sampled_energy_scale_moment_bound_from_row_column_energy_moment
DisprovedRole. It controls sampled row/column counts or energies, which feed the moment bounds for random sampled matrices.
Problem and notation. Exact matrix completion asks when an unknown low-rank real matrix can be recovered from a random subset of its entries. Here has rank , entries are observed, and . Recovery means nuclear-norm minimization: minimize among matrices agreeing with on the observed entries. Probability notation. is the fixed-cardinality success probability: is chosen uniformly among all subsets of entries with , and the event is that the convex program uniquely returns . In Bernoulli nodes, or means each entry is sampled independently with probability , usually . Coherence notation. The object records SVD/singular-vector data for . The hypotheses and are the Candes-Recht incoherence assumptions: measures how spread out the singular vector spaces are, and measures the largest entry of the sign matrix . The parameter controls polynomial failure probabilities such as . For sampled row/column nodes, counts observed entries in row , counts observed entries in column , and the corresponding energies sum over sampled entries. These estimates feed the noncommutative Khintchine and spectral-norm concentration bounds.
Claim. Algebraic moment conversion after Khintchine: a moment bound for the maximum sampled row/column energy gives the moment bound for , at the displayed scale.
Lecture-note formulation:
The constants in this node are universal existential constants; the theorem asserts that some positive constants with these roles exist.
Decomposition status. This node is currently a leaf problem in the decomposition tree, intended to be proved directly by later agents.
import Definitions.Def_matrix_completion_rademacher open MatrixCompletion
theorem sampled_energy_scale_moment_bound_from_row_column_energy_moment
(Cenergy Ckh : ℝ) :
0 < Cenergy →
0 < Ckh →
∃ Crad : ℝ, 0 < Crad ∧
∀ (n₁ n₂ m q : ℕ) (X : Matrix (Fin n₁) (Fin n₂) ℝ),
bernoulliExpectation ((m : ℝ) / ((n₁ : ℝ) * (n₂ : ℝ)))
(fun Omega =>
(max (sampledRowEnergyMax Omega X)
(sampledColumnEnergyMax Omega X)) ^ q) ≤
(Cenergy * ((m : ℝ) / ((n₁ : ℝ) * (n₂ : ℝ))) *
(↑(max n₁ n₂)) * entrySupNorm X ^ 2) ^ q →
bernoulliExpectation ((m : ℝ) / ((n₁ : ℝ) * (n₂ : ℝ)))
(fun Omega =>
(Ckh * Real.sqrt (q : ℝ) *
(((m : ℝ) / ((n₁ : ℝ) * (n₂ : ℝ)))⁻¹) *
Real.sqrt
(max (sampledRowEnergyMax Omega X)
(sampledColumnEnergyMax Omega X))) ^ q) ≤
(Crad * Real.sqrt
(((q : ℝ) * (↑(max n₁ n₂))) /
((m : ℝ) / ((n₁ : ℝ) * (n₂ : ℝ)))) *
entrySupNorm X) ^ q := by
sorry