Anderson's paving conjecture, now known to hold due to the resolution of the Kadison-Singer problem asserts that every zero diagonal Hermitian matrix admits non-trivial pavings with dimension independent bounds. In this paper, we develop a technique extending the arguments of Marcus, Spielman and Srivastava in their solution of the Kadison-Singer problem to show the existence of non-trivial pavings for collections of matrices. We show that given zero diagonal Hermitian contractions A (1) , · · · , A (k) ∈ M n (C) and ǫ > 0, one may find a pavingAs a consequence, we get the correct asymptotic estimates for paving general zero diagonal matrices; zero diagonal contractions can be (O(ǫ −2 ), ǫ) paved. As an application, we give a simplified proof wth slightly better estimates of a theorem of Johnson, Ozawa and Schechtman concerning commutator representations of zero trace matrices.
IntroductionAnderson's Paving conjecture [And79] asserts that for every ǫ > 0 there is an integer r such that every zero diagonal complex n × n matrix M admits a paving X 1 ∪ . . . ∪ X r = [n], which satisfies:where the P X i are diagonal projections. This conjecture, which implies a positive solution to the Kadison-Singer Problem, is proven in [MSS15] with the dependence r = (6/ǫ) 8 which arises as follows. First, the "one-sided" boundis obtained with r = (6/ǫ) 2 for Hermitian M , via the method of interlacing families of polynomials. Then this is extended to a two-sided bound by taking a product of pavings for M and −M , and finally to non-Hermitian M = A+iB by taking yet another product of two-sided pavings of A and B. Each product increases r quadratically, and in general this approach allows one to simultaneously pave any k Hermitian matrices in the one-sided sense with the dependence (6/ǫ) 2k . Our main theorem says, Theorem 1 Given zero diagonal Hermitian contractions A (1) , · · · , A (k) ∈ M n (C) and ǫ > 0, there exists a paving X 1 ∐ · · · ∐ X r = [n] where r ≤ 18kǫ −2 such that,