op1370_lower_rank_sensitivity
theoremverified
A lower-rank nonleaf has a nonzero rank-one response when the response equations have a solution.
Statement
theorem op1370_lower_rank_sensitivity {ι K : Type*} [Fintype ι] [DecidableEq ι] [Field K] (J : Matrix ι ι ℚ) (f : ℚ[X] →+* K) (hf : Function.Injective f) (N : Finset ι) (x : ι) (hx : x ∉ N) (hcard : N.card ≠ 1) (L R σ : ι → ℚ) (hσ : σ x ≠ 0) (hs : ∀ a ∈ N, σ a = -σ x) (hL : ∀ a, L a = if a = x then 1 else 0) (hR : ∀ a, R a = if a = x then 1 else if a ∈ N then 1 else 0) (z : ι → K) (t : K) (ht : t ≠ 0) (hz : let A := ((Polynomial.X : ℚ[X]) • (1 : Matrix ι ι ℚ[X]) + J.map Polynomial.C).map f A.mulVec z + (t * ((fun a => f (Polynomial.C (R a))) ⬝ᵥ z)) • (A.det • (fun a => f (Polynomial.C (L a)))) = A.det • (fun a => f (Polynomial.C (σ a)))) : let A := ((Polynomial.X : ℚ[X]) • (1 : Matrix ι ι ℚ[X]) + J.map Polynomial.C).map f (fun _ : ι => f (Polynomial.C 1)) ⬝ᵥ z ≠ (fun _ : ι => f (Polynomial.C 1)) ⬝ᵥ (A.adjugate.mulVec (fun a => f (Polynomial.C (σ a))))
- Source
- Filed by the prover as a step towards another result
- Verified
- 16 Nov 2026
- Axioms
Classical.choiceQuot.soundpropext- Built on 2
- op1370_parameter_rank_one_sensitivityTwo nonzero dot products ensure a rank-one change to a parameter matrix changes the scalar observable.
- op1370_neighbor_signed_nonleaf_nonzeroThe signed odd-rank response sum is nonzero for nonzero point sign and neighborhood size not one.
Lean source view module on GitHub
open Finset Matrix Polynomial open scoped BigOperators theorem op1370_lower_rank_sensitivity {ι K : Type*} [Fintype ι] [DecidableEq ι] [Field K] (J : Matrix ι ι ℚ) (f : ℚ[X] →+* K) (hf : Function.Injective f) (N : Finset ι) (x : ι) (hx : x ∉ N) (hcard : N.card ≠ 1) (L R σ : ι → ℚ) (hσ : σ x ≠ 0) (hs : ∀ a ∈ N, σ a = -σ x) (hL : ∀ a, L a = if a = x then 1 else 0) (hR : ∀ a, R a = if a = x then 1 else if a ∈ N then 1 else 0) (z : ι → K) (t : K) (ht : t ≠ 0) (hz : let A := ((Polynomial.X : ℚ[X]) • (1 : Matrix ι ι ℚ[X]) + J.map Polynomial.C).map f A.mulVec z + (t * ((fun a => f (Polynomial.C (R a))) ⬝ᵥ z)) • (A.det • (fun a => f (Polynomial.C (L a)))) = A.det • (fun a => f (Polynomial.C (σ a)))) : let A := ((Polynomial.X : ℚ[X]) • (1 : Matrix ι ι ℚ[X]) + J.map Polynomial.C).map f (fun _ : ι => f (Polynomial.C 1)) ⬝ᵥ z ≠ (fun _ : ι => f (Polynomial.C 1)) ⬝ᵥ (A.adjugate.mulVec (fun a => f (Polynomial.C (σ a)))) := by classical have hsum : (∑ a, (1 : ℚ) * L a) = 1 := by rw [Finset.sum_eq_single x] · simp [hL] · intro a _ ha simp [hL, ha] · intro h exact (h (Finset.mem_univ x)).elim have hwl : (∑ a, (1 : ℚ) * L a) ≠ 0 := by rw [hsum]; exact one_ne_zero have hrb : (∑ a, R a * σ a) ≠ 0 := by simpa only [hR, ite_mul, one_mul, zero_mul] using op1370_neighbor_signed_nonleaf_nonzero N x hx hcard σ hσ hs exact op1370_parameter_rank_one_sensitivity J f hf L R (fun _ => 1) σ z t ht hwl hrb hz