OAI/Analysis/KadisonKastler/CorrectionBounds.lean
6 declarations, 0 contain sorry, 0 declared as axioms
Bounds for correcting a transported multiplication
These are the inverse and composition estimates used in the UCP criterion of §2 of Universal strong Kadison–Kastler stability (2026). The estimates are formulated for an explicitly supplied correction map.
- theorem
OAI.KadisonKastler.norm_inverse_le_of_close_to_idline 21Source documentation
The inverse norm estimate for a linear equivalence within
η < 1of the identity.theorem norm_inverse_le_of_close_to_id (L : E ≃ₗ[ℂ] E) {η : ℝ} (hη : η < 1) (hclose : ∀ x, ‖L x - x‖ ≤ η * ‖x‖) (x : E) : ‖L.symm x‖ ≤ (1 - η)⁻¹ * ‖x‖ := by … - theorem
OAI.KadisonKastler.norm_inverse_sub_le_of_close_to_idline 33Source documentation
The inverse displacement estimate corresponding to a near-identity linear equivalence.
theorem norm_inverse_sub_le_of_close_to_id (L : E ≃ₗ[ℂ] E) {η : ℝ} (hη₀ : 0 ≤ η) (hη : η < 1) (hclose : ∀ x, ‖L x - x‖ ≤ η * ‖x‖) (x : E) : ‖L.symm x - x‖ ≤ (η / (1 - η)) * ‖x‖ := by … - theorem
OAI.KadisonKastler.norm_correctedMap_sub_leline 46Source documentation
Composing a contraction with an inverse correction adds at most the inverse displacement to the original comparison error.
theorem norm_correctedMap_sub_le (L : E ≃ₗ[ℂ] E) (F ρ : E →ₗ[ℂ] G) {η e : ℝ} (hη₀ : 0 ≤ η) (hη : η < 1) (hL : ∀ x, ‖L x - x‖ ≤ η * ‖x‖) (hF : ∀ x, ‖F x‖ ≤ ‖x‖) (hclose : ∀ x, ‖F x - ρ x‖ ≤ e * ‖x‖) (x : E) : ‖F (L.symm x) - ρ x‖ ≤ (e + η / (1 - η)) * ‖x‖ := by … - theorem
OAI.KadisonKastler.ucp_correction_parameter_boundsline 63Source documentation
The correction parameter furnished by the multiplication estimate is small enough for inversion at the manuscript's UCP tolerance.
theorem ucp_correction_parameter_bounds {e : ℝ} (he : 0 ≤ e) (he₁ : e < 1 / 1000) : 0 ≤ 30 * e / (1 - e) ∧ 30 * e / (1 - e) < 1 := by … - theorem
OAI.KadisonKastler.ucp_correction_constant_leline 73Source documentation
The numerical estimate in the UCP criterion:
e + η / (1 - η) ≤ 32eforη = 30e / (1 - e).theorem ucp_correction_constant_le {e : ℝ} (he : 0 ≤ e) (he₁ : e < 1 / 1000) : e + (30 * e / (1 - e)) / (1 - 30 * e / (1 - e)) ≤ 32 * e := by … - theorem
OAI.KadisonKastler.norm_correctedMap_sub_le_thirty_two_mulline 93Source documentation
The corrected map is within
32eof the comparison map under the explicit correction bound appearing in the UCP criterion.theorem norm_correctedMap_sub_le_thirty_two_mul (L : E ≃ₗ[ℂ] E) (F ρ : E →ₗ[ℂ] G) {e : ℝ} (he : 0 ≤ e) (he₁ : e < 1 / 1000) (hL : ∀ x, ‖L x - x‖ ≤ (30 * e / (1 - e)) * ‖x‖) (hF : ∀ x, ‖F x‖ ≤ ‖x‖) (hclose : ∀ x, ‖F x - ρ x‖ ≤ e * ‖x‖) (x : E) : ‖F (L.symm x) - ρ x‖ ≤ (32 * e) * ‖x‖ := by …