AFTD Auto-Formalizing Theoretical Domains

op_1693_noon_witness_symm

theoremverifiedformalization

For every natural number N, balance parameter q in Real, and transmissivities x, y in Real, the realignment witness satisfies noon_realign_witness N (1 - q) y x = noon_realign_witness N q x y.

Statement

theorem op_1693_noon_witness_symm (N : ℕ) (q x y : ℝ) :
    noon_realign_witness N (1 - q) y x = noon_realign_witness N q x y
Formalizes
arXiv:2611.17021
Verified
15 Nov 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 1
  • noon_realign_witnessFor any natural number N, real balance parameter q, and real transmissivities x, y, the second-moment realignment witness W_N(q, x, y) is…

Read back from the Lean

For every natural number $N$ and all real numbers $q$, $x$, and $y$, noon_realign_witness N (1 - q) y x equals noon_realign_witness N q x y.

Written by a model that saw only the Lean, never the English above. If the two disagree, that is worth a challenge.

The proof, explained

The theorem asserts that the NOON realignment witness is symmetric under simultaneously swapping the balance parameter $q \leftrightarrow 1 - q$ and the transmissivities $x \leftrightarrow y$. Unfolding the definition reveals that the witness consists of a sum of squares and an isolated boundary product term. The boundary terms match by commutativity of multiplication and the relation $1 - (1 - q) = q$, and the sums match term-by-term because swapping $q \leftrightarrow 1 - q$ and $x \leftrightarrow y$ simply transposes the two terms inside each square.

  1. The definition of the realignment witness is unfolded on both sides of the equation.

    Unfolding `noon_realign_witness` expresses each side as a sum of squares over $k \in \{0, \dots, N\}$ plus an additional product term of the form $4 c (1 - c) (u v)^N$.

    line 4

      unfold noon_realign_witness
  2. The additional product terms on both sides are equal: $4 (1 - q)(1 - (1 - q))(y x)^N = 4 q (1 - q)(x y)^N$.

    Rewriting with commutativity of multiplication (`mul_comm`) changes $(y x)^N$ to $(x y)^N$. Ring arithmetic (`ring`) then simplifies the prefactor, using $1 - (1 - q) = q$ and commutativity to establish the identity.

    lines 5–7

    uses mul_comm

      have h1 : 4 * (1 - q) * (1 - (1 - q)) * (y * x) ^ N = 4 * q * (1 - q) * (x * y) ^ N := by
        rw [mul_comm y x]
        ring
  3. The equality reduces to showing that the remaining sums are equal.

    Substituting the equality of the product terms (`rw [h1]`) matches the second terms on both sides. Congruence (`congr 1`) then cancels them, leaving only the equality between the two sums.

    lines 8–9

      rw [h1]
      congr 1
  4. The two sums are equal term-by-term, completing the proof.

    By `Finset.sum_congr`, it suffices to check that the summands coincide for each $k \in \{0, \dots, N\}$. For any index $k$, the summand on the left is $((1 - q) A + (1 - (1 - q)) B)^2$ and on the right is $(q B + (1 - q) A)^2$, where $A = \binom{N}{k} y^k (1 - y)^{N - k}$ and $B = \binom{N}{k} x^k (1 - x)^{N - k}$. Since $1 - (1 - q) = q$ and addition is commutative, polynomial expansion (`ring`) shows that the two expressions are identical.

    lines 10–12

    uses Finset.sum_congr

      apply Finset.sum_congr rfl
      intro k hk
      ring
mul_comm
Multiplication is commutative: $a \cdot b = b \cdot a$.
Finset.sum_congr
Two sums over finite sets are equal if the index sets are equal and their summands agree at every element.

Written by a model (gemini-3.8-flash) from the Lean proof and the goal states Lean reported at each step. Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.

Lean source view module on GitHub

/-- Symmetry of the NOON realignment witness under exchange of modes and balance parameter. -/theorem op_1693_noon_witness_symm (N : ℕ) (q x y : ℝ) :    noon_realign_witness N (1 - q) y x = noon_realign_witness N q x y := by  unfold noon_realign_witness  have h1 : 4 * (1 - q) * (1 - (1 - q)) * (y * x) ^ N = 4 * q * (1 - q) * (x * y) ^ N := by    rw [mul_comm y x]    ring  rw [h1]  congr 1  apply Finset.sum_congr rfl  intro k hk  ring
See it in the graphDiscuss this resultChallenge it