rank_one_outer_product_square_collapse
Provedcandes-rechtlinear-algebramatrix-completionreferencerudelson
Rank-one square collapse (yy*)^2 = ||y||^2 (yy*). For the rank-one outer product y (x) y = vecMulVec y y, its matrix square equals the scalar ||y||^2 = y . y times itself: (y(x)y)(y(x)y) = (y . y) . (y(x)y). This is the algebraic identity that makes the Gram square of the rank-one tensor operators collapse in Rudelson's noncommutative-Khintchine selection lemma. Proof: Matrix.vecMulVec_mul_vecMulVec gives vecMulVec y y * vecMulVec y y = vecMulVec y ((y . y) . y), then entrywise ring.
Preamble
import Mathlib.LinearAlgebra.Matrix.Trace import Mathlib.Data.Matrix.Mul import Mathlib.Data.Real.Basic open scoped BigOperators Matrix
Formal statement
theorem rank_one_outer_product_square_collapse {d : Nat} (y : Fin d → Real) : Matrix.vecMulVec y y * Matrix.vecMulVec y y = (y ⬝ᵥ y) • Matrix.vecMulVec y y := by sorrySource
Rudelson, 'Random vectors in the isotropic position', J. Funct. Anal. 164 (1999), Thm 1 pp.3-6; van Handel, 'Structured Random Matrices', arXiv:1610.05200 Sec. 3 (Lemma 3.5); Candes-Recht 2009 (arXiv:0805.4471) Sec. 4.2 eq (4.9) p.18. These are the rank-one (outer-product) operator identities used by the noncommutative-Khintchine step of Rudelson's selection lemma; the rank-one summands are X_c = y_c (x) y_c on the Hilbert-Schmidt space, and the Gram square collapses by (yy*)^2 = ||y||^2 (yy*).