Eigenvectors and eigenvalues in a random subspace of a tensor product (Q695296)

\textit{B. Collins} and \textit{I. Nechita} [Adv. Math. 226, No. 2, 1181--1201 (2011; Zbl 1203.81022)] observed that with very high probability some sequences of numbers in \(\mathbb R^k_+\) never occur as singular values of elements in a randomly chosen subspace \(V_n\) (of relative dimension \(t\) for large \(n\) and fixed \(k\)) of \(\mathbb C^k \otimes \mathbb C^n\) when \(n\) becomes large. Motivated by this, the authors in this paper consider the following setting. Given two positive integers \(n\) and \(k\) and a parameter \(t \in (0, 1)\), choose at random a vector subspace \(V_n \subset \mathbb C^k\otimes \mathbb C^n\) of dimension \(N\sim tnk\). They show that the set of \(k\)-tuples of singular values of all unit vectors in \(V_n\) fills asymptotically (as \(n\) tends to infinity) a deterministic convex set \(K_{k,t}\) that is described by a new norm \(\|\cdot\|_{(t)}\) in \(\mathbb R^k\) (see Section 3). The norm is obtained via an operator algebraic construction and its restriction on \(\mathbb R^k\) interpolates between the \(l^1\)-norm and \(l^\infty\)-norm. The proof relies on free probability, random matrix theory, complex analysis and matrix analysis techniques. The main result comes together with a law of large numbers for the singular value decomposition of the eigenvectors corresponding to large eigenvalues of a random truncation of a matrix with high eigenvalue degeneracy. Of independent interest, it is shown that in some sense the eigenvectors of some random matrices are much more deterministic than one might expect (Theorem 1.3). The description makes use of the unitary group.