Exact twelve-point instance of Theorem 5.5
DefinitionSampleRemoval_Supplement_points55least-squaresoptimizationp2o-gran-per-chapterp2o-plan-paperp2o-v1robustness-auditsample-removal
These are the twelve planar points displayed in Theorem 5.5. Every terminating decimal is stored as the exact corresponding rational number in , so the audit threshold and range obstruction are not floating-point claims.
Definition code
import Mathlib set_option autoImplicit false namespace SampleRemoval.Supplement /-- The exact twelve-point rational audit instance of Theorem 5.5. Each printed terminating decimal is interpreted as its exact rational value (Xu, Gao, Hu, STOC 2027 submission, Theorem 5.5, p.11 (PDF p.13)). -/ noncomputable def points55 : Fin 12 → ℝ × ℝ := fun i => match i.1 with | 0 => (2082 / 1000, 240 / 1000) | 1 => (3235 / 1000, -(14 : ℝ) / 1000) | 2 => (189 / 1000, 1636 / 1000) | 3 => (-(4066 : ℝ) / 1000, 2816 / 1000) | 4 => (6438 / 1000, -(320 : ℝ) / 1000) | 5 => (-(1749 : ℝ) / 1000, -(1697 : ℝ) / 1000) | 6 => (-(1757 : ℝ) / 1000, -(1703 : ℝ) / 1000) | 7 => (-(1744 : ℝ) / 1000, -(1692 : ℝ) / 1000) | 8 => (-(1739 : ℝ) / 1000, -(1695 : ℝ) / 1000) | 9 => (-(1752 : ℝ) / 1000, -(1702 : ℝ) / 1000) | 10 => (-(1730 : ℝ) / 1000, -(1678 : ℝ) / 1000) | _ => (-(1761 : ℝ) / 1000, -(1678 : ℝ) / 1000) end SampleRemoval.Supplement
Source
Xu, Gao, Hu, Rounding, Ranges, and Exact Sample-Removal Audits for Least Squares, STOC 2027 submission (manuscript sha256 dcd33acebc1ccdc626c7ee41bfabc29795abe2e1cd9267d8cbd66d98f81c4b26), Theorem 5.5, p.11 (PDF p.13)