qtrue=(0,0,1)q_{\mathrm{true}}=(0,0,1)qtrue​=(0,0,1). For every Boolean message bbb, these weights satisfy ∑x=02qb(x)=1\sum_{x=0}^{2}q_b(x)=1∑x=02​qb​(x)=1, qb(x)≥0q_b(x)\ge0qb​(x)≥0 for every xxx, and

∑x=02qb(x)Rx=∣b⟩⟨b∣,∑x=02qb(x)Ex=12I2.\sum_{x=0}^{2}q_b(x)R_x=|b\rangle\langle b|, \qquad \sum_{x=0}^{2}q_b(x)E_x=\tfrac12I_2.x=0∑2​qb​(x)Rx​=∣b⟩⟨b∣,x=0∑2​qb​(x)Ex​=21​I2​.

In addition, for every probability vector p=(p0,p1,p2)p=(p_0,p_1,p_2)p=(p0​,p1​,p2​), let ωpE=∑xpx∣x⟩⟨x∣⊗Ex\omega^E_p=\sum_xp_x|x\rangle\langle x|\otimes E_xωpE​=∑x​px​∣x⟩⟨x∣⊗Ex​ and ωpR=∑xpx∣x⟩⟨x∣⊗Rx\omega^R_p=\sum_xp_x|x\rangle\langle x|\otimes R_xωpR​=∑x​px​∣x⟩⟨x∣⊗Rx​, and write S(X∣B)ω=S(ω)−S(tr⁡Xω)S(X\mid B)_\omega=S(\omega)-S(\operatorname{tr}_X\omega)S(X∣B)ω​=S(ω)−S(trX​ω), where (tr⁡Xω)bb′=∑xω(x,b),(x,b′)(\operatorname{tr}_X\omega)_{bb'}=\sum_x\omega_{(x,b),(x,b')}(trX​ω)bb′​=∑x​ω(x,b),(x,b′)​ and S(ρ)=−Re⁡tr⁡[g(ρ)]S(\rho)=-\operatorname{Re}\operatorname{tr}[g(\rho)]S(ρ)=−Retr[g(ρ)] for g(t)=tlog⁡2tg(t)=t\log_2tg(t)=tlog2​t applied by continuous functional calculus. The extended-real supremum satisfies

sup⁡px≥0p0+p1+p2=1[S(X∣E)ωpE−S(X∣R)ωpR]<1.\sup_{\substack{p_x\ge0\\p_0+p_1+p_2=1}} \left[S(X\mid E)_{\omega^E_p}-S(X\mid R)_{\omega^R_p}\right]<1.px​≥0p0​+p1​+p2​=1​sup​[S(X∣E)ωpE​​−S(X∣R)ωpR​​]<1.

Probability vectors with zero coordinates are included.

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