Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Plaintext evolution of bootstrapped evaluation: after `k` blocks the

Proved
FHENoise.bootIter_plain

by raver1975 · Sep 17, 2026 · Mathlib c5ea003 (Lean v4.30.0)

aether-catalogcryptography

Plaintext evolution of bootstrapped evaluation: after k blocks the plaintext has been raised to the power 2^(L·k).

theorem FHENoise.bootIter_plain{M : Type*} [CommRing M] {D T Bref : ℝ} (G : NoiseGauge R) (s : R)
    (pi : R →+* M) (relin refresh : Cipher R → Cipher R)
    (hrelin : ∀ c, G.nu (phase s (relin c) - phase s c) ≤ D) (hD : 0 ≤ D)
    (hrelinP : ∀ c, pi (phase s (relin c)) = pi (phase s c))
    (hrefN : ∀ c, noise G s c < T → noise G s (refresh c) ≤ Bref)
    (hrefP : ∀ c, noise G s c < T → pi (phase s (refresh c)) = pi (phase s c))
    (hBref : 0 ≤ Bref) (L : ℕ) (hsafe : iterD G.gamma D L Bref < T)
    (c : Cipher R) (hc : noise G s c ≤ Bref) :
    ∀ k, pi (phase s (bootIter relin refresh L k c)) = (pi (phase s c)) ^ (2 ^ (L * k))
  := by sorry

Formalization Note Transplanted verbatim from the Aether Catalog source Cryptography/FHE/Bootstrapping.lean; the statement is byte-identical to the source declaration, elaborated with autoImplicit disabled in the platform environment.

Preamble
-- Thm stub generated from Cryptography/FHE/Bootstrapping.lean
import Mathlib
import Definitions.Def_Cryptography_FHE_Bootstrapping
import Definitions.Def_Cryptography_FHE_NoiseDichotomy
import Definitions.Def_Cryptography_FHE_NoiseGauge
import Definitions.Def_Cryptography_FHE_NoiseGrowth

/-!
# Quantitative bootstrapping: unbounded depth at bounded noise

`Cryptography.FHE.RingLWE` states Gentry's bootstrapping principle in an
idealized form: a refresh operation that is *always* correct yields unbounded
depth.  Real bootstrapping is not unconditional — the refresh procedure is
itself a homomorphic evaluation of the decryption circuit, so it only works when
its input is still decryptable, i.e. when the input noise is below the decoding
radius `T`.  This file formalizes exactly that conditional statement and derives
the two quantities a parameter designer actually needs:

* how many multiplication levels `L` fit between two bootstraps
  (`levels_between_bootstraps`, a logarithmic formula), and
* the fact that the *bootstrapped* evaluation of an arbitrarily deep squaring
  chain keeps noise `≤ B_ref` at every stage and decrypts correctly
  (`bootIter_decrypt`).

We also record how modulus switching moves the stability threshold of the
dichotomy of `NoiseDichotomy` (`modSwitch_dichotomy`): dividing the noise by `p`
at each level multiplies the tolerable key-switching noise by `p`.
-/

open FHENoise

-- open removed: section is not a namespace

noncomputable section

variable {R : Type*} [CommRing R]

/-! ## 1. Blocks of squarings between refreshes -/






/-! ## 2. How many levels fit between two bootstraps -/


/-! ## 3. Bootstrapped evaluation of unboundedly deep chains -/
Formal statement
theorem FHENoise.bootIter_plain{M : Type*} [CommRing M] {D T Bref : ℝ} (G : NoiseGauge R) (s : R)
    (pi : R →+* M) (relin refresh : Cipher R → Cipher R)
    (hrelin : ∀ c, G.nu (phase s (relin c) - phase s c) ≤ D) (hD : 0 ≤ D)
    (hrelinP : ∀ c, pi (phase s (relin c)) = pi (phase s c))
    (hrefN : ∀ c, noise G s c < T → noise G s (refresh c) ≤ Bref)
    (hrefP : ∀ c, noise G s c < T → pi (phase s (refresh c)) = pi (phase s c))
    (hBref : 0 ≤ Bref) (L : ℕ) (hsafe : iterD G.gamma D L Bref < T)
    (c : Cipher R) (hc : noise G s c ≤ Bref) :
    ∀ k, pi (phase s (bootIter relin refresh L k c)) = (pi (phase s c)) ^ (2 ^ (L * k))
  := by sorry
Source
https://github.com/paulklemstine/Lean/blob/53c2925a02/Catalog/Cryptography/FHE/Bootstrapping.lean#L130

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