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.

The next gap in the subrank of 3-tensors

Discreteness of asymptotic tensor ranks

A Gap in the Subrank of Tensors

Countably many asymptotic tensor ranks

Geometric rank of tensors and subrank of matrix multiplication

Asymptotic tensor rank is characterized by polynomials

A note on the gap between rank and border rank

Typical ranks of random order-three tensors

Subrank and optimal reduction of scalar multiplications to generic tensors

Semialgebraic Geometry of Nonnegative Tensor Rank

Notions of Tensor Rank

Subtracting a best rank-1 approximation may increase tensor rank

The span of singular tuples of a tensor beyond the boundary format

The average condition number of most tensor rank decomposition problems is infinite

Exact values of generic subrank

Maximum relative distance between real rank-two and rank-one tensors

The Inverse, Rank and Product of Tensors

Persistent Tensors and Multiqudit Entanglement Transformation

The number of singular vector tuples and uniqueness of best rank one approximation of tensors

Strassen's rank additivity for small tensors, including tensors of rank less or equal 7

Tensor Representation of Rank-Metric Codes