Exact inverse-transpose coordinates from a variable-dimensional integer basis
ProvedInfoLossLAR.WorstCase.RationalInverse.Native.current_correctcomplexitylatticesquantum-computing
For every nonsingular integer matrix of dimension , the row-major rational matrix returned by the native basis algorithm is exactly . More explicitly, the returned entry in row , column is the th Cramer coordinate for , divided by . This identifies the vectors used to encode fine-lattice queries with the exact dual basis in every dimension.
Preamble
import Definitions.Def_InfoLossLAR_ReciprocallyScaledGaussianSource set_option autoImplicit false
Formal statement
namespace InfoLossLAR.WorstCase.RationalInverse.Native
theorem current_correct {n : ℕ} (B : Matrix (Fin n) (Fin n) ℤ) (hdet : B.det≠0) :
current (inputRows B) = List.ofFn (fun i : Fin n => List.ofFn (fun j : Fin n =>
(Matrix.cramer B.transpose (Pi.single j 1) i : ℚ)/(B.det : ℚ))) := by sorry
end InfoLossLAR.WorstCase.RationalInverse.Native
Source
InfoLossLAR formalization of physical Fourier sampling from a shared bounded-distance decoder. https://prove2.me/theorems/d106aba4-0ad7-4239-9c6b-4a9dc8c9ebc4