quadratic_neumann_first_index_distinct_mean_coefficient_lemma68_bound_min_dim
Disprovedcandes-rechtcoefficient-boundlemma-68matrix-completionsection-63source-faithful
Source: Candes-Recht 2008, Section 6.3, PDF pp. 31--32, Lemma 6.8, equation (6.22).
This theorem records the source-faithful rectangular entrywise bound for the deterministic coefficient matrix in the mean part of the quadratic Neumann case. It keeps both terms in Lemma 6.8, including the -dependent contribution, and uses the repaired rectangular scale instead of the deprecated -free max-denominator collapse.
Preamble
import Definitions.Def_matrix_completion_neumann open MatrixCompletion
Formal statement
theorem quadratic_neumann_first_index_distinct_mean_coefficient_lemma68_bound_min_dim :
∃ C68 : ℝ, 0 < C68 ∧
∀ (n₁ n₂ r : ℕ) (M : Matrix (Fin n₁) (Fin n₂) ℝ)
(μ₀ μ₁ : ℝ) (S : SVD M r) (p : ℝ),
0 < n₁ → 0 < n₂ → 0 < r →
1 ≤ μ₀ → 1 ≤ μ₁ →
A0 S μ₀ → A1 S μ₁ →
entrySupNorm (quadraticFirstIndexDistinctMeanCoefficientMatrix S p) ≤
C68 * p⁻¹ *
(μ₀ * (r : ℝ) / (↑(min n₁ n₂))) *
(μ₁ * Real.sqrt ((r : ℝ) / ((n₁ : ℝ) * (n₂ : ℝ))) +
μ₀ * (r : ℝ) / (↑(min n₁ n₂))) := by
sorry
Source
Deprecated by matrix-completion-leaf-batch-solver on 2026-07-01 after source audit: the statement is not source-faithful as written. It uses a rectangular `max(n_1,n_2)` denominator or otherwise omits required positivity/scale hypotheses; Candes-Recht Section 6.3 and Lemma 6.4/Lemma 6.8 give the corresponding `min(n_1,n_2)`/positive-rate scales. No replacement theorem is activated here because the parent rate absorption also needs a corrected route.