quadratic_neumann_all_distinct_inner_base_entry_sup_norm_bound_min_dim

Source: Candès--Recht, Exact Matrix Completion via Convex Optimization, PDF p. 28, Section 6.2, Lemma 6.6, equations (6.15)--(6.17), and PDF p. 30, Section 6.3, equation (6.20).

Mathematical statement: Let $S$ be rank-$r$ SVD data for $M\in\mathbb R^{n_1\times n_2}$ satisfying $A0(S,\mu_0)$ and $A1(S,\mu_1)$. For coordinates $w_1,w_2$, let $B^{\mathrm{all}}_{w_1,w_2}$ be quadraticAllDistinctInnerBaseMatrix, the all-distinct inner base matrix in the equation (6.20) branch. Then there is a universal positive constant $C_{\rm entry}$ such that

Notation context: Downstream, $n=\max(n_1,n_2)$, $p=m/(n_1n_2)$ is the Bernoulli rate, and $\Omega$ is the sample set. This deterministic base-bound theorem itself does not quantify over $p$, $\Omega$, or $Z(\Omega)$. The parameters $\mu_0$ and $\mu_1$ are the Candès--Recht incoherence parameters from $A0$ and $A1$.

Formalization note: This is a formal bridge, not a theorem stated verbatim in the paper. It composes the source-backed linear off-diagonal entry base bound linear_neumann_off_diagonal_coefficient_base_entry_sup_norm_bound_min_dim with the source-backed formal bridge quadratic_neumann_all_distinct_inner_base_norms_le_linear_offdiag_base, which says the all-distinct equation (6.20) base matrix is the Lemma 6.6 linear off-diagonal base matrix with one extra coordinate zeroed. This is the corrected min-dimension replacement for the disproved old max-denominator node quadratic_neumann_all_distinct_inner_base_entry_sup_norm_bound.