AFTD Auto-Formalizing Theoretical Domains

range_criterion_subspaces_sporadic_five_five_three_five

theoremverifiedformalization

There exist complex subspaces D and E in C^5 ⊗ C^5 of codimensions 3 and 5, respectively, satisfying the range criterion (no nonzero product vector has its tensor in D and its partial conjugate in E).

Statement

theorem range_criterion_subspaces_sporadic_five_five_three_five :
    has_range_criterion_subspaces 5 5 3 5
Formalizes
arXiv:2611.08518 Sec. 1
Verified
09 Nov 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 3

Read back from the Lean

There exist $\mathbb{C}$-submodules (subspaces) $D$ and $E$ of $(\mathbb{C}^5) \otimes_{\mathbb{C}} (\mathbb{C}^5)$ such that: 1. The $\mathbb{C}$-codimension of $D$ (i.e., the dimension of the quotient space $((\mathbb{C}^5) \otimes_{\mathbb{C}} (\mathbb{C}^5)) / D$) is equal to 3; 2. The $\mathbb{C}$-codimension of $E$ (i.e., the dimension of the quotient space $((\mathbb{C}^5) \otimes_{\mathbb{C}} (\mathbb{C}^5)) / E$) is equal to 5; and 3. For all vectors $x \in \mathbb{C}^5$ and $y \in \mathbb{C}^5$, if the elementary tensor $x \otimes y$ belongs to $D$ and $(\operatorname{star} x) \otimes y$ belongs to $E$ (where $\operatorname{star} x$ is the entrywise complex conjugate of $x$), then $x = 0$ or $y = 0$.

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

To construct range criterion subspaces of codimensions 3 and 5 in C^5 (x) C^5, it suffices by linear duality to exhibit two linearly independent matrix pencils A (size 3) and B (size 5) in C^(5x5) that share no nontrivial bilinear zeros (x^T A_a y = 0 and x* B_j y = 0 implies x = 0 or y = 0). Pencils A and B with Gaussian integer entries are explicitly specified, and their linear independence is certified by dual coordinate inverse matrices verified via kernel computation. The nonexistence of nontrivial common zeros is proved using a Sum-of-Squares (SOS) Positivstellensatz certificate: a Hermitian tensor H formed from the pencils evaluates to -sum_k |cterm(k)|^2 <= 0 on rank-one vectors v = x* y^T, while an approximate Cholesky decomposition 16 H = 3200 I + M M* + E with kernel-verified error bounds shows that the same form is at least 1980 ||v||^2, forcing v = 0 and hence x = 0 or y = 0.

  1. The certificate matrices over Gaussian integers are defined and the row and column norm bounds on the error matrix E are verified. automation

    We define explicit matrices A (3 of size 5x5), B (5 of size 5x5), C (14 of size 5x5), and M (25x25) over Gaussian integers, together with the tensor H and error matrix E = 16 H - 3200 I - M M*. The boolean test r5chk verifies that every row and column of E has integer l1-norm at most 1220; this is checked by kernel evaluation.

    lines 1–85

    abbrev r5ZI := GaussianInt
    def r5dA : List (List (ℤ × ℤ)) := [
      [(-7,23),(-8,5),(-4,-2),(-3,0),(12,14),(-2,-7),(17,6),(-2,11),(3,18),(5,-8),(0,-12),(-23,4),(-4,-9),(14,-1),(14,12),(10,-7),(-2,-5),(-13,3),(-5,1),(-15,9),(-8,-8),(-13,14),(16,4),(0,0),(-9,-14)],
      [(-1,-8),(-9,-5),(11,10),(-25,-3),(2,-18),(31,-1),(-8,7),(15,13),(13,-1),(3,1),(0,-5),(-11,3),(15,5),(-6,5),(2,21),(-13,-5),(-19,14),(-18,1),(15,1),(-11,-7),(-2,-8),(7,-8),(-5,16),(-16,3),(-2,-4)],
      [(4,5),(-14,2),(-21,-15),(1,-3),(-16,-5),(1,5),(9,-9),(1,-11),(3,-24),(-1,-12),(-15,16),(-13,-2),(12,5),(-10,-10),(5,15),(6,7),(-11,8),(10,-5),(-16,-7),(-14,-9),(9,-10),(3,-3),(8,2),(4,-20),(-4,7)]]
    def r5dB : List (List (ℤ × ℤ)) := [
      [(-7,7),(6,12),(11,8),(10,3),(5,-1),(-8,-8),(-9,9),(-9,11),(9,-8),(-9,8),(-8,-1),(2,-11),(-3,-12),(-12,0),(7,11),(-7,7),(-6,-8),(-9,-7),(-12,-8),(6,9),(-6,-9),(-10,-8),(-7,10),(2,-11),(-8,9)],
      [(-6,11),(11,4),(-12,-1),(-11,11),(-4,-6),(10,-7),(6,-2),(11,-7),(-7,10),(-7,-11),(0,13),(1,-14),(-2,-7),(1,-5),(12,-1),(13,1),(-5,6),(11,2),(12,-4),(10,1),(-6,-8),(-11,3),(-8,-9),(11,-1),(-1,13)],
      [(3,11),(-10,9),(-1,11),(12,-4),(-1,-9),(7,8),(10,5),(1,-10),(11,7),(-5,14),(-7,6),(0,1),(9,-5),(15,-3),(-6,-6),(7,8),(5,5),(-10,0),(10,-1),(12,3),(17,3),(1,11),(-8,8),(-10,6),(-9,4)],
      [(-6,-5),(10,-7),(-14,2),(6,6),(-5,-8),(-11,-3),(12,0),(3,12),(-1,-9),(-8,-7),(10,-4),(-2,12),(-11,6),(-14,-5),(-12,6),(8,10),(8,-11),(-12,6),(12,2),(-7,-7),(3,-6),(2,5),(-11,-5),(-2,5),(-9,9)],
      [(-5,-12),(-4,8),(1,-10),(13,-3),(9,-8),(2,-5),(6,-11),(-3,9),(-9,-5),(7,10),(5,-13),(-7,-11),(6,5),(-9,-8),(5,-10),(11,12),(-5,14),(0,-12),(9,-3),(1,12),(-8,-10),(10,8),(-10,1),(7,1),(4,-12)]]
    def r5dC : List (List (ℤ × ℤ)) := [
      [(0,-6),(-2,7),(-2,-9),(8,6),(4,-6),(1,-2),(3,-2),(-3,-11),(11,4),(-1,-2),(6,-2),(-8,3),(-5,1),(-3,8),(-2,-7),(-5,7),(-8,17),(0,7),(-9,-8),(-5,4),(-2,2),(3,-2),(-4,-3),(-3,6),(-5,-2)],
      [(5,3),(6,2),(5,11),(2,-1),(1,11),(-8,-7),(2,7),(-6,-3),(-6,5),(0,0),(-8,1),(3,0),(-1,-7),(1,9),(-8,5),(-2,2),(1,4),(4,-4),(-1,6),(6,1),(1,4),(-4,-3),(-3,11),(-1,4),(1,-10)],
      [(2,5),(1,11),(8,0),(1,0),(5,-4),(-1,-4),(-8,-8),(-2,-10),(0,-2),(-1,-4),(-2,4),(8,4),(3,-4),(4,-5),(-6,-2),(-4,-4),(0,4),(1,-2),(7,3),(5,4),(2,-1),(1,-2),(-4,-3),(-5,-2),(0,-4)],
      [(0,3),(-5,1),(6,1),(6,-3),(-2,-1),(-7,-5),(4,1),(4,2),(0,2),(-4,-6),(-1,-4),(1,-3),(-4,4),(5,9),(-2,8),(2,1),(-2,2),(11,-6),(-7,6),(3,-5),(-6,-8),(1,-3),(5,4),(-3,-2),(2,1)],
      [(0,2),(1,0),(5,0),(-2,6),(2,-3),(9,-8),(-6,2),(2,0),(3,-13),(2,-2),(1,6),(-2,-5),(-1,-5),(-3,-1),(-7,-3),(-2,4),(-3,-2),(0,-1),(0,-5),(5,0),(4,-9),(6,-5),(-3,0),(-5,0),(-1,9)],
      [(-10,-7),(-4,2),(-7,6),(0,2),(-6,-5),(-7,-3),(7,1),(-4,-2),(-1,1),(5,-1),(-2,-3),(-4,0),(2,2),(4,5),(-7,-3),(13,3),(-4,4),(8,0),(-3,5),(0,0),(-4,1),(-6,4),(4,-2),(5,6),(3,5)],
      [(-2,-7),(3,-4),(6,-5),(2,-9),(-1,-7),(-12,6),(1,3),(-2,-1),(-6,-3),(-2,5),(0,1),(-5,-5),(-4,4),(3,0),(8,6),(-1,3),(2,-1),(-3,-2),(0,-1),(4,-11),(-5,-3),(-11,-2),(9,2),(-4,1),(0,4)],
      [(-1,-9),(-2,-4),(0,5),(5,-4),(5,-1),(-3,2),(9,-3),(7,-1),(4,5),(8,3),(-2,-2),(2,1),(1,3),(0,5),(-2,-3),(4,5),(4,-2),(3,4),(-2,6),(8,6),(4,-5),(-3,3),(6,5),(7,-1),(-5,-4)],
      [(-1,7),(-1,5),(8,3),(-9,2),(-2,5),(7,5),(-1,1),(2,9),(12,0),(-4,1),(-7,4),(3,-7),(1,2),(6,0),(4,-5),(-8,2),(-4,3),(6,-10),(1,3),(-7,1),(6,0),(4,3),(-2,-4),(-8,-1),(-1,5)],
      [(-5,3),(5,3),(7,9),(0,-2),(-2,8),(3,1),(-2,0),(-3,-8),(2,0),(7,-2),(-6,4),(-2,1),(2,-10),(-5,5),(-9,-4),(-6,-1),(-1,3),(-1,-1),(2,-1),(8,1),(3,2),(1,3),(-5,2),(0,-1),(-2,-6)],
      [(5,0),(-4,-1),(-3,2),(3,-9),(-2,4),(-1,3),(3,1),(6,6),(3,6),(-4,6),(-6,-7),(3,4),(1,7),(3,5),(10,-1),(-5,1),(4,-2),(6,-1),(-8,10),(-5,2),(1,1),(-1,-4),(3,2),(0,-2),(-1,-3)],
      [(-2,7),(2,-5),(-5,-7),(2,-6),(-1,-3),(4,0),(-6,-6),(-4,-2),(-12,-6),(0,-3),(6,-4),(2,12),(2,3),(-1,-5),(-1,5),(6,-3),(12,-5),(-5,8),(-1,0),(2,-2),(-2,1),(1,0),(-2,0),(3,-4),(4,-1)],
      [(3,1),(4,3),(-7,-1),(-6,8),(-3,11),(4,-6),(3,7),(-5,2),(6,8),(1,-6),(4,2),(6,-1),(-1,-4),(0,5),(-12,-1),(5,7),(-3,4),(3,-3),(3,3),(-5,6),(-3,4),(4,5),(-8,1),(7,8),(-2,-3)],
      [(-4,-10),(4,-5),(-7,6),(2,-7),(-6,1),(-8,5),(3,6),(-1,5),(7,-1),(-6,4),(-1,-8),(-1,-13),(-3,5),(-2,3),(7,9),(0,-1),(1,-3),(0,-4),(0,-2),(3,-4),(-6,3),(-5,-3),(3,4),(4,-6),(-4,6)]]
    def r5dM : List (List (ℤ × ℤ)) := [
      [(81,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(2,-7),(84,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(9,0),(-3,4),(89,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(1,-4),(-18,2),(-2,-5),(89,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(-7,-3),(2,0),(-13,7),(-6,-3),(86,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(3,1),(-18,4),(8,5),(14,-4),(-4,-4),(89,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(5,2),(-2,-6),(7,-12),(-2,8),(-6,-1),(2,2),(81,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(-1,-2),(-9,8),(-9,0),(14,7),(4,-9),(14,5),(8,-8),(82,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(-13,-4),(-6,-3),(-18,-4),(0,3),(14,5),(5,9),(0,4),(-1,13),(87,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(3,0),(5,-3),(3,6),(-9,13),(-1,-6),(2,1),(3,6),(-4,8),(4,-10),(77,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(-7,6),(1,13),(-6,13),(-2,-11),(10,-20),(17,-10),(-5,-6),(5,-9),(2,-8),(3,10),(82,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(-5,8),(14,0),(-13,14),(-19,1),(13,-8),(-12,-5),(-12,4),(-2,2),(12,-7),(3,8),(1,2),(86,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(4,-11),(-14,-2),(2,-16),(7,3),(-7,7),(-1,2),(-4,5),(-1,-3),(-3,7),(0,-6),(-4,-2),(-5,0),(75,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(1,10),(7,1),(-9,14),(-7,-1),(8,-6),(-9,0),(-2,5),(2,-3),(12,-2),(-5,13),(-5,2),(2,-2),(1,-4),(79,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(-2,-5),(-1,13),(-5,-4),(16,-14),(5,-6),(8,-7),(0,-3),(-2,-3),(4,6),(-12,1),(-5,5),(-8,3),(-2,-8),(6,7),(82,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(-3,2),(13,-3),(-3,6),(-2,-8),(-2,-1),(-6,-3),(-4,-4),(-3,10),(3,2),(-1,4),(-3,-1),(0,1),(-1,-4),(1,1),(2,-7),(78,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(-1,-11),(13,-6),(2,-10),(-9,8),(7,10),(-12,3),(-2,0),(-10,13),(18,1),(-5,-3),(-7,1),(-3,2),(9,-9),(-5,1),(-5,-1),(0,-3),(75,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(-7,-3),(0,-10),(-17,4),(2,0),(12,1),(-4,-2),(-10,-3),(3,2),(10,-2),(-6,6),(10,11),(1,1),(-10,-5),(5,2),(-1,0),(-7,3),(-1,0),(78,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(2,9),(7,-11),(1,13),(-1,11),(0,-13),(-6,5),(-3,1),(2,4),(8,-8),(6,-5),(6,1),(6,-6),(-2,1),(4,-10),(-9,2),(3,0),(-7,-1),(3,-5),(75,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(-4,-11),(-5,-20),(-8,-5),(13,10),(13,-4),(-1,20),(2,-4),(7,18),(11,-2),(2,4),(2,6),(-3,0),(0,2),(-3,-2),(3,12),(3,-5),(3,-5),(4,-1),(3,4),(76,0),(0,0),(0,0),(0,0),(0,0),(0,0)],
      [(10,-5),(4,19),(4,-7),(1,-3),(-8,16),(-2,-22),(-2,4),(-11,0),(-2,3),(-5,-1),(-1,-7),(-5,-7),(7,-2),(-4,1),(0,0),(-1,2),(-8,2),(-3,-5),(-5,-2),(-5,-10),(73,0),(0,0),(0,0),(0,0),(0,0)],
      [(-9,-2),(-10,-1),(-12,-4),(8,-3),(13,4),(5,7),(-6,-2),(9,5),(12,-7),(1,3),(8,3),(5,4),(-5,-3),(-3,-2),(2,3),(-3,-2),(0,-1),(8,-2),(-4,0),(3,2),(-1,-2),(78,0),(0,0),(0,0),(0,0)],
      [(-4,-3),(1,-24),(-1,-1),(-4,23),(10,-5),(-7,18),(7,7),(-3,10),(7,-7),(5,5),(-4,2),(4,0),(-2,5),(1,-3),(-1,1),(6,3),(10,2),(2,14),(0,5),(1,8),(-3,3),(-10,1),(78,0),(0,0),(0,0)],
      [(4,-2),(-2,-1),(-8,-11),(7,-2),(6,18),(8,-3),(7,1),(-5,1),(6,1),(-8,-12),(-3,4),(-4,5),(6,-3),(-3,0),(5,-6),(2,4),(-6,2),(4,-2),(5,-4),(6,-7),(0,3),(5,6),(-4,-3),(73,0),(0,0)],
      [(0,-4),(-2,5),(0,-12),(3,6),(1,1),(4,0),(-2,-2),(-3,-8),(3,11),(-3,4),(-10,-5),(1,0),(8,1),(-3,2),(4,-6),(6,-3),(-3,4),(2,-3),(8,-8),(4,6),(-3,-1),(5,-2),(-2,5),(-8,2),(74,0)]]
    def r5WA : List (List (ℤ × ℤ)) := [
      [(-210,-286),(30,804),(208,-264)],
      [(379,-3),(55,206),(-752,76)],
      [(51,-811),(87,6),(-144,-644)]]
    def r5WB : List (List (ℤ × ℤ)) := [
      [(-154936,-27196),(150438,147826),(-179405,-154381),(-95592,24217),(-40325,-177110)],
      [(-2648,-199174),(209087,-158991),(61822,27502),(-75544,194724),(102740,-195290)],
      [(91874,12808),(-50504,82662),(-275116,-140302),(220702,39504),(162841,43967)],
      [(221285,105097),(-160034,-152028),(-45046,-101434),(-188159,69054),(-209570,-38755)],
      [(139619,177267),(-31003,143159),(82887,162515),(-23530,145829),(25363,-196314)]]
    def r5dDA : r5ZI := ⟨1680,27712⟩
    def r5dDB : r5ZI := ⟨-9293469,-4452693⟩
    def r5cell (l : List (List (ℤ × ℤ))) (i j : ℕ) : r5ZI :=
      let v := (l.getD i []).getD j (0, 0)
      ⟨v.1, v.2⟩
    def r5A (a : Fin 3) (r s : Fin 5) : r5ZI := r5cell r5dA a (5 * r + s)
    def r5B (j : Fin 5) (r s : Fin 5) : r5ZI := r5cell r5dB j (5 * r + s)
    def r5C (k : Fin 14) (r s : Fin 5) : r5ZI := r5cell r5dC k (5 * r + s)
    def r5M (p q : Fin 5) (u : Fin 25) : r5ZI := r5cell r5dM (5 * p + q) u
    def r5star (z : r5ZI) : r5ZI := ⟨z.re, -z.im⟩
    def r5H (p q p' q' : Fin 5) : r5ZI :=
      (∑ a, r5A a p q' * r5star (r5A a p' q)) + (∑ j, r5star (r5B j p q) * r5B j p' q')
        - ∑ k, r5C k p' q * r5star (r5C k p q')
    def r5E (p q p' q' : Fin 5) : r5ZI :=
      (16 : r5ZI) * r5H p q p' q' - (if p = p' ∧ q = q' then (3200 : r5ZI) else 0)
        - ∑ u, r5M p q u * r5star (r5M p' q' u)
    def r5bd (z : r5ZI) : ℕ := z.re.natAbs + z.im.natAbs
    def r5chk : Bool :=
      decide (∀ p q : Fin 5, ∑ p', ∑ q', r5bd (r5E p q p' q') ≤ 1220) &&
      decide (∀ p' q' : Fin 5, ∑ p, ∑ q, r5bd (r5E p q p' q') ≤ 1220)
    set_option maxRecDepth 1000000 in
    theorem r5chk_true : r5chk = true := by decide +kernel
    
  2. The canonical embedding of Gaussian integers into C preserves conjugation and satisfies an l1 norm bound.

    The embedding r5i maps real and imaginary parts to R, intertwines integer conjugation (r5star) with complex conjugation, and bounds the complex norm by the integer l1-norm |Re(z)| + |Im(z)|.

    lines 86–104

    uses Complex.ext, r5i_re, r5i_im, Complex.star_def, Complex.conj_re, Complex.conj_im, Complex.norm_le_abs_re_add_abs_im

    noncomputable def r5i : r5ZI →+* ℂ := GaussianInt.toComplex
    
    lemma r5i_re (z : r5ZI) : (r5i z).re = (z.re : ℝ) := (GaussianInt.intCast_re z).symm
    lemma r5i_im (z : r5ZI) : (r5i z).im = (z.im : ℝ) := (GaussianInt.intCast_im z).symm
    
    lemma r5i_star (z : r5ZI) : r5i (r5star z) = star (r5i z) := by
      apply Complex.ext
      · rw [r5i_re, Complex.star_def, Complex.conj_re, r5i_re]; rfl
      · rw [r5i_im, Complex.star_def, Complex.conj_im, r5i_im]
        show (((-z.im : ℤ)) : ℝ) = _
        push_cast; ring
    
    lemma r5i_norm (z : r5ZI) : ‖r5i z‖ ≤ (r5bd z : ℝ) := by
      have h := Complex.norm_le_abs_re_add_abs_im (r5i z)
      rw [r5i_re, r5i_im] at h
      have hb : (r5bd z : ℝ) = |(z.re : ℝ)| + |(z.im : ℝ)| := by
        simp [r5bd, Nat.cast_add, Nat.cast_natAbs, Int.cast_abs]
      rw [hb]; exact h
    
  3. Multiple summations over Fin 5 can have their summation order permuted.

    Lemmas r5perm1, r5perm2, r5perm3, and r5swap4 establish index permutation and summation interchange identities for 4-fold and 5-fold sums over Fin 5 by repeatedly applying the commutativity of finite sums.

    lines 105–141

    uses Finset.sum_comm, Finset.sum_congr

    lemma r5perm1 (f : Fin 5 → Fin 5 → Fin 5 → Fin 5 → ℂ) :
        (∑ r, ∑ s, ∑ r', ∑ s', f r s r' s') = ∑ p, ∑ q, ∑ p', ∑ q', f p q' p' q := by
      refine Finset.sum_congr rfl fun r _ => ?_
      calc (∑ s, ∑ r', ∑ s', f r s r' s') = ∑ r', ∑ s, ∑ s', f r s r' s' := Finset.sum_comm
        _ = ∑ r', ∑ s', ∑ s, f r s r' s' := Finset.sum_congr rfl fun _ _ => Finset.sum_comm
        _ = ∑ s', ∑ r', ∑ s, f r s r' s' := Finset.sum_comm
    
    lemma r5perm2 (f : Fin 5 → Fin 5 → Fin 5 → Fin 5 → ℂ) :
        (∑ r, ∑ s, ∑ r', ∑ s', f r s r' s') = ∑ p, ∑ q, ∑ p', ∑ q', f p' q' p q := by
      calc (∑ r, ∑ s, ∑ r', ∑ s', f r s r' s') = ∑ r, ∑ r', ∑ s, ∑ s', f r s r' s' :=
            Finset.sum_congr rfl fun _ _ => Finset.sum_comm
        _ = ∑ r', ∑ r, ∑ s, ∑ s', f r s r' s' := Finset.sum_comm
        _ = ∑ r', ∑ r, ∑ s', ∑ s, f r s r' s' :=
            Finset.sum_congr rfl fun _ _ => Finset.sum_congr rfl fun _ _ => Finset.sum_comm
        _ = ∑ r', ∑ s', ∑ r, ∑ s, f r s r' s' :=
            Finset.sum_congr rfl fun _ _ => Finset.sum_comm
    
    lemma r5perm3 (f : Fin 5 → Fin 5 → Fin 5 → Fin 5 → ℂ) :
        (∑ r, ∑ s, ∑ r', ∑ s', f r s r' s') = ∑ p, ∑ q, ∑ p', ∑ q', f p' q p q' := by
      calc (∑ r, ∑ s, ∑ r', ∑ s', f r s r' s') = ∑ r, ∑ r', ∑ s, ∑ s', f r s r' s' :=
            Finset.sum_congr rfl fun _ _ => Finset.sum_comm
        _ = ∑ r', ∑ r, ∑ s, ∑ s', f r s r' s' := Finset.sum_comm
        _ = ∑ r', ∑ s, ∑ r, ∑ s', f r s r' s' :=
            Finset.sum_congr rfl fun _ _ => Finset.sum_comm
    
    lemma r5swap4 {n : ℕ} (h : Fin n → Fin 5 → Fin 5 → Fin 5 → Fin 5 → ℂ) :
        (∑ p, ∑ q, ∑ p', ∑ q', ∑ i, h i p q p' q') = ∑ i, ∑ p, ∑ q, ∑ p', ∑ q', h i p q p' q' := by
      calc (∑ p, ∑ q, ∑ p', ∑ q', ∑ i, h i p q p' q')
          = ∑ p, ∑ q, ∑ p', ∑ i, ∑ q', h i p q p' q' :=
            Finset.sum_congr rfl fun _ _ => Finset.sum_congr rfl fun _ _ =>
              Finset.sum_congr rfl fun _ _ => Finset.sum_comm
        _ = ∑ p, ∑ q, ∑ i, ∑ p', ∑ q', h i p q p' q' :=
            Finset.sum_congr rfl fun _ _ => Finset.sum_congr rfl fun _ _ => Finset.sum_comm
        _ = ∑ p, ∑ i, ∑ q, ∑ p', ∑ q', h i p q p' q' :=
            Finset.sum_congr rfl fun _ _ => Finset.sum_comm
        _ = ∑ i, ∑ p, ∑ q, ∑ p', ∑ q', h i p q p' q' := Finset.sum_comm
    
  4. The definitions of tensors H and E map into C compatibly with ring operations and complex conjugation.

    Because the embedding r5i is a ring homomorphism that commutes with conjugation, the complex images of H and E expand into the expected algebraic expressions over C.

    lines 142–156

    uses map_ofNat, map_sub, map_add, map_sum, map_mul, r5i_star

    lemma r5i_H (p q p' q' : Fin 5) : r5i (r5H p q p' q') =
        (∑ a, r5i (r5A a p q') * star (r5i (r5A a p' q)))
          + (∑ j, star (r5i (r5B j p q)) * r5i (r5B j p' q'))
          - ∑ k, r5i (r5C k p' q) * star (r5i (r5C k p q')) := by
      simp only [r5H, map_sub, map_add, map_sum, map_mul, r5i_star]
    
    lemma r5i_E (p q p' q' : Fin 5) : r5i (r5E p q p' q') =
        16 * r5i (r5H p q p' q') - (if p = p' ∧ q = q' then (3200 : ℂ) else 0)
          - ∑ u, r5i (r5M p q u) * star (r5i (r5M p' q' u)) := by
      have h16 : r5i (16 : r5ZI) = 16 := map_ofNat r5i 16
      have h3200 : r5i (3200 : r5ZI) = 3200 := map_ofNat r5i 3200
      simp only [r5E, map_sub, map_mul, map_sum, r5i_star, h16]
      congr 2
      split_ifs <;> simp [h3200]
    
  5. The row and column sums of entry norms of E in C are at most 1220.

    Combining the kernel-verified boolean certificate r5chk_true with the norm bound r5i_norm shows that the sum of the complex norms along any row or column of E is bounded by 1220.

    lines 157–170

    uses r5chk_true, of_decide_eq_true, Bool.and_eq_true, r5i_norm, Finset.sum_le_sum

    lemma r5i_rowsum (p q : Fin 5) : (∑ p', ∑ q', ‖r5i (r5E p q p' q')‖) ≤ 1220 := by
      have h1 := of_decide_eq_true ((Bool.and_eq_true _ _).mp r5chk_true).1
      have h : (∑ p', ∑ q', ((r5bd (r5E p q p' q') : ℕ) : ℝ)) ≤ 1220 := by
        exact_mod_cast h1 p q
      refine le_trans ?_ h
      exact Finset.sum_le_sum fun _ _ => Finset.sum_le_sum fun _ _ => r5i_norm _
    
    lemma r5i_colsum (p' q' : Fin 5) : (∑ p, ∑ q, ‖r5i (r5E p q p' q')‖) ≤ 1220 := by
      have h1 := of_decide_eq_true ((Bool.and_eq_true _ _).mp r5chk_true).2
      have h : (∑ p, ∑ q, ((r5bd (r5E p q p' q') : ℕ) : ℝ)) ≤ 1220 := by
        exact_mod_cast h1 p' q'
      refine le_trans ?_ h
      exact Finset.sum_le_sum fun _ _ => Finset.sum_le_sum fun _ _ => r5i_norm _
    
  6. For any vector v and matrix e with row and column l1-norms at most 1220, the real part of v* e v is bounded below by -1220 ||v||^2.

    For each term, |Re(v_pq* e v_p'q')| <= ||e|| (|v_pq|^2 + |v_p'q'|^2)/2 by Cauchy-Schwarz. Summing over all indices and using the row and column sum bounds on e establishes that Re(v* e v) >= -1220 sum_pq |v_pq|^2.

    lines 171–230

    uses Complex.abs_re_le_norm, norm_mul, norm_star, norm_nonneg, sq_nonneg, Complex.re_sum, Finset.sum_le_sum, Finset.mul_sum, Finset.sum_mul, mul_le_mul_of_nonneg_right

    lemma r5quad_lower (v : Fin 5 → Fin 5 → ℂ) (e : Fin 5 → Fin 5 → Fin 5 → Fin 5 → ℂ)
        (hrow : ∀ p q, (∑ p', ∑ q', ‖e p q p' q'‖) ≤ 1220)
        (hcol : ∀ p' q', (∑ p, ∑ q, ‖e p q p' q'‖) ≤ 1220) :
        -(1220 * ∑ p, ∑ q, ‖v p q‖ ^ 2)
          ≤ (∑ p, ∑ q, ∑ p', ∑ q', star (v p q) * e p q p' q' * v p' q').re := by
      have hterm : ∀ p q p' q', -(‖e p q p' q'‖ * (‖v p q‖ ^ 2 + ‖v p' q'‖ ^ 2) / 2)
          ≤ (star (v p q) * e p q p' q' * v p' q').re := by
        intro p q p' q'
        have h1 : -‖star (v p q) * e p q p' q' * v p' q'‖ ≤ (star (v p q) * e p q p' q' * v p' q').re := by
          have := Complex.abs_re_le_norm (star (v p q) * e p q p' q' * v p' q')
          linarith [(abs_le.mp this).1]
        have h2 : ‖star (v p q) * e p q p' q' * v p' q'‖ = ‖v p q‖ * ‖e p q p' q'‖ * ‖v p' q'‖ := by
          rw [norm_mul, norm_mul, norm_star]
        rw [h2] at h1
        have h3 : 0 ≤ ‖e p q p' q'‖ := norm_nonneg _
        nlinarith [mul_nonneg h3 (sq_nonneg (‖v p q‖ - ‖v p' q'‖))]
      have hsum : (∑ p, ∑ q, ∑ p', ∑ q', -(‖e p q p' q'‖ * (‖v p q‖ ^ 2 + ‖v p' q'‖ ^ 2) / 2))
          ≤ (∑ p, ∑ q, ∑ p', ∑ q', star (v p q) * e p q p' q' * v p' q').re := by
        simp only [Complex.re_sum]
        exact Finset.sum_le_sum fun p _ => Finset.sum_le_sum fun q _ => Finset.sum_le_sum fun p' _ =>
          Finset.sum_le_sum fun q' _ => hterm p q p' q'
      set S : ℝ := ∑ p, ∑ q, ‖v p q‖ ^ 2 with hS
      have hfirst : (∑ p, ∑ q, ∑ p', ∑ q', ‖e p q p' q'‖ * ‖v p q‖ ^ 2) ≤ 1220 * S := by
        rw [hS, Finset.mul_sum]
        refine Finset.sum_le_sum fun p _ => ?_
        rw [Finset.mul_sum]
        refine Finset.sum_le_sum fun q _ => ?_
        calc (∑ p', ∑ q', ‖e p q p' q'‖ * ‖v p q‖ ^ 2)
            = (∑ p', ∑ q', ‖e p q p' q'‖) * ‖v p q‖ ^ 2 := by
              simp only [Finset.sum_mul]
          _ ≤ 1220 * ‖v p q‖ ^ 2 := mul_le_mul_of_nonneg_right (hrow p q) (sq_nonneg _)
      have hsecond : (∑ p, ∑ q, ∑ p', ∑ q', ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2) ≤ 1220 * S := by
        have hswap : (∑ p, ∑ q, ∑ p', ∑ q', ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2)
            = ∑ p', ∑ q', ∑ p, ∑ q, ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2 := by
          calc (∑ p, ∑ q, ∑ p', ∑ q', ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2)
              = ∑ p, ∑ p', ∑ q, ∑ q', ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2 :=
                Finset.sum_congr rfl fun _ _ => Finset.sum_comm
            _ = ∑ p', ∑ p, ∑ q, ∑ q', ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2 := Finset.sum_comm
            _ = ∑ p', ∑ p, ∑ q', ∑ q, ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2 :=
                Finset.sum_congr rfl fun _ _ => Finset.sum_congr rfl fun _ _ => Finset.sum_comm
            _ = ∑ p', ∑ q', ∑ p, ∑ q, ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2 :=
                Finset.sum_congr rfl fun _ _ => Finset.sum_comm
        rw [hswap, hS, Finset.mul_sum]
        refine Finset.sum_le_sum fun p' _ => ?_
        rw [Finset.mul_sum]
        refine Finset.sum_le_sum fun q' _ => ?_
        calc (∑ p, ∑ q, ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2)
            = (∑ p, ∑ q, ‖e p q p' q'‖) * ‖v p' q'‖ ^ 2 := by
              simp only [Finset.sum_mul]
          _ ≤ 1220 * ‖v p' q'‖ ^ 2 := mul_le_mul_of_nonneg_right (hcol p' q') (sq_nonneg _)
      have h : ∀ p q p' q', -(‖e p q p' q'‖ * (‖v p q‖ ^ 2 + ‖v p' q'‖ ^ 2) / 2)
          = (-1 / 2) * (‖e p q p' q'‖ * ‖v p q‖ ^ 2) + (-1 / 2) * (‖e p q p' q'‖ * ‖v p' q'‖ ^ 2) := by
        intro p q p' q'; ring
      have heq : (∑ p, ∑ q, ∑ p', ∑ q', -(‖e p q p' q'‖ * (‖v p q‖ ^ 2 + ‖v p' q'‖ ^ 2) / 2))
          = (-1 / 2) * (∑ p, ∑ q, ∑ p', ∑ q', ‖e p q p' q'‖ * ‖v p q‖ ^ 2)
            + (-1 / 2) * ∑ p, ∑ q, ∑ p', ∑ q', ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2 := by
        simp only [h, Finset.sum_add_distrib, ← Finset.mul_sum]
      rw [heq] at hsum
      linarith
    
  7. A product of two double sums equals a 4-fold sum of products.

    Distributing multiplication across the sums establishes the standard factorization identity for finite double sums.

    lines 231–240

    uses Finset.sum_mul, Finset.mul_sum, Finset.sum_congr

    lemma r5mul (f g : Fin 5 → Fin 5 → ℂ) :
        (∑ r, ∑ s, f r s) * (∑ r', ∑ s', g r' s') = ∑ r, ∑ s, ∑ r', ∑ s', f r s * g r' s' := by
      rw [Finset.sum_mul]
      refine Finset.sum_congr rfl fun r _ => ?_
      rw [Finset.sum_mul]
      refine Finset.sum_congr rfl fun s _ => ?_
      rw [Finset.mul_sum]
      refine Finset.sum_congr rfl fun r' _ => ?_
      rw [Finset.mul_sum]
    
  8. For a rank-one vector v_pq = (x_p)* y_q, the squared pencil evaluations expand as quadratic forms in v.

    With v_pq = (x_p)* y_q, the squared norms |x^T A_a y|^2, |x* B_j y|^2, and |x* y* C_k|^2 match the 4-fold contractions of v* v with the tensors formed from A_a, B_j, and C_k.

    lines 241–284

    uses star_mul', star_star, star_sum, r5mul, r5perm1, r5perm2, r5perm3

    lemma r5main (x y : Fin 5 → ℂ)
        (hA : ∀ a : Fin 3, ∑ r, ∑ s, x r * r5i (r5A a r s) * y s = 0)
        (hB : ∀ j : Fin 5, ∑ r, ∑ s, star (x r) * r5i (r5B j r s) * y s = 0) :
        x = 0 ∨ y = 0 := by
      set v : Fin 5 → Fin 5 → ℂ := fun p q => star (x p) * y q with hv
      have hstarv : ∀ p q, star (v p q) = x p * star (y q) := by
        intro p q; simp only [hv, star_mul', star_star]
      -- the three squares
      let aterm : Fin 3 → ℂ := fun a => ∑ r, ∑ s, x r * r5i (r5A a r s) * y s
      let bterm : Fin 5 → ℂ := fun j => ∑ r, ∑ s, star (x r) * r5i (r5B j r s) * y s
      let cterm : Fin 14 → ℂ := fun k => ∑ r, ∑ s, star (x r) * star (y s) * r5i (r5C k r s)
      have ha : ∀ a, (∑ p, ∑ q, ∑ p', ∑ q', star (v p q)
          * (r5i (r5A a p q') * star (r5i (r5A a p' q))) * v p' q') = aterm a * star (aterm a) := by
        intro a
        have e : (∑ r, ∑ s, ∑ r', ∑ s', (x r * r5i (r5A a r s) * y s) *
            (star (x r') * star (r5i (r5A a r' s')) * star (y s'))) = aterm a * star (aterm a) := by
          simp only [aterm, star_sum, star_mul', r5mul]
        rw [← e, r5perm1]
        refine Finset.sum_congr rfl fun p _ => Finset.sum_congr rfl fun q _ => Finset.sum_congr rfl fun p' _ =>
          Finset.sum_congr rfl fun q' _ => ?_
        rw [hstarv]; simp only [hv]; ring
      have hb : ∀ j, (∑ p, ∑ q, ∑ p', ∑ q', star (v p q)
          * (star (r5i (r5B j p q)) * r5i (r5B j p' q')) * v p' q') = bterm j * star (bterm j) := by
        intro j
        have e : (∑ r, ∑ s, ∑ r', ∑ s', (star (x r) * r5i (r5B j r s) * y s) *
            (x r' * star (r5i (r5B j r' s')) * star (y s'))) = bterm j * star (bterm j) := by
          simp only [bterm, star_sum, star_mul', star_star, r5mul]
        rw [← e, r5perm2]
        refine Finset.sum_congr rfl fun p _ => Finset.sum_congr rfl fun q _ => Finset.sum_congr rfl fun p' _ =>
          Finset.sum_congr rfl fun q' _ => ?_
        rw [hstarv]; simp only [hv]; ring
      have hc : ∀ k, (∑ p, ∑ q, ∑ p', ∑ q', star (v p q)
          * (r5i (r5C k p' q) * star (r5i (r5C k p q'))) * v p' q') = cterm k * star (cterm k) := by
        intro k
        have e : (∑ r, ∑ s, ∑ r', ∑ s', (star (x r) * star (y s) * r5i (r5C k r s)) *
            (x r' * y s' * star (r5i (r5C k r' s')))) = cterm k * star (cterm k) := by
          simp only [cterm, star_sum, star_mul', star_star, r5mul]
        rw [← e, r5perm3]
        refine Finset.sum_congr rfl fun p _ => Finset.sum_congr rfl fun q _ => Finset.sum_congr rfl fun p' _ =>
          Finset.sum_congr rfl fun q' _ => ?_
        rw [hstarv]; simp only [hv]; ring
      have ha0 : ∀ a, aterm a = 0 := fun a => hA a
      have hb0 : ∀ j, bterm j = 0 := fun j => hB j
      -- the quadratic form Q
  9. Under the vanishing assumptions on pencils A and B, the quadratic form Q = v* H v satisfies Re(Q) <= 0.

    Because x^T A_a y = 0 and x* B_j y = 0 for all a and j, the positive terms in Q vanish, leaving Q = - sum_k |cterm(k)|^2. The real part of this sum of squares is non-positive.

    lines 285–320

    uses r5i_H, r5swap4, Complex.star_def, Complex.mul_conj, Complex.normSq_nonneg, Finset.sum_nonneg

      set Q : ℂ := ∑ p, ∑ q, ∑ p', ∑ q', star (v p q) * r5i (r5H p q p' q') * v p' q' with hQ
      have hQval : Q = -∑ k, cterm k * star (cterm k) := by
        have hexp : ∀ p q p' q', star (v p q) * r5i (r5H p q p' q') * v p' q' =
            (∑ a, star (v p q) * (r5i (r5A a p q') * star (r5i (r5A a p' q))) * v p' q')
            + (∑ j, star (v p q) * (star (r5i (r5B j p q)) * r5i (r5B j p' q')) * v p' q')
            - ∑ k, star (v p q) * (r5i (r5C k p' q) * star (r5i (r5C k p q'))) * v p' q' := by
          intro p q p' q'
          rw [r5i_H]
          rw [mul_sub, sub_mul, mul_add, add_mul]
          simp only [Finset.mul_sum, Finset.sum_mul]
        have hsum : Q = (∑ p, ∑ q, ∑ p', ∑ q', ∑ a, star (v p q) * (r5i (r5A a p q') * star (r5i (r5A a p' q))) * v p' q')
            + (∑ p, ∑ q, ∑ p', ∑ q', ∑ j, star (v p q) * (star (r5i (r5B j p q)) * r5i (r5B j p' q')) * v p' q')
            - ∑ p, ∑ q, ∑ p', ∑ q', ∑ k, star (v p q) * (r5i (r5C k p' q) * star (r5i (r5C k p q'))) * v p' q' := by
          rw [hQ]
          simp only [hexp, Finset.sum_add_distrib, Finset.sum_sub_distrib]
        have e1 : (∑ p, ∑ q, ∑ p', ∑ q', ∑ a, star (v p q) * (r5i (r5A a p q') * star (r5i (r5A a p' q))) * v p' q')
            = ∑ a, ∑ p, ∑ q, ∑ p', ∑ q', star (v p q) * (r5i (r5A a p q') * star (r5i (r5A a p' q))) * v p' q' :=
          r5swap4 (fun a p q p' q' => star (v p q) * (r5i (r5A a p q') * star (r5i (r5A a p' q))) * v p' q')
        have e2 : (∑ p, ∑ q, ∑ p', ∑ q', ∑ j, star (v p q) * (star (r5i (r5B j p q)) * r5i (r5B j p' q')) * v p' q')
            = ∑ j, ∑ p, ∑ q, ∑ p', ∑ q', star (v p q) * (star (r5i (r5B j p q)) * r5i (r5B j p' q')) * v p' q' :=
          r5swap4 (fun j p q p' q' => star (v p q) * (star (r5i (r5B j p q)) * r5i (r5B j p' q')) * v p' q')
        have e3 : (∑ p, ∑ q, ∑ p', ∑ q', ∑ k, star (v p q) * (r5i (r5C k p' q) * star (r5i (r5C k p q'))) * v p' q')
            = ∑ k, ∑ p, ∑ q, ∑ p', ∑ q', star (v p q) * (r5i (r5C k p' q) * star (r5i (r5C k p q'))) * v p' q' :=
          r5swap4 (fun k p q p' q' => star (v p q) * (r5i (r5C k p' q) * star (r5i (r5C k p q'))) * v p' q')
        rw [hsum, e1, e2, e3]
        simp only [ha, hb, hc, ha0, hb0, zero_mul, Finset.sum_const_zero, zero_add]
        ring
      have hQre : Q.re ≤ 0 := by
        rw [hQval, Complex.neg_re, Complex.re_sum]
        have : ∀ k, 0 ≤ (cterm k * star (cterm k)).re := by
          intro k
          rw [Complex.star_def, Complex.mul_conj]
          simp [Complex.normSq_nonneg]
        have := Finset.sum_nonneg (fun k (_ : k ∈ Finset.univ) => this k)
        linarith
      -- the certificate
  10. The scaled quadratic form 16 Q decomposes into 3200 ||v||^2, a positive semidefinite term, and the error quadratic form.

    From the definition 16 H = 3200 I + M M* + E, contracting both sides with v yields 16 Q = 3200 sum_pq |v_pq|^2 + sum_u |m_u|^2 + v* E v, where m_u = sum_pq (v_pq)* M_pq,u.

    lines 321–353

    uses r5swap4, Finset.sum_eq_single, r5i_E

      let e : Fin 5 → Fin 5 → Fin 5 → Fin 5 → ℂ := fun p q p' q' => r5i (r5E p q p' q')
      let m : Fin 25 → ℂ := fun u => ∑ p, ∑ q, star (v p q) * r5i (r5M p q u)
      have hmsum : (∑ p, ∑ q, ∑ p', ∑ q', ∑ u, star (v p q) *
          (r5i (r5M p q u) * star (r5i (r5M p' q' u))) * v p' q') = ∑ u, m u * star (m u) := by
        rw [r5swap4]
        refine Finset.sum_congr rfl fun u _ => ?_
        simp only [m, star_sum, star_mul', star_star, r5mul]
        refine Finset.sum_congr rfl fun p _ => Finset.sum_congr rfl fun q _ => Finset.sum_congr rfl fun p' _ =>
          Finset.sum_congr rfl fun q' _ => ?_
        ring
      have hdiag : (∑ p, ∑ q, ∑ p', ∑ q', star (v p q) * (if p = p' ∧ q = q' then (3200 : ℂ) else 0) * v p' q')
          = 3200 * ∑ p, ∑ q, star (v p q) * v p q := by
        rw [Finset.mul_sum]
        refine Finset.sum_congr rfl fun p _ => ?_
        rw [Finset.mul_sum]
        refine Finset.sum_congr rfl fun q _ => ?_
        rw [Finset.sum_eq_single p, Finset.sum_eq_single q]
        · simp; ring
        · intro b _ hb; simp [Ne.symm hb]
        · simp
        · intro b _ hb
          refine Finset.sum_eq_zero fun c _ => ?_
          simp [Ne.symm hb]
        · simp
      have h16 : 16 * Q = 3200 * (∑ p, ∑ q, star (v p q) * v p q) + (∑ u, m u * star (m u))
          + ∑ p, ∑ q, ∑ p', ∑ q', star (v p q) * e p q p' q' * v p' q' := by
        rw [← hdiag, ← hmsum, hQ]
        simp only [Finset.mul_sum, ← Finset.sum_add_distrib]
        refine Finset.sum_congr rfl fun p _ => Finset.sum_congr rfl fun q _ => Finset.sum_congr rfl fun p' _ =>
          Finset.sum_congr rfl fun q' _ => ?_
        simp only [e, r5i_E, ← Finset.mul_sum, ← Finset.sum_mul]
        ring
      have hlow := r5quad_lower v e (fun p q => r5i_rowsum p q) (fun p' q' => r5i_colsum p' q')
  11. The norm squared of v must be zero, forcing v = 0.

    Because sum_u |m_u|^2 >= 0 and Re(v* E v) >= -1220 ||v||^2, taking real parts gives 16 Re(Q) >= (3200 - 1220) ||v||^2 = 1980 ||v||^2. Since Re(Q) <= 0 and ||v||^2 >= 0, this forces ||v||^2 = 0, so v_pq = 0 for all p and q.

    lines 354–378

    uses r5quad_lower, r5i_rowsum, r5i_colsum, Complex.mul_conj, Complex.normSq_eq_norm_sq, Complex.normSq_nonneg, Finset.sum_eq_zero_iff_of_nonneg

      set S : ℝ := ∑ p, ∑ q, ‖v p q‖ ^ 2 with hS
      have hSre : (∑ p, ∑ q, star (v p q) * v p q).re = S := by
        simp only [Complex.re_sum, hS]
        refine Finset.sum_congr rfl fun p _ => Finset.sum_congr rfl fun q _ => ?_
        rw [Complex.star_def, mul_comm, Complex.mul_conj, Complex.normSq_eq_norm_sq]
        exact Complex.ofReal_re _
      have hm : 0 ≤ (∑ u, m u * star (m u)).re := by
        rw [Complex.re_sum]
        refine Finset.sum_nonneg fun u _ => ?_
        rw [Complex.star_def, Complex.mul_conj]
        simp [Complex.normSq_nonneg]
      have hre := congrArg Complex.re h16
      have h16re : (16 * Q).re = 16 * Q.re := by simp [Complex.mul_re]
      have h3200re : (3200 * ∑ p, ∑ q, star (v p q) * v p q).re = 3200 * S := by
        rw [← hSre]; simp [Complex.mul_re]
      rw [h16re, Complex.add_re, Complex.add_re, h3200re] at hre
      have hS0 : S ≤ 0 := by nlinarith
      have hSnn : 0 ≤ S := Finset.sum_nonneg fun p _ => Finset.sum_nonneg fun q _ => sq_nonneg _
      have hS00 : S = 0 := le_antisymm hS0 hSnn
      have hv0 : ∀ p q, v p q = 0 := by
        intro p q
        have h1 := (Finset.sum_eq_zero_iff_of_nonneg
          (fun p _ => Finset.sum_nonneg fun q _ => sq_nonneg _)).mp hS00 p (Finset.mem_univ _)
        have h2 := (Finset.sum_eq_zero_iff_of_nonneg (fun q _ => sq_nonneg _)).mp h1 q (Finset.mem_univ _)
        simpa using h2
  12. Any common zero of the two pencils must be trivial: x = 0 or y = 0.

    Since v_pq = (x_p)* y_q = 0 for all p, q, if x is nonzero then there is some index p with x_p != 0; this forces y_q = 0 for all q, so y = 0.

    lines 379–392

    uses mul_eq_zero, star_eq_zero

      by_cases hx : x = 0
      · exact Or.inl hx
      · right
        obtain ⟨p, hp⟩ : ∃ p, x p ≠ 0 := by
          by_contra hne
          push Not at hne
          exact hx (funext hne)
        ext q
        have := hv0 p q
        simp only [hv] at this
        rcases mul_eq_zero.mp this with h | h
        · exact absurd (star_eq_zero.mp h) hp
        · simpa using h
    
  13. A family of matrices with an explicit dual inverse certificate at pivot positions is linearly independent.

    Given a vanishing linear combination sum_a g_a F_a = 0, evaluating at the pivot positions and multiplying by the certificate matrix W isolates each coefficient g_b times a scalar d; since d != 0, every g_b must vanish.

    lines 393–420

    uses Fintype.linearIndependent_iff, Matrix.sum_apply, Finset.sum_comm, mul_eq_zero

    lemma r5li {k : ℕ} (F : Fin k → Fin 5 → Fin 5 → r5ZI) (piv : Fin k → Fin 5 × Fin 5)
        (W : Fin k → Fin k → r5ZI) (d : r5ZI) (hd : r5i d ≠ 0)
        (hW : ∀ a b, ∑ i, F a (piv i).1 (piv i).2 * W i b = if a = b then d else 0) :
        LinearIndependent ℂ (fun a => (Matrix.of fun r s => r5i (F a r s) : Matrix (Fin 5) (Fin 5) ℂ)) := by
      rw [Fintype.linearIndependent_iff]
      intro g hg b
      have heval : ∀ i, ∑ a, g a * r5i (F a (piv i).1 (piv i).2) = 0 := by
        intro i
        have := congrFun (congrFun hg (piv i).1) (piv i).2
        simpa [Matrix.sum_apply] using this
      have h1 : ∑ i, (∑ a, g a * r5i (F a (piv i).1 (piv i).2)) * r5i (W i b) = 0 := by
        simp [heval]
      have h2 : (∑ i, (∑ a, g a * r5i (F a (piv i).1 (piv i).2)) * r5i (W i b))
          = ∑ a, g a * ∑ i, r5i (F a (piv i).1 (piv i).2) * r5i (W i b) := by
        simp only [Finset.sum_mul, Finset.mul_sum]
        rw [Finset.sum_comm]
        refine Finset.sum_congr rfl fun a _ => Finset.sum_congr rfl fun i _ => ?_
        ring
      have h3 : ∀ a, ∑ i, r5i (F a (piv i).1 (piv i).2) * r5i (W i b) = if a = b then r5i d else 0 := by
        intro a
        have := congrArg r5i (hW a b)
        simp only [map_sum, map_mul] at this
        rw [this]
        split_ifs <;> simp
      rw [h2] at h1
      simp only [h3, mul_ite, mul_zero, Finset.sum_ite_eq', Finset.mem_univ, if_true] at h1
      exact (mul_eq_zero.mp h1).resolve_right hd
    
  14. The dual inverse identities for pencils A and B are verified. automation

    Kernel computation evaluates the exact integer matrix products contracting A with WA at pivots r5pA to obtain r5dDA * I_3, and contracting B with WB at pivots r5pB to obtain r5dDB * I_5.

    lines 421–432

    def r5pA : Fin 3 → Fin 5 × Fin 5 := ![(1,0),(1,3),(2,1)]
    def r5pB : Fin 5 → Fin 5 × Fin 5 := ![(0,3),(2,1),(2,3),(3,3),(4,4)]
    def r5WAf (i j : Fin 3) : r5ZI := r5cell r5WA i j
    def r5WBf (i j : Fin 5) : r5ZI := r5cell r5WB i j
    
    set_option maxRecDepth 1000000 in
    theorem r5hWA : ∀ a b : Fin 3, ∑ i, r5A a (r5pA i).1 (r5pA i).2 * r5WAf i b = if a = b then r5dDA else 0 := by
      decide +kernel
    set_option maxRecDepth 1000000 in
    theorem r5hWB : ∀ a b : Fin 5, ∑ i, r5B a (r5pB i).1 (r5pB i).2 * r5WBf i b = if a = b then r5dDB else 0 := by
      decide +kernel
    
  15. Pencils A and B are linearly independent over C.

    Applying r5li with the verified dual certificates r5hWA and r5hWB, and noting that r5dDA and r5dDB have nonzero real parts (1680 and -9293469), proves linear independence of pencil A (size 3) and pencil B (size 5).

    lines 433–452

    uses r5li, r5hWA, r5hWB, r5i_re

    theorem r5hcond (x : Fin 5 → ℂ) (y : Fin 5 → ℂ)
        (hA : ∀ i : Fin 3, ∑ r : Fin 5, ∑ s : Fin 5, x r * (Matrix.of fun r s => r5i (r5A i r s) : Matrix (Fin 5) (Fin 5) ℂ) r s * y s = 0)
        (hB : ∀ j : Fin 5, ∑ r : Fin 5, ∑ s : Fin 5, star (x r) * (Matrix.of fun r s => r5i (r5B j r s) : Matrix (Fin 5) (Fin 5) ℂ) r s * y s = 0) :
        x = 0 ∨ y = 0 := r5main x y hA hB
    
    theorem r5li_A : LinearIndependent ℂ (fun a : Fin 3 => (Matrix.of fun r s => r5i (r5A a r s) : Matrix (Fin 5) (Fin 5) ℂ)) := by
      refine r5li r5A r5pA r5WAf r5dDA ?_ r5hWA
      intro h
      have := congrArg Complex.re h
      rw [r5i_re] at this
      simp [r5dDA] at this
    
    theorem r5li_B : LinearIndependent ℂ (fun a : Fin 5 => (Matrix.of fun r s => r5i (r5B a r s) : Matrix (Fin 5) (Fin 5) ℂ)) := by
      refine r5li r5B r5pB r5WBf r5dDB ?_ r5hWB
      intro h
      have := congrArg Complex.re h
      rw [r5i_re] at this
      simp [r5dDB] at this
    
    /-- Existence of range-criterion subspaces for the minimal sporadic quadruple (m, n, k, ℓ) = (5, 5, 3, 5). -/
  16. There exist range criterion subspaces in C^5 (x) C^5 of codimensions 3 and 5.

    Applying has_range_criterion_of_matrix_pencil to the linearly independent pencils A (size 3) and B (size 5) and the nonexistence of nontrivial common zeros (r5hcond) completes the proof.

    lines 453–455

    uses has_range_criterion_of_matrix_pencil, r5li_A, r5li_B, r5hcond

    theorem range_criterion_subspaces_sporadic_five_five_three_five :
        has_range_criterion_subspaces 5 5 3 5 := by
      exact has_range_criterion_of_matrix_pencil 5 5 3 5 (fun a => Matrix.of fun r s => r5i (r5A a r s)) (fun j => Matrix.of fun r s => r5i (r5B j r s)) r5li_A r5li_B r5hcond
Complex.ext
two complex numbers are equal if and only if their real and imaginary parts are equal
r5i_re
the real part of a Gaussian integer embedded in C is its integer real part cast to R
r5i_im
the imaginary part of a Gaussian integer embedded in C is its integer imaginary part cast to R
Complex.star_def
the star operation on C is complex conjugation
Complex.conj_re
the real part of the complex conjugate of z equals the real part of z
Complex.conj_im
the imaginary part of the complex conjugate of z equals the negative of the imaginary part of z
Complex.norm_le_abs_re_add_abs_im
the complex absolute value is bounded by the sum of the absolute values of the real and imaginary parts
Finset.sum_comm
finite sums over two index sets can be interchanged
Finset.sum_congr
two sums over the same finite set are equal if their summands are pointwise equal
map_ofNat
a ring homomorphism maps positive integer numerals to the corresponding numerals in the target ring
map_sub
an additive group homomorphism preserves subtraction
map_add
an additive map preserves addition
map_sum
an additive homomorphism distributes over finite sums
map_mul
a multiplicative map preserves multiplication
r5i_star
the Gaussian integer embedding into C commutes with the star involution (complex conjugation)
r5chk_true
the boolean certificate r5chk evaluates to true by kernel computation
of_decide_eq_true
if a decidable proposition evaluates to true via decide, the proposition holds
Bool.and_eq_true
a boolean conjunction evaluates to true if and only if both conjuncts evaluate to true
r5i_norm
the complex norm of an embedded Gaussian integer is bounded by its integer l1-norm
Finset.sum_le_sum
if f(i) <= g(i) pointwise on a finite set, then the sum of f is at most the sum of g
Complex.abs_re_le_norm
the absolute value of the real part of a complex number is bounded by its norm
norm_mul
the norm of a product is the product of the norms
norm_star
the norm of the complex conjugate of x equals the norm of x
norm_nonneg
norms are always non-negative
sq_nonneg
the square of any real number is non-negative
Complex.re_sum
the real part of a finite sum of complex numbers is the sum of the real parts
Finset.mul_sum
multiplication by a constant distributes over a finite sum from the left
Finset.sum_mul
multiplication by a constant distributes over a finite sum from the right
mul_le_mul_of_nonneg_right
multiplying an inequality on the right by a non-negative element preserves the inequality
star_mul'
in a commutative ring, the star of a product is the product of the stars
star_star
the star involution applied twice is the identity
star_sum
the star operation distributes over finite sums
r5mul
the product of two double sums factors into a 4-fold sum
r5perm1
a specific permutation of summation indices for 4-fold sums over Fin 5
r5perm2
a second permutation of summation indices for 4-fold sums over Fin 5
r5perm3
a third permutation of summation indices for 4-fold sums over Fin 5
r5i_H
the complex embedding of the Hermitian tensor H expands as a sum of products of A, B, and C with their conjugates
r5swap4
an outer sum over a finite index set can be interchanged with a 4-fold sum over Fin 5
Complex.mul_conj
a complex number times its conjugate equals its squared modulus
Complex.normSq_nonneg
the squared modulus of a complex number is non-negative
Finset.sum_nonneg
a sum of non-negative real numbers is non-negative
Finset.sum_eq_single
a finite sum whose terms vanish away from a single index equals the value at that index
r5i_E
the complex embedding of the error tensor E expands into 16 H minus a diagonal term minus M M*
r5quad_lower
a matrix whose row and column l1-norms are at most 1220 has quadratic form real part bounded below by -1220 times the squared norm
r5i_rowsum
the row sums of entry norms of the complex error matrix E are bounded by 1220
r5i_colsum
the column sums of entry norms of the complex error matrix E are bounded by 1220
Complex.normSq_eq_norm_sq
the squared modulus of a complex number equals its norm squared
Finset.sum_eq_zero_iff_of_nonneg
a sum of non-negative real terms is zero if and only if each term is zero
mul_eq_zero
a product in a domain is zero if and only if at least one factor is zero
star_eq_zero
the conjugate of x is zero if and only if x is zero
Fintype.linearIndependent_iff
a finite collection of vectors is linearly independent if and only if any vanishing linear combination has all zero coefficients
Matrix.sum_apply
evaluating a sum of matrices at entry (i, j) gives the sum of the entries at (i, j)
r5li
if a family of matrices admits an explicit dual certificate matrix at chosen pivot positions with a nonzero diagonal scalar, the family is linearly independent
r5hWA
the dual certificate WA satisfies the diagonal inverse relation with scalar r5dDA on pencil A
r5hWB
the dual certificate WB satisfies the diagonal inverse relation with scalar r5dDB on pencil B
has_range_criterion_of_matrix_pencil
linearly independent matrix pencils of sizes k and l with no nontrivial common zeros yield range criterion subspaces of codimensions k and l
r5li_A
pencil A is linearly independent over C
r5li_B
pencil B is linearly independent over C
r5hcond
any common zero of pencils A and B has x = 0 or y = 0

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

abbrev r5ZI := GaussianIntdef r5dA : List (List (ℤ × ℤ)) := [  [(-7,23),(-8,5),(-4,-2),(-3,0),(12,14),(-2,-7),(17,6),(-2,11),(3,18),(5,-8),(0,-12),(-23,4),(-4,-9),(14,-1),(14,12),(10,-7),(-2,-5),(-13,3),(-5,1),(-15,9),(-8,-8),(-13,14),(16,4),(0,0),(-9,-14)],  [(-1,-8),(-9,-5),(11,10),(-25,-3),(2,-18),(31,-1),(-8,7),(15,13),(13,-1),(3,1),(0,-5),(-11,3),(15,5),(-6,5),(2,21),(-13,-5),(-19,14),(-18,1),(15,1),(-11,-7),(-2,-8),(7,-8),(-5,16),(-16,3),(-2,-4)],  [(4,5),(-14,2),(-21,-15),(1,-3),(-16,-5),(1,5),(9,-9),(1,-11),(3,-24),(-1,-12),(-15,16),(-13,-2),(12,5),(-10,-10),(5,15),(6,7),(-11,8),(10,-5),(-16,-7),(-14,-9),(9,-10),(3,-3),(8,2),(4,-20),(-4,7)]]def r5dB : List (List (ℤ × ℤ)) := [  [(-7,7),(6,12),(11,8),(10,3),(5,-1),(-8,-8),(-9,9),(-9,11),(9,-8),(-9,8),(-8,-1),(2,-11),(-3,-12),(-12,0),(7,11),(-7,7),(-6,-8),(-9,-7),(-12,-8),(6,9),(-6,-9),(-10,-8),(-7,10),(2,-11),(-8,9)],  [(-6,11),(11,4),(-12,-1),(-11,11),(-4,-6),(10,-7),(6,-2),(11,-7),(-7,10),(-7,-11),(0,13),(1,-14),(-2,-7),(1,-5),(12,-1),(13,1),(-5,6),(11,2),(12,-4),(10,1),(-6,-8),(-11,3),(-8,-9),(11,-1),(-1,13)],  [(3,11),(-10,9),(-1,11),(12,-4),(-1,-9),(7,8),(10,5),(1,-10),(11,7),(-5,14),(-7,6),(0,1),(9,-5),(15,-3),(-6,-6),(7,8),(5,5),(-10,0),(10,-1),(12,3),(17,3),(1,11),(-8,8),(-10,6),(-9,4)],  [(-6,-5),(10,-7),(-14,2),(6,6),(-5,-8),(-11,-3),(12,0),(3,12),(-1,-9),(-8,-7),(10,-4),(-2,12),(-11,6),(-14,-5),(-12,6),(8,10),(8,-11),(-12,6),(12,2),(-7,-7),(3,-6),(2,5),(-11,-5),(-2,5),(-9,9)],  [(-5,-12),(-4,8),(1,-10),(13,-3),(9,-8),(2,-5),(6,-11),(-3,9),(-9,-5),(7,10),(5,-13),(-7,-11),(6,5),(-9,-8),(5,-10),(11,12),(-5,14),(0,-12),(9,-3),(1,12),(-8,-10),(10,8),(-10,1),(7,1),(4,-12)]]def r5dC : List (List (ℤ × ℤ)) := [  [(0,-6),(-2,7),(-2,-9),(8,6),(4,-6),(1,-2),(3,-2),(-3,-11),(11,4),(-1,-2),(6,-2),(-8,3),(-5,1),(-3,8),(-2,-7),(-5,7),(-8,17),(0,7),(-9,-8),(-5,4),(-2,2),(3,-2),(-4,-3),(-3,6),(-5,-2)],  [(5,3),(6,2),(5,11),(2,-1),(1,11),(-8,-7),(2,7),(-6,-3),(-6,5),(0,0),(-8,1),(3,0),(-1,-7),(1,9),(-8,5),(-2,2),(1,4),(4,-4),(-1,6),(6,1),(1,4),(-4,-3),(-3,11),(-1,4),(1,-10)],  [(2,5),(1,11),(8,0),(1,0),(5,-4),(-1,-4),(-8,-8),(-2,-10),(0,-2),(-1,-4),(-2,4),(8,4),(3,-4),(4,-5),(-6,-2),(-4,-4),(0,4),(1,-2),(7,3),(5,4),(2,-1),(1,-2),(-4,-3),(-5,-2),(0,-4)],  [(0,3),(-5,1),(6,1),(6,-3),(-2,-1),(-7,-5),(4,1),(4,2),(0,2),(-4,-6),(-1,-4),(1,-3),(-4,4),(5,9),(-2,8),(2,1),(-2,2),(11,-6),(-7,6),(3,-5),(-6,-8),(1,-3),(5,4),(-3,-2),(2,1)],  [(0,2),(1,0),(5,0),(-2,6),(2,-3),(9,-8),(-6,2),(2,0),(3,-13),(2,-2),(1,6),(-2,-5),(-1,-5),(-3,-1),(-7,-3),(-2,4),(-3,-2),(0,-1),(0,-5),(5,0),(4,-9),(6,-5),(-3,0),(-5,0),(-1,9)],  [(-10,-7),(-4,2),(-7,6),(0,2),(-6,-5),(-7,-3),(7,1),(-4,-2),(-1,1),(5,-1),(-2,-3),(-4,0),(2,2),(4,5),(-7,-3),(13,3),(-4,4),(8,0),(-3,5),(0,0),(-4,1),(-6,4),(4,-2),(5,6),(3,5)],  [(-2,-7),(3,-4),(6,-5),(2,-9),(-1,-7),(-12,6),(1,3),(-2,-1),(-6,-3),(-2,5),(0,1),(-5,-5),(-4,4),(3,0),(8,6),(-1,3),(2,-1),(-3,-2),(0,-1),(4,-11),(-5,-3),(-11,-2),(9,2),(-4,1),(0,4)],  [(-1,-9),(-2,-4),(0,5),(5,-4),(5,-1),(-3,2),(9,-3),(7,-1),(4,5),(8,3),(-2,-2),(2,1),(1,3),(0,5),(-2,-3),(4,5),(4,-2),(3,4),(-2,6),(8,6),(4,-5),(-3,3),(6,5),(7,-1),(-5,-4)],  [(-1,7),(-1,5),(8,3),(-9,2),(-2,5),(7,5),(-1,1),(2,9),(12,0),(-4,1),(-7,4),(3,-7),(1,2),(6,0),(4,-5),(-8,2),(-4,3),(6,-10),(1,3),(-7,1),(6,0),(4,3),(-2,-4),(-8,-1),(-1,5)],  [(-5,3),(5,3),(7,9),(0,-2),(-2,8),(3,1),(-2,0),(-3,-8),(2,0),(7,-2),(-6,4),(-2,1),(2,-10),(-5,5),(-9,-4),(-6,-1),(-1,3),(-1,-1),(2,-1),(8,1),(3,2),(1,3),(-5,2),(0,-1),(-2,-6)],  [(5,0),(-4,-1),(-3,2),(3,-9),(-2,4),(-1,3),(3,1),(6,6),(3,6),(-4,6),(-6,-7),(3,4),(1,7),(3,5),(10,-1),(-5,1),(4,-2),(6,-1),(-8,10),(-5,2),(1,1),(-1,-4),(3,2),(0,-2),(-1,-3)],  [(-2,7),(2,-5),(-5,-7),(2,-6),(-1,-3),(4,0),(-6,-6),(-4,-2),(-12,-6),(0,-3),(6,-4),(2,12),(2,3),(-1,-5),(-1,5),(6,-3),(12,-5),(-5,8),(-1,0),(2,-2),(-2,1),(1,0),(-2,0),(3,-4),(4,-1)],  [(3,1),(4,3),(-7,-1),(-6,8),(-3,11),(4,-6),(3,7),(-5,2),(6,8),(1,-6),(4,2),(6,-1),(-1,-4),(0,5),(-12,-1),(5,7),(-3,4),(3,-3),(3,3),(-5,6),(-3,4),(4,5),(-8,1),(7,8),(-2,-3)],  [(-4,-10),(4,-5),(-7,6),(2,-7),(-6,1),(-8,5),(3,6),(-1,5),(7,-1),(-6,4),(-1,-8),(-1,-13),(-3,5),(-2,3),(7,9),(0,-1),(1,-3),(0,-4),(0,-2),(3,-4),(-6,3),(-5,-3),(3,4),(4,-6),(-4,6)]]def r5dM : List (List (ℤ × ℤ)) := [  [(81,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(2,-7),(84,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(9,0),(-3,4),(89,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(1,-4),(-18,2),(-2,-5),(89,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(-7,-3),(2,0),(-13,7),(-6,-3),(86,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(3,1),(-18,4),(8,5),(14,-4),(-4,-4),(89,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(5,2),(-2,-6),(7,-12),(-2,8),(-6,-1),(2,2),(81,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(-1,-2),(-9,8),(-9,0),(14,7),(4,-9),(14,5),(8,-8),(82,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(-13,-4),(-6,-3),(-18,-4),(0,3),(14,5),(5,9),(0,4),(-1,13),(87,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(3,0),(5,-3),(3,6),(-9,13),(-1,-6),(2,1),(3,6),(-4,8),(4,-10),(77,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(-7,6),(1,13),(-6,13),(-2,-11),(10,-20),(17,-10),(-5,-6),(5,-9),(2,-8),(3,10),(82,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(-5,8),(14,0),(-13,14),(-19,1),(13,-8),(-12,-5),(-12,4),(-2,2),(12,-7),(3,8),(1,2),(86,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(4,-11),(-14,-2),(2,-16),(7,3),(-7,7),(-1,2),(-4,5),(-1,-3),(-3,7),(0,-6),(-4,-2),(-5,0),(75,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(1,10),(7,1),(-9,14),(-7,-1),(8,-6),(-9,0),(-2,5),(2,-3),(12,-2),(-5,13),(-5,2),(2,-2),(1,-4),(79,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(-2,-5),(-1,13),(-5,-4),(16,-14),(5,-6),(8,-7),(0,-3),(-2,-3),(4,6),(-12,1),(-5,5),(-8,3),(-2,-8),(6,7),(82,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(-3,2),(13,-3),(-3,6),(-2,-8),(-2,-1),(-6,-3),(-4,-4),(-3,10),(3,2),(-1,4),(-3,-1),(0,1),(-1,-4),(1,1),(2,-7),(78,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(-1,-11),(13,-6),(2,-10),(-9,8),(7,10),(-12,3),(-2,0),(-10,13),(18,1),(-5,-3),(-7,1),(-3,2),(9,-9),(-5,1),(-5,-1),(0,-3),(75,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(-7,-3),(0,-10),(-17,4),(2,0),(12,1),(-4,-2),(-10,-3),(3,2),(10,-2),(-6,6),(10,11),(1,1),(-10,-5),(5,2),(-1,0),(-7,3),(-1,0),(78,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(2,9),(7,-11),(1,13),(-1,11),(0,-13),(-6,5),(-3,1),(2,4),(8,-8),(6,-5),(6,1),(6,-6),(-2,1),(4,-10),(-9,2),(3,0),(-7,-1),(3,-5),(75,0),(0,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(-4,-11),(-5,-20),(-8,-5),(13,10),(13,-4),(-1,20),(2,-4),(7,18),(11,-2),(2,4),(2,6),(-3,0),(0,2),(-3,-2),(3,12),(3,-5),(3,-5),(4,-1),(3,4),(76,0),(0,0),(0,0),(0,0),(0,0),(0,0)],  [(10,-5),(4,19),(4,-7),(1,-3),(-8,16),(-2,-22),(-2,4),(-11,0),(-2,3),(-5,-1),(-1,-7),(-5,-7),(7,-2),(-4,1),(0,0),(-1,2),(-8,2),(-3,-5),(-5,-2),(-5,-10),(73,0),(0,0),(0,0),(0,0),(0,0)],  [(-9,-2),(-10,-1),(-12,-4),(8,-3),(13,4),(5,7),(-6,-2),(9,5),(12,-7),(1,3),(8,3),(5,4),(-5,-3),(-3,-2),(2,3),(-3,-2),(0,-1),(8,-2),(-4,0),(3,2),(-1,-2),(78,0),(0,0),(0,0),(0,0)],  [(-4,-3),(1,-24),(-1,-1),(-4,23),(10,-5),(-7,18),(7,7),(-3,10),(7,-7),(5,5),(-4,2),(4,0),(-2,5),(1,-3),(-1,1),(6,3),(10,2),(2,14),(0,5),(1,8),(-3,3),(-10,1),(78,0),(0,0),(0,0)],  [(4,-2),(-2,-1),(-8,-11),(7,-2),(6,18),(8,-3),(7,1),(-5,1),(6,1),(-8,-12),(-3,4),(-4,5),(6,-3),(-3,0),(5,-6),(2,4),(-6,2),(4,-2),(5,-4),(6,-7),(0,3),(5,6),(-4,-3),(73,0),(0,0)],  [(0,-4),(-2,5),(0,-12),(3,6),(1,1),(4,0),(-2,-2),(-3,-8),(3,11),(-3,4),(-10,-5),(1,0),(8,1),(-3,2),(4,-6),(6,-3),(-3,4),(2,-3),(8,-8),(4,6),(-3,-1),(5,-2),(-2,5),(-8,2),(74,0)]]def r5WA : List (List (ℤ × ℤ)) := [  [(-210,-286),(30,804),(208,-264)],  [(379,-3),(55,206),(-752,76)],  [(51,-811),(87,6),(-144,-644)]]def r5WB : List (List (ℤ × ℤ)) := [  [(-154936,-27196),(150438,147826),(-179405,-154381),(-95592,24217),(-40325,-177110)],  [(-2648,-199174),(209087,-158991),(61822,27502),(-75544,194724),(102740,-195290)],  [(91874,12808),(-50504,82662),(-275116,-140302),(220702,39504),(162841,43967)],  [(221285,105097),(-160034,-152028),(-45046,-101434),(-188159,69054),(-209570,-38755)],  [(139619,177267),(-31003,143159),(82887,162515),(-23530,145829),(25363,-196314)]]def r5dDA : r5ZI := ⟨1680,27712⟩def r5dDB : r5ZI := ⟨-9293469,-4452693⟩def r5cell (l : List (List (ℤ × ℤ))) (i j : ℕ) : r5ZI :=  let v := (l.getD i []).getD j (0, 0)  ⟨v.1, v.2⟩def r5A (a : Fin 3) (r s : Fin 5) : r5ZI := r5cell r5dA a (5 * r + s)def r5B (j : Fin 5) (r s : Fin 5) : r5ZI := r5cell r5dB j (5 * r + s)def r5C (k : Fin 14) (r s : Fin 5) : r5ZI := r5cell r5dC k (5 * r + s)def r5M (p q : Fin 5) (u : Fin 25) : r5ZI := r5cell r5dM (5 * p + q) udef r5star (z : r5ZI) : r5ZI := ⟨z.re, -z.im⟩def r5H (p q p' q' : Fin 5) : r5ZI :=  (∑ a, r5A a p q' * r5star (r5A a p' q)) + (∑ j, r5star (r5B j p q) * r5B j p' q')    - ∑ k, r5C k p' q * r5star (r5C k p q')def r5E (p q p' q' : Fin 5) : r5ZI :=  (16 : r5ZI) * r5H p q p' q' - (if p = p' ∧ q = q' then (3200 : r5ZI) else 0)    - ∑ u, r5M p q u * r5star (r5M p' q' u)def r5bd (z : r5ZI) : ℕ := z.re.natAbs + z.im.natAbsdef r5chk : Bool :=  decide (∀ p q : Fin 5, ∑ p', ∑ q', r5bd (r5E p q p' q') ≤ 1220) &&  decide (∀ p' q' : Fin 5, ∑ p, ∑ q, r5bd (r5E p q p' q') ≤ 1220)set_option maxRecDepth 1000000 intheorem r5chk_true : r5chk = true := by decide +kernelnoncomputable def r5i : r5ZI →+* ℂ := GaussianInt.toComplexlemma r5i_re (z : r5ZI) : (r5i z).re = (z.re : ℝ) := (GaussianInt.intCast_re z).symmlemma r5i_im (z : r5ZI) : (r5i z).im = (z.im : ℝ) := (GaussianInt.intCast_im z).symmlemma r5i_star (z : r5ZI) : r5i (r5star z) = star (r5i z) := by  apply Complex.ext  · rw [r5i_re, Complex.star_def, Complex.conj_re, r5i_re]; rfl  · rw [r5i_im, Complex.star_def, Complex.conj_im, r5i_im]    show (((-z.im : ℤ)) : ℝ) = _    push_cast; ringlemma r5i_norm (z : r5ZI) : ‖r5i z‖ ≤ (r5bd z : ℝ) := by  have h := Complex.norm_le_abs_re_add_abs_im (r5i z)  rw [r5i_re, r5i_im] at h  have hb : (r5bd z : ℝ) = |(z.re : ℝ)| + |(z.im : ℝ)| := by    simp [r5bd, Nat.cast_add, Nat.cast_natAbs, Int.cast_abs]  rw [hb]; exact hlemma r5perm1 (f : Fin 5 → Fin 5 → Fin 5 → Fin 5 → ℂ) :    (∑ r, ∑ s, ∑ r', ∑ s', f r s r' s') = ∑ p, ∑ q, ∑ p', ∑ q', f p q' p' q := by  refine Finset.sum_congr rfl fun r _ => ?_  calc (∑ s, ∑ r', ∑ s', f r s r' s') = ∑ r', ∑ s, ∑ s', f r s r' s' := Finset.sum_comm    _ = ∑ r', ∑ s', ∑ s, f r s r' s' := Finset.sum_congr rfl fun _ _ => Finset.sum_comm    _ = ∑ s', ∑ r', ∑ s, f r s r' s' := Finset.sum_commlemma r5perm2 (f : Fin 5 → Fin 5 → Fin 5 → Fin 5 → ℂ) :    (∑ r, ∑ s, ∑ r', ∑ s', f r s r' s') = ∑ p, ∑ q, ∑ p', ∑ q', f p' q' p q := by  calc (∑ r, ∑ s, ∑ r', ∑ s', f r s r' s') = ∑ r, ∑ r', ∑ s, ∑ s', f r s r' s' :=        Finset.sum_congr rfl fun _ _ => Finset.sum_comm    _ = ∑ r', ∑ r, ∑ s, ∑ s', f r s r' s' := Finset.sum_comm    _ = ∑ r', ∑ r, ∑ s', ∑ s, f r s r' s' :=        Finset.sum_congr rfl fun _ _ => Finset.sum_congr rfl fun _ _ => Finset.sum_comm    _ = ∑ r', ∑ s', ∑ r, ∑ s, f r s r' s' :=        Finset.sum_congr rfl fun _ _ => Finset.sum_commlemma r5perm3 (f : Fin 5 → Fin 5 → Fin 5 → Fin 5 → ℂ) :    (∑ r, ∑ s, ∑ r', ∑ s', f r s r' s') = ∑ p, ∑ q, ∑ p', ∑ q', f p' q p q' := by  calc (∑ r, ∑ s, ∑ r', ∑ s', f r s r' s') = ∑ r, ∑ r', ∑ s, ∑ s', f r s r' s' :=        Finset.sum_congr rfl fun _ _ => Finset.sum_comm    _ = ∑ r', ∑ r, ∑ s, ∑ s', f r s r' s' := Finset.sum_comm    _ = ∑ r', ∑ s, ∑ r, ∑ s', f r s r' s' :=        Finset.sum_congr rfl fun _ _ => Finset.sum_commlemma r5swap4 {n : ℕ} (h : Fin n → Fin 5 → Fin 5 → Fin 5 → Fin 5 → ℂ) :    (∑ p, ∑ q, ∑ p', ∑ q', ∑ i, h i p q p' q') = ∑ i, ∑ p, ∑ q, ∑ p', ∑ q', h i p q p' q' := by  calc (∑ p, ∑ q, ∑ p', ∑ q', ∑ i, h i p q p' q')      = ∑ p, ∑ q, ∑ p', ∑ i, ∑ q', h i p q p' q' :=        Finset.sum_congr rfl fun _ _ => Finset.sum_congr rfl fun _ _ =>          Finset.sum_congr rfl fun _ _ => Finset.sum_comm    _ = ∑ p, ∑ q, ∑ i, ∑ p', ∑ q', h i p q p' q' :=        Finset.sum_congr rfl fun _ _ => Finset.sum_congr rfl fun _ _ => Finset.sum_comm    _ = ∑ p, ∑ i, ∑ q, ∑ p', ∑ q', h i p q p' q' :=        Finset.sum_congr rfl fun _ _ => Finset.sum_comm    _ = ∑ i, ∑ p, ∑ q, ∑ p', ∑ q', h i p q p' q' := Finset.sum_commlemma r5i_H (p q p' q' : Fin 5) : r5i (r5H p q p' q') =    (∑ a, r5i (r5A a p q') * star (r5i (r5A a p' q)))      + (∑ j, star (r5i (r5B j p q)) * r5i (r5B j p' q'))      - ∑ k, r5i (r5C k p' q) * star (r5i (r5C k p q')) := by  simp only [r5H, map_sub, map_add, map_sum, map_mul, r5i_star]lemma r5i_E (p q p' q' : Fin 5) : r5i (r5E p q p' q') =    16 * r5i (r5H p q p' q') - (if p = p' ∧ q = q' then (3200 : ℂ) else 0)      - ∑ u, r5i (r5M p q u) * star (r5i (r5M p' q' u)) := by  have h16 : r5i (16 : r5ZI) = 16 := map_ofNat r5i 16  have h3200 : r5i (3200 : r5ZI) = 3200 := map_ofNat r5i 3200  simp only [r5E, map_sub, map_mul, map_sum, r5i_star, h16]  congr 2  split_ifs <;> simp [h3200]lemma r5i_rowsum (p q : Fin 5) : (∑ p', ∑ q', ‖r5i (r5E p q p' q')‖) ≤ 1220 := by  have h1 := of_decide_eq_true ((Bool.and_eq_true _ _).mp r5chk_true).1  have h : (∑ p', ∑ q', ((r5bd (r5E p q p' q') : ℕ) : ℝ)) ≤ 1220 := by    exact_mod_cast h1 p q  refine le_trans ?_ h  exact Finset.sum_le_sum fun _ _ => Finset.sum_le_sum fun _ _ => r5i_norm _lemma r5i_colsum (p' q' : Fin 5) : (∑ p, ∑ q, ‖r5i (r5E p q p' q')‖) ≤ 1220 := by  have h1 := of_decide_eq_true ((Bool.and_eq_true _ _).mp r5chk_true).2  have h : (∑ p, ∑ q, ((r5bd (r5E p q p' q') : ℕ) : ℝ)) ≤ 1220 := by    exact_mod_cast h1 p' q'  refine le_trans ?_ h  exact Finset.sum_le_sum fun _ _ => Finset.sum_le_sum fun _ _ => r5i_norm _lemma r5quad_lower (v : Fin 5 → Fin 5 → ℂ) (e : Fin 5 → Fin 5 → Fin 5 → Fin 5 → ℂ)    (hrow : ∀ p q, (∑ p', ∑ q', ‖e p q p' q'‖) ≤ 1220)    (hcol : ∀ p' q', (∑ p, ∑ q, ‖e p q p' q'‖) ≤ 1220) :    -(1220 * ∑ p, ∑ q, ‖v p q‖ ^ 2)      ≤ (∑ p, ∑ q, ∑ p', ∑ q', star (v p q) * e p q p' q' * v p' q').re := by  have hterm : ∀ p q p' q', -(‖e p q p' q'‖ * (‖v p q‖ ^ 2 + ‖v p' q'‖ ^ 2) / 2)      ≤ (star (v p q) * e p q p' q' * v p' q').re := by    intro p q p' q'    have h1 : -‖star (v p q) * e p q p' q' * v p' q'‖ ≤ (star (v p q) * e p q p' q' * v p' q').re := by      have := Complex.abs_re_le_norm (star (v p q) * e p q p' q' * v p' q')      linarith [(abs_le.mp this).1]    have h2 : ‖star (v p q) * e p q p' q' * v p' q'‖ = ‖v p q‖ * ‖e p q p' q'‖ * ‖v p' q'‖ := by      rw [norm_mul, norm_mul, norm_star]    rw [h2] at h1    have h3 : 0 ≤ ‖e p q p' q'‖ := norm_nonneg _    nlinarith [mul_nonneg h3 (sq_nonneg (‖v p q‖ - ‖v p' q'‖))]  have hsum : (∑ p, ∑ q, ∑ p', ∑ q', -(‖e p q p' q'‖ * (‖v p q‖ ^ 2 + ‖v p' q'‖ ^ 2) / 2))      ≤ (∑ p, ∑ q, ∑ p', ∑ q', star (v p q) * e p q p' q' * v p' q').re := by    simp only [Complex.re_sum]    exact Finset.sum_le_sum fun p _ => Finset.sum_le_sum fun q _ => Finset.sum_le_sum fun p' _ =>      Finset.sum_le_sum fun q' _ => hterm p q p' q'  set S : ℝ := ∑ p, ∑ q, ‖v p q‖ ^ 2 with hS  have hfirst : (∑ p, ∑ q, ∑ p', ∑ q', ‖e p q p' q'‖ * ‖v p q‖ ^ 2) ≤ 1220 * S := by    rw [hS, Finset.mul_sum]    refine Finset.sum_le_sum fun p _ => ?_    rw [Finset.mul_sum]    refine Finset.sum_le_sum fun q _ => ?_    calc (∑ p', ∑ q', ‖e p q p' q'‖ * ‖v p q‖ ^ 2)        = (∑ p', ∑ q', ‖e p q p' q'‖) * ‖v p q‖ ^ 2 := by          simp only [Finset.sum_mul]      _ ≤ 1220 * ‖v p q‖ ^ 2 := mul_le_mul_of_nonneg_right (hrow p q) (sq_nonneg _)  have hsecond : (∑ p, ∑ q, ∑ p', ∑ q', ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2) ≤ 1220 * S := by    have hswap : (∑ p, ∑ q, ∑ p', ∑ q', ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2)        = ∑ p', ∑ q', ∑ p, ∑ q, ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2 := by      calc (∑ p, ∑ q, ∑ p', ∑ q', ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2)          = ∑ p, ∑ p', ∑ q, ∑ q', ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2 :=            Finset.sum_congr rfl fun _ _ => Finset.sum_comm        _ = ∑ p', ∑ p, ∑ q, ∑ q', ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2 := Finset.sum_comm        _ = ∑ p', ∑ p, ∑ q', ∑ q, ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2 :=            Finset.sum_congr rfl fun _ _ => Finset.sum_congr rfl fun _ _ => Finset.sum_comm        _ = ∑ p', ∑ q', ∑ p, ∑ q, ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2 :=            Finset.sum_congr rfl fun _ _ => Finset.sum_comm    rw [hswap, hS, Finset.mul_sum]    refine Finset.sum_le_sum fun p' _ => ?_    rw [Finset.mul_sum]    refine Finset.sum_le_sum fun q' _ => ?_    calc (∑ p, ∑ q, ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2)        = (∑ p, ∑ q, ‖e p q p' q'‖) * ‖v p' q'‖ ^ 2 := by          simp only [Finset.sum_mul]      _ ≤ 1220 * ‖v p' q'‖ ^ 2 := mul_le_mul_of_nonneg_right (hcol p' q') (sq_nonneg _)  have h : ∀ p q p' q', -(‖e p q p' q'‖ * (‖v p q‖ ^ 2 + ‖v p' q'‖ ^ 2) / 2)      = (-1 / 2) * (‖e p q p' q'‖ * ‖v p q‖ ^ 2) + (-1 / 2) * (‖e p q p' q'‖ * ‖v p' q'‖ ^ 2) := by    intro p q p' q'; ring  have heq : (∑ p, ∑ q, ∑ p', ∑ q', -(‖e p q p' q'‖ * (‖v p q‖ ^ 2 + ‖v p' q'‖ ^ 2) / 2))      = (-1 / 2) * (∑ p, ∑ q, ∑ p', ∑ q', ‖e p q p' q'‖ * ‖v p q‖ ^ 2)        + (-1 / 2) * ∑ p, ∑ q, ∑ p', ∑ q', ‖e p q p' q'‖ * ‖v p' q'‖ ^ 2 := by    simp only [h, Finset.sum_add_distrib, ← Finset.mul_sum]  rw [heq] at hsum  linarithlemma r5mul (f g : Fin 5 → Fin 5 → ℂ) :    (∑ r, ∑ s, f r s) * (∑ r', ∑ s', g r' s') = ∑ r, ∑ s, ∑ r', ∑ s', f r s * g r' s' := by  rw [Finset.sum_mul]  refine Finset.sum_congr rfl fun r _ => ?_  rw [Finset.sum_mul]  refine Finset.sum_congr rfl fun s _ => ?_  rw [Finset.mul_sum]  refine Finset.sum_congr rfl fun r' _ => ?_  rw [Finset.mul_sum]lemma r5main (x y : Fin 5 → ℂ)    (hA : ∀ a : Fin 3, ∑ r, ∑ s, x r * r5i (r5A a r s) * y s = 0)    (hB : ∀ j : Fin 5, ∑ r, ∑ s, star (x r) * r5i (r5B j r s) * y s = 0) :    x = 0 ∨ y = 0 := by  set v : Fin 5 → Fin 5 → ℂ := fun p q => star (x p) * y q with hv  have hstarv : ∀ p q, star (v p q) = x p * star (y q) := by    intro p q; simp only [hv, star_mul', star_star]  -- the three squares  let aterm : Fin 3 → ℂ := fun a => ∑ r, ∑ s, x r * r5i (r5A a r s) * y s  let bterm : Fin 5 → ℂ := fun j => ∑ r, ∑ s, star (x r) * r5i (r5B j r s) * y s  let cterm : Fin 14 → ℂ := fun k => ∑ r, ∑ s, star (x r) * star (y s) * r5i (r5C k r s)  have ha : ∀ a, (∑ p, ∑ q, ∑ p', ∑ q', star (v p q)      * (r5i (r5A a p q') * star (r5i (r5A a p' q))) * v p' q') = aterm a * star (aterm a) := by    intro a    have e : (∑ r, ∑ s, ∑ r', ∑ s', (x r * r5i (r5A a r s) * y s) *        (star (x r') * star (r5i (r5A a r' s')) * star (y s'))) = aterm a * star (aterm a) := by      simp only [aterm, star_sum, star_mul', r5mul]    rw [← e, r5perm1]    refine Finset.sum_congr rfl fun p _ => Finset.sum_congr rfl fun q _ => Finset.sum_congr rfl fun p' _ =>      Finset.sum_congr rfl fun q' _ => ?_    rw [hstarv]; simp only [hv]; ring  have hb : ∀ j, (∑ p, ∑ q, ∑ p', ∑ q', star (v p q)      * (star (r5i (r5B j p q)) * r5i (r5B j p' q')) * v p' q') = bterm j * star (bterm j) := by    intro j    have e : (∑ r, ∑ s, ∑ r', ∑ s', (star (x r) * r5i (r5B j r s) * y s) *        (x r' * star (r5i (r5B j r' s')) * star (y s'))) = bterm j * star (bterm j) := by      simp only [bterm, star_sum, star_mul', star_star, r5mul]    rw [← e, r5perm2]    refine Finset.sum_congr rfl fun p _ => Finset.sum_congr rfl fun q _ => Finset.sum_congr rfl fun p' _ =>      Finset.sum_congr rfl fun q' _ => ?_    rw [hstarv]; simp only [hv]; ring  have hc : ∀ k, (∑ p, ∑ q, ∑ p', ∑ q', star (v p q)      * (r5i (r5C k p' q) * star (r5i (r5C k p q'))) * v p' q') = cterm k * star (cterm k) := by    intro k    have e : (∑ r, ∑ s, ∑ r', ∑ s', (star (x r) * star (y s) * r5i (r5C k r s)) *        (x r' * y s' * star (r5i (r5C k r' s')))) = cterm k * star (cterm k) := by      simp only [cterm, star_sum, star_mul', star_star, r5mul]    rw [← e, r5perm3]    refine Finset.sum_congr rfl fun p _ => Finset.sum_congr rfl fun q _ => Finset.sum_congr rfl fun p' _ =>      Finset.sum_congr rfl fun q' _ => ?_    rw [hstarv]; simp only [hv]; ring  have ha0 : ∀ a, aterm a = 0 := fun a => hA a  have hb0 : ∀ j, bterm j = 0 := fun j => hB j  -- the quadratic form Q  set Q : ℂ := ∑ p, ∑ q, ∑ p', ∑ q', star (v p q) * r5i (r5H p q p' q') * v p' q' with hQ  have hQval : Q = -∑ k, cterm k * star (cterm k) := by    have hexp : ∀ p q p' q', star (v p q) * r5i (r5H p q p' q') * v p' q' =        (∑ a, star (v p q) * (r5i (r5A a p q') * star (r5i (r5A a p' q))) * v p' q')        + (∑ j, star (v p q) * (star (r5i (r5B j p q)) * r5i (r5B j p' q')) * v p' q')        - ∑ k, star (v p q) * (r5i (r5C k p' q) * star (r5i (r5C k p q'))) * v p' q' := by      intro p q p' q'      rw [r5i_H]      rw [mul_sub, sub_mul, mul_add, add_mul]      simp only [Finset.mul_sum, Finset.sum_mul]    have hsum : Q = (∑ p, ∑ q, ∑ p', ∑ q', ∑ a, star (v p q) * (r5i (r5A a p q') * star (r5i (r5A a p' q))) * v p' q')        + (∑ p, ∑ q, ∑ p', ∑ q', ∑ j, star (v p q) * (star (r5i (r5B j p q)) * r5i (r5B j p' q')) * v p' q')        - ∑ p, ∑ q, ∑ p', ∑ q', ∑ k, star (v p q) * (r5i (r5C k p' q) * star (r5i (r5C k p q'))) * v p' q' := by      rw [hQ]      simp only [hexp, Finset.sum_add_distrib, Finset.sum_sub_distrib]    have e1 : (∑ p, ∑ q, ∑ p', ∑ q', ∑ a, star (v p q) * (r5i (r5A a p q') * star (r5i (r5A a p' q))) * v p' q')        = ∑ a, ∑ p, ∑ q, ∑ p', ∑ q', star (v p q) * (r5i (r5A a p q') * star (r5i (r5A a p' q))) * v p' q' :=      r5swap4 (fun a p q p' q' => star (v p q) * (r5i (r5A a p q') * star (r5i (r5A a p' q))) * v p' q')    have e2 : (∑ p, ∑ q, ∑ p', ∑ q', ∑ j, star (v p q) * (star (r5i (r5B j p q)) * r5i (r5B j p' q')) * v p' q')        = ∑ j, ∑ p, ∑ q, ∑ p', ∑ q', star (v p q) * (star (r5i (r5B j p q)) * r5i (r5B j p' q')) * v p' q' :=      r5swap4 (fun j p q p' q' => star (v p q) * (star (r5i (r5B j p q)) * r5i (r5B j p' q')) * v p' q')    have e3 : (∑ p, ∑ q, ∑ p', ∑ q', ∑ k, star (v p q) * (r5i (r5C k p' q) * star (r5i (r5C k p q'))) * v p' q')        = ∑ k, ∑ p, ∑ q, ∑ p', ∑ q', star (v p q) * (r5i (r5C k p' q) * star (r5i (r5C k p q'))) * v p' q' :=      r5swap4 (fun k p q p' q' => star (v p q) * (r5i (r5C k p' q) * star (r5i (r5C k p q'))) * v p' q')    rw [hsum, e1, e2, e3]    simp only [ha, hb, hc, ha0, hb0, zero_mul, Finset.sum_const_zero, zero_add]    ring  have hQre : Q.re ≤ 0 := by    rw [hQval, Complex.neg_re, Complex.re_sum]    have : ∀ k, 0 ≤ (cterm k * star (cterm k)).re := by      intro k      rw [Complex.star_def, Complex.mul_conj]      simp [Complex.normSq_nonneg]    have := Finset.sum_nonneg (fun k (_ : k ∈ Finset.univ) => this k)    linarith  -- the certificate  let e : Fin 5 → Fin 5 → Fin 5 → Fin 5 → ℂ := fun p q p' q' => r5i (r5E p q p' q')  let m : Fin 25 → ℂ := fun u => ∑ p, ∑ q, star (v p q) * r5i (r5M p q u)  have hmsum : (∑ p, ∑ q, ∑ p', ∑ q', ∑ u, star (v p q) *      (r5i (r5M p q u) * star (r5i (r5M p' q' u))) * v p' q') = ∑ u, m u * star (m u) := by    rw [r5swap4]    refine Finset.sum_congr rfl fun u _ => ?_    simp only [m, star_sum, star_mul', star_star, r5mul]    refine Finset.sum_congr rfl fun p _ => Finset.sum_congr rfl fun q _ => Finset.sum_congr rfl fun p' _ =>      Finset.sum_congr rfl fun q' _ => ?_    ring  have hdiag : (∑ p, ∑ q, ∑ p', ∑ q', star (v p q) * (if p = p' ∧ q = q' then (3200 : ℂ) else 0) * v p' q')      = 3200 * ∑ p, ∑ q, star (v p q) * v p q := by    rw [Finset.mul_sum]    refine Finset.sum_congr rfl fun p _ => ?_    rw [Finset.mul_sum]    refine Finset.sum_congr rfl fun q _ => ?_    rw [Finset.sum_eq_single p, Finset.sum_eq_single q]    · simp; ring    · intro b _ hb; simp [Ne.symm hb]    · simp    · intro b _ hb      refine Finset.sum_eq_zero fun c _ => ?_      simp [Ne.symm hb]    · simp  have h16 : 16 * Q = 3200 * (∑ p, ∑ q, star (v p q) * v p q) + (∑ u, m u * star (m u))      + ∑ p, ∑ q, ∑ p', ∑ q', star (v p q) * e p q p' q' * v p' q' := by    rw [← hdiag, ← hmsum, hQ]    simp only [Finset.mul_sum, ← Finset.sum_add_distrib]    refine Finset.sum_congr rfl fun p _ => Finset.sum_congr rfl fun q _ => Finset.sum_congr rfl fun p' _ =>      Finset.sum_congr rfl fun q' _ => ?_    simp only [e, r5i_E, ← Finset.mul_sum, ← Finset.sum_mul]    ring  have hlow := r5quad_lower v e (fun p q => r5i_rowsum p q) (fun p' q' => r5i_colsum p' q')  set S : ℝ := ∑ p, ∑ q, ‖v p q‖ ^ 2 with hS  have hSre : (∑ p, ∑ q, star (v p q) * v p q).re = S := by    simp only [Complex.re_sum, hS]    refine Finset.sum_congr rfl fun p _ => Finset.sum_congr rfl fun q _ => ?_    rw [Complex.star_def, mul_comm, Complex.mul_conj, Complex.normSq_eq_norm_sq]    exact Complex.ofReal_re _  have hm : 0 ≤ (∑ u, m u * star (m u)).re := by    rw [Complex.re_sum]    refine Finset.sum_nonneg fun u _ => ?_    rw [Complex.star_def, Complex.mul_conj]    simp [Complex.normSq_nonneg]  have hre := congrArg Complex.re h16  have h16re : (16 * Q).re = 16 * Q.re := by simp [Complex.mul_re]  have h3200re : (3200 * ∑ p, ∑ q, star (v p q) * v p q).re = 3200 * S := by    rw [← hSre]; simp [Complex.mul_re]  rw [h16re, Complex.add_re, Complex.add_re, h3200re] at hre  have hS0 : S ≤ 0 := by nlinarith  have hSnn : 0 ≤ S := Finset.sum_nonneg fun p _ => Finset.sum_nonneg fun q _ => sq_nonneg _  have hS00 : S = 0 := le_antisymm hS0 hSnn  have hv0 : ∀ p q, v p q = 0 := by    intro p q    have h1 := (Finset.sum_eq_zero_iff_of_nonneg      (fun p _ => Finset.sum_nonneg fun q _ => sq_nonneg _)).mp hS00 p (Finset.mem_univ _)    have h2 := (Finset.sum_eq_zero_iff_of_nonneg (fun q _ => sq_nonneg _)).mp h1 q (Finset.mem_univ _)    simpa using h2  by_cases hx : x = 0  · exact Or.inl hx  · right    obtain ⟨p, hp⟩ : ∃ p, x p ≠ 0 := by      by_contra hne      push Not at hne      exact hx (funext hne)    ext q    have := hv0 p q    simp only [hv] at this    rcases mul_eq_zero.mp this with h | h    · exact absurd (star_eq_zero.mp h) hp    · simpa using hlemma r5li {k : ℕ} (F : Fin k → Fin 5 → Fin 5 → r5ZI) (piv : Fin k → Fin 5 × Fin 5)    (W : Fin k → Fin k → r5ZI) (d : r5ZI) (hd : r5i d ≠ 0)    (hW : ∀ a b, ∑ i, F a (piv i).1 (piv i).2 * W i b = if a = b then d else 0) :    LinearIndependent ℂ (fun a => (Matrix.of fun r s => r5i (F a r s) : Matrix (Fin 5) (Fin 5) ℂ)) := by  rw [Fintype.linearIndependent_iff]  intro g hg b  have heval : ∀ i, ∑ a, g a * r5i (F a (piv i).1 (piv i).2) = 0 := by    intro i    have := congrFun (congrFun hg (piv i).1) (piv i).2    simpa [Matrix.sum_apply] using this  have h1 : ∑ i, (∑ a, g a * r5i (F a (piv i).1 (piv i).2)) * r5i (W i b) = 0 := by    simp [heval]  have h2 : (∑ i, (∑ a, g a * r5i (F a (piv i).1 (piv i).2)) * r5i (W i b))      = ∑ a, g a * ∑ i, r5i (F a (piv i).1 (piv i).2) * r5i (W i b) := by    simp only [Finset.sum_mul, Finset.mul_sum]    rw [Finset.sum_comm]    refine Finset.sum_congr rfl fun a _ => Finset.sum_congr rfl fun i _ => ?_    ring  have h3 : ∀ a, ∑ i, r5i (F a (piv i).1 (piv i).2) * r5i (W i b) = if a = b then r5i d else 0 := by    intro a    have := congrArg r5i (hW a b)    simp only [map_sum, map_mul] at this    rw [this]    split_ifs <;> simp  rw [h2] at h1  simp only [h3, mul_ite, mul_zero, Finset.sum_ite_eq', Finset.mem_univ, if_true] at h1  exact (mul_eq_zero.mp h1).resolve_right hddef r5pA : Fin 3 → Fin 5 × Fin 5 := ![(1,0),(1,3),(2,1)]def r5pB : Fin 5 → Fin 5 × Fin 5 := ![(0,3),(2,1),(2,3),(3,3),(4,4)]def r5WAf (i j : Fin 3) : r5ZI := r5cell r5WA i jdef r5WBf (i j : Fin 5) : r5ZI := r5cell r5WB i jset_option maxRecDepth 1000000 intheorem r5hWA : ∀ a b : Fin 3, ∑ i, r5A a (r5pA i).1 (r5pA i).2 * r5WAf i b = if a = b then r5dDA else 0 := by  decide +kernelset_option maxRecDepth 1000000 intheorem r5hWB : ∀ a b : Fin 5, ∑ i, r5B a (r5pB i).1 (r5pB i).2 * r5WBf i b = if a = b then r5dDB else 0 := by  decide +kerneltheorem r5hcond (x : Fin 5 → ℂ) (y : Fin 5 → ℂ)    (hA : ∀ i : Fin 3, ∑ r : Fin 5, ∑ s : Fin 5, x r * (Matrix.of fun r s => r5i (r5A i r s) : Matrix (Fin 5) (Fin 5) ℂ) r s * y s = 0)    (hB : ∀ j : Fin 5, ∑ r : Fin 5, ∑ s : Fin 5, star (x r) * (Matrix.of fun r s => r5i (r5B j r s) : Matrix (Fin 5) (Fin 5) ℂ) r s * y s = 0) :    x = 0 ∨ y = 0 := r5main x y hA hBtheorem r5li_A : LinearIndependent ℂ (fun a : Fin 3 => (Matrix.of fun r s => r5i (r5A a r s) : Matrix (Fin 5) (Fin 5) ℂ)) := by  refine r5li r5A r5pA r5WAf r5dDA ?_ r5hWA  intro h  have := congrArg Complex.re h  rw [r5i_re] at this  simp [r5dDA] at thistheorem r5li_B : LinearIndependent ℂ (fun a : Fin 5 => (Matrix.of fun r s => r5i (r5B a r s) : Matrix (Fin 5) (Fin 5) ℂ)) := by  refine r5li r5B r5pB r5WBf r5dDB ?_ r5hWB  intro h  have := congrArg Complex.re h  rw [r5i_re] at this  simp [r5dDB] at this/-- Existence of range-criterion subspaces for the minimal sporadic quadruple (m, n, k, ℓ) = (5, 5, 3, 5). -/theorem range_criterion_subspaces_sporadic_five_five_three_five :    has_range_criterion_subspaces 5 5 3 5 := by  exact has_range_criterion_of_matrix_pencil 5 5 3 5 (fun a => Matrix.of fun r s => r5i (r5A a r s)) (fun j => Matrix.of fun r s => r5i (r5B j r s)) r5li_A r5li_B r5hcond
See it in the graphDiscuss this resultChallenge it