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.

Estimates for the quantized tensor train ranks for the power functions

Multiscale interpolative construction of quantized tensor trains

Estimation of QTT Ranks of Regular Functions on a Uniform Square Grid

Quantitative rank distribution conjecture over $\mathbb{F}_q(t)$

A Geometric Description of Feasible Singular Values in the Tensor Train Format

A Gap in the Subrank of Tensors

The interplay between bounded ranks of tensors arising from partitions

Rank Is All You Need: Estimating the Trace of Powers of Density Matrices

Discreteness of asymptotic tensor ranks

Two-particle calculations with quantics tensor trains -- solving the parquet equations

Randomized algorithms for computing the tensor train approximation and their applications

Operator Quantum Geometric Tensor and Quantum Phase Transitions

Asymptotic tensor rank is characterized by polynomials

On estimating the trace of quantum state powers

Compactness of quantics tensor train representations of local imaginary-time propagators

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

Persistent Tensors and Multiqudit Entanglement Transformation

Probabilistic bounds on best rank-1 approximation ratio

A universal sequence of tensors for the asymptotic rank conjecture

On the tensor rank of multiplication in any extension of $\F_2$