Abstract:Asymptotic tensor rank is notoriously difficult to determine. Indeed, determining its value for the $2\times 2$ matrix multiplication tensor would determine the matrix multiplication exponent, a long-standing open problem. On the other hand, Strassen's asymptotic rank conjecture makes the bold claim that asymptotic tensor rank equals the largest dimension of the tensor and is thus as easy to compute as matrix rank. Despite tremendous interest, much is still unknown about the structural and computational properties of asymptotic rank; for instance whether it is computable. We prove that asymptotic tensor rank is "computable from above", that is, for any real number $r$ there is an (efficient) algorithm that determines, given a tensor $T$, if the asymptotic tensor rank of $T$ is at most $r$. The algorithm has a simple structure; it consists of evaluating a finite list of polynomials on the tensor. Indeed, we prove that the sublevel sets of asymptotic rank are Zariski-closed (just like matrix rank). While we do not exhibit these polynomials explicitly, their mere existence has strong implications on the structure of asymptotic rank. As one such implication, we find that the values that asymptotic tensor rank takes, on all tensors, is a well-ordered set. In other words, any non-increasing sequence of asymptotic ranks stabilizes ("discreteness from above"). In particular, for the matrix multiplication exponent (which is an asymptotic rank) there is no sequence of exponents of bilinear maps that approximates it arbitrarily closely from above without being eventually constant. In other words, any upper bound on the matrix multiplication exponent that is close enough, will "snap" to it. Previously such discreteness results were only known for finite fields or for other tensor parameters (e.g., asymptotic slice rank). We obtain them for infinite fields like the complex numbers.

Rank two perturbations of matrices and operators and operator model for t-transformation of probability measures

Eigenvalue and pseudospectrum processes generated by nonnormal Toeplitz matrices with rank 1 perturbations

Symmetries and Wigner representations of operational theories

Finite rank perturbations of normal operators: hyperinvariant subspaces and a problem of Pearcy

Operator means of probability measures and generalized Karcher equations

Global properties of eigenvalues of parametric rank one perturbations for unstructured and structured matrices II

Perturbation and Inverse Problems of Stochastic Matrices

Boundedness of Metaplectic Operators Within Spaces, Applications to Pseudodifferential Calculus, and Time–Frequency Representations

Finite rank perturbation of non-Hermitian random matrices: heavy tail and sparse regimes

Metric Transforms and Low Rank Matrices via Representation Theory of the Real Hyperrectangle

Representation Theory for Risk On Markowitz-Tversky-Kahneman Topology

On the effect of perturbation of conditional probabilities in total variation

Outliers and bounded rank perturbation for non-Hermitian random band matrices

A Riemann--Hilbert approach to the perturbation theory for orthogonal polynomials: Applications to numerical linear algebra and random matrix theory

Determinantal Formulas for Rational Perturbations of Multiple Orthogonality Measures

Expansion of eigenvalues of rank-one perturbations of the discrete bilaplacian

Probabilistic perturbation bounds of matrix decompositions

Understanding of linear operators through Wigner analysis

A geometric approach to finite rank unitary perturbations

Asymptotic tensor rank is characterized by polynomials

A Class of Transformation Matrices and Its Applications