Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 1.1 — efficient perfectly correct encryption with the exact unclonable bound

Proved
PauliUnclonable531.Pauli.headline_encryption

by mikedeng1 · Nov 8, 2026 · Mathlib 0df444a (Lean v4.33.1)

cryptographyp2o-batch-stoc27ap2o-gran-per-chapterp2o-plan-paperp2o-v1quantum-information

There exists the explicit Section 3 family of one-time private-key schemes for one-bit messages. For every n≥1 its key is exactly uniform on the 4ⁿ−1 nonidentity n-qubit Pauli labels, encoded injectively by 2n classical bits, and its ciphertext is an n-qubit density operator. Its decryption effects form a valid binary POVM, with perfect correctness for every supported key and message.

The generated Section 3.1 encryption instructions have their actual H,S,X matrix semantics and realize the prescribed ciphertext ensemble. They use exactly n−1 independent fair bits, n computational-basis preparations, and at most 2n H/S gates; each basis preparation uses at most one X on a fresh zero qubit. The pivot is the least nonidentity position. These single-qubit gates are unitary and normalize the Pauli family up to sign, giving the asserted Clifford implementation. Generated decryption uses at most n local Pauli measurements and its Born outcome law equals the global two-projector measurement for every input density operator, including entangled inputs. Under the declared address-bit cost model both classical costs are bounded uniformly by 100(n+1)².

Against every finite-dimensional information-theoretic attack, including randomized challenge-message selection with associated auxiliary preparation, the full game satisfies

Pr⁡[bB=bC=b]≤12+122n4n−1.\Pr[b_B=b_C=b]\le\frac12+\frac12\sqrt{\frac{2^n}{4^n-1}}.Pr[bB​=bC​=b]≤21​+21​4n−12n​​.

Correctness, physical validity, operational semantics, resource counts and security are conclusions for the constructed family. None is assumed in a scheme or attack record.

Differences from the source:

  • encoding: finite-dimensional complex Hilbert-space operators and binary strings → complex matrices indexed by finite computational bases, Fin 2 bits, real Born probabilities, Euclidean spectral norm and PSD-difference order; reason: these are the same finite-dimensional objects and conventions after choosing the source basis
  • encoding: efficient encryption and decryption by the p.16–17 instructions → the concrete uniform instruction generators and Born interpreters, with exact draw/preparation/measurement counts and both address-bit costs at most 100(n+1)²; reason: 100 bounds the charged lengths of these particular procedures in the declared bit model; it is independent of n, key, message and attack, and changes no security constant
  • encoding: the general one-bit message-selection game → a distribution on four classical message pairs and arbitrary conditional CPTP/local-POVM attacks; reason: conditioning absorbs the auxiliary preparation exactly as in Remark 2.12
Preamble
import Mathlib
import Definitions.Def_PauliUnclonable531_Pauli_Game
import Definitions.Def_PauliUnclonable531_Pauli_Ensemble

open scoped Matrix Matrix.Norms.L2Operator ComplexOrder
set_option autoImplicit false
Formal statement
namespace PauliUnclonable531.Pauli

/-- Full Theorem 1.1, p.3, with the concrete §3.1 implementation. The fixed coefficient
100 is a conservative bookkeeping bound for the declared address-bit cost of the actual
lists; it is not a security constant. H,S,X have their actual matrices, and local
measurement semantics include entangled inputs. -/
theorem headline_encryption :
    ∃ family : (n : ℕ) → Scheme n, family = explicitScheme ∧
      (∀ g : Gate, (gateMatrix g).conjTranspose * gateMatrix g = 1 ∧
        ∀ p : Fin 4, ∃ q : Fin 4, ∃ b : Fin 2,
          gateMatrix g * QMaxCutSDP.Rounding.pauli p * (gateMatrix g).conjTranspose =
            sign b • QMaxCutSDP.Rounding.pauli q) ∧
      ∀ n : ℕ, 1 ≤ n →
        Fintype.card (Key n) = 4^n - 1 ∧
        Function.Injective (@keyEncode n) ∧
        Fintype.card (Fin (n+n)) = 2*n ∧
        Fintype.card (Bits n) = 2^n ∧
        (∀ k, (family n).gen k = (Fintype.card (Key n) : ℝ)⁻¹ ∧ 0 < (family n).gen k) ∧
        (∑ k : Key n, (family n).gen k) = 1 ∧
        (∀ k m, Density ((family n).enc k m)) ∧
        (∀ k, (∀ m, ((family n).dec k m).PosSemidef) ∧ (∑ m, (family n).dec k m) = 1) ∧
        (∀ k m, born ((family n).dec k m) ((family n).enc k m) = 1) ∧
        (∀ k m, operationalEnc k m = (family n).enc k m) ∧
        (∀ k : Key n, (∃ j, pivot k = some j ∧ active k.val j = true ∧
                    ∀ i : Fin n, i < j → active k.val i = false) ∧
          Fintype.card (Tape k) = 2^(n-1) ∧
          fairCount (encProgram k) = n-1 ∧ preparationCount (encProgram k) = n ∧
          gateCount (encProgram k) ≤ 2*n ∧ encBitCost k ≤ 100*(n+1)^2) ∧
        (∀ k (σ : Op (Bits n)), Density σ → ∀ b : Fin 2,
          operationalDec k σ b = born ((family n).dec k b) σ) ∧
        (∀ k : Key n, (decProgram k).length ≤ n ∧ decBitCost k ≤ 100*(n+1)^2) ∧
        (∀ (B C : Type) [Fintype B] [Fintype C] [DecidableEq B] [DecidableEq C]
            [Nonempty B] [Nonempty C] (a : GeneralAttack n B C),
          generalWin (family n) a ≤ 1/2 + (1/2) * Real.sqrt ((2 : ℝ)^n / ((4 : ℝ)^n - 1))) := by sorry

end PauliUnclonable531.Pauli
Source
Prabhanjan Ananth, Seyoon Ragavan, Amit Sahai, STOC 2027 submission #531 (October 2026), Theorem 1.1 (Informal), p.3; Definitions 2.9–2.11 and Remark 2.12, pp.14–15; §3.1, pp.16–17; Theorem 4.1, p.17
Human review
  • Endorsed by Shuze Chen · Nov 8, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Nov 8, 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