Proposition A.6 — a rank-one nonnegative matrix with entry sum 1 is the product of its row and column sums
ProvedLowRankMisspec.RankOne.proposition_A_6distribution-ranklinear-algebrap2o-gran-per-chapterp2o-plan-paperp2o-v1
Let be a real matrix with finite row and column index sets, nonnegative entries and . If , then every entry equals its row sum times its column sum:
This is the step that turns rank one of each into the factorization in Proposition A.7.
Differences from the source:
- encoding: a matrix A (finite index sets) → A : Matrix m n ℝ with [Fintype m] [Fintype n], m n : Type; reason: any finite index sets; the universe level changes no instance
Preamble
import Mathlib
Formal statement
namespace LowRankMisspec.RankOne
/-- Proposition A.6 (entries of a rank-one probability matrix): a real matrix with nonnegative
entries summing to `1` and rank exactly `1` has every entry equal to its row sum times its column
sum. -/
theorem proposition_A_6 {m n : Type} [Fintype m] [Fintype n] (A : Matrix m n ℝ)
(hnonneg : ∀ i j, 0 ≤ A i j) (hsum : ∑ i, ∑ j, A i j = 1) (hrank : A.rank = 1) :
∀ i j, A i j = (∑ k, A i k) * (∑ l, A l j) := by sorry
end LowRankMisspec.RankOne
Source
Hasrati, Risteski, Understanding Model Misspecification in Low-Rank Distributions: Tolerant Testing and Agnostic Learning, STOC 2027 submission (manuscript sha256 235f2a4f…), p. 15 (PDF 16), Proposition A.6
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.