proofs
Provedalphabaguettebaguette-zquantizer

Values already on the grid are stored exactly: Q(m/2ᵏ) = m

For all k∈N\htmlData{mx=1}{k \in \htmlData{mx=0}{\mathbb{N}}} and m∈Z\htmlData{mx=1}{m \in \htmlData{mx=0}{\mathbb{Z}}}:

⌊m2k⋅2k⌋=m\htmlData{mx=8}{\left\lfloor \htmlData{mx=7}{\htmlData{mx=4}{\frac{\htmlData{mx=1}{\htmlData{mx=0}{m}}}{\htmlData{mx=3}{2^{\htmlData{mx=2}{k}}}}} \cdot \htmlData{mx=6}{2^{\htmlData{mx=5}{k}}}} \right\rfloor} \mathrel{\htmlData{mx=10}{=}} \htmlData{mx=9}{m}
What the notation means · or point at any symbol
NNatural numbers
The natural numbers 0, 1, 2, …. In Lean they start at 0. (ℕ)
ZIntegers
The integers …, −2, −1, 0, 1, 2, …. (ℤ)
⌊x⌋Floor
The largest integer not above m/2^k · 2^k. (⌊((m : ℝ) / 2 ^ k) * 2 ^ k⌋)
=Equality
⌊m/2^k · 2^k⌋ and m are the same. (⌊((m : ℝ) / 2 ^ k) * 2 ^ k⌋ = m)
·Multiplication
m/2^k times 2^k. (((m : ℝ) / 2 ^ k) * 2 ^ k)
(x : T)As real numbers
m read as an element of real numbers. Mathematicians leave this implicit; Lean needs it to know which kind of number is meant. ((m : ℝ))
a/bDivision
m divided by 2^k. In Lean, dividing by 0 gives 0. ((m : ℝ) / 2 ^ k)
aᵇPower
2 raised to the power k. (2 ^ k)

Re-quantizing a grid value is the identity, so integer state round-trips through the real-valued view without loss. Part of the plan 'Baguette-Z: closed-loop certified integer-state training'.

How it was proved

Proved
3 October 2026
By
Ajax, with the agent “Baguette math & proofs (Claude Code, 2026-10-03)”
Checked by
Lean's kernel (Lean 4.33.1, with Mathlib); the proof is signed and replays offline
Axioms
Classical.choice, Quot.sound, propext
Proof
13 lines of Lean
Sign in to see the proofThe proof, its signed report and its history.

More proved statements