Maximal border subrank tensors

Chia-Yu Chang

DOI: https://doi.org/10.1080/03081087.2024.2352456

2024-05-28

Linear and Multilinear Algebra

Abstract:We prove a lower bound on the dimension of the set of maximal border subrank tensors. This is the first such bound of its type.

mathematics

What problem does this paper attempt to address?

Border subrank via a generalised Hilbert-Mumford criterion

Benjamin Biaggi,Chia-Yu Chang,Jan Draisma,Filip Rupniewski

2024-02-16

Abstract:We show that the border subrank of a sufficiently general tensor in $(\mathbb{C}^n)^{\otimes d}$ is $\mathcal{O}(n^{1/(d-1)})$ for $n \to \infty$. Since this matches the growth rate $\Theta(n^{1/(d-1)})$ for the generic (non-border) subrank recently established by Derksen-Makam-Zuiddam, we find that the generic border subrank has the same growth rate. In our proof, we use a generalisation of the Hilbert-Mumford criterion that we believe will be of independent interest.

Algebraic Geometry,Computational Complexity
Border rank bounds for $GL(V)$-invariant tensors arising from matrices of constant rank

Derek Wu

2024-05-10

Abstract:We prove border rank bounds for a class of $GL(V)$-invariant tensors in $V^*\otimes U\otimes W$, where $U$ and $W$ are $GL(V)$-modules. These tensors correspond to spaces of matrices of constant rank. In particular we prove lower bounds for tensors in $\mathbb{C}^l\otimes\mathbb{C}^m\otimes\mathbb{C}^n$ that are not $1_A$-generic, where no nontrivial bounds were known, and also when $l,m\ll n$, where previously only bounds for unbalanced matrix multiplication tensors were known. We give the first explicit use of Young flattenings for tensors beyond Koszul to obtain border rank lower bounds, and determine the border rank of three tensors.

Algebraic Geometry
Concise tensors of minimal border rank

Joachim Jelisiejew,J.M. Landsberg,Arpan Pal

2024-08-27

Abstract:We determine defining equations for the set of concise tensors of minimal border rank in $C^m\otimes C^m\otimes C^m$ when $m=5$ and the set of concise minimal border rank $1_*$-generic tensors when $m=5,6$. We solve this classical problem in algebraic complexity theory with the aid of two recent developments: the 111-equations defined by Buczyńska-Buczyński and results of Jelisiejew-Šivic on the variety of commuting matrices. We introduce a new algebraic invariant of a concise tensor, its 111-algebra, and exploit it to give a strengthening of Friedland's normal form for $1$-degenerate tensors satisfying Strassen's equations. We use the 111-algebra to characterize wild minimal border rank tensors and classify them in $C^5\otimes C^5\otimes C^5$.

Algebraic Geometry,Computational Complexity,Commutative Algebra
Symmetrization maps and minimal border rank Comon's conjecture

Tomasz Mańdziuk,Emanuele Ventura

2024-11-09

Abstract:One of the fundamental open problems in the field of tensors is the border Comon's conjecture: given a symmetric tensor $F\in(\mathbb{C}^n)^{\otimes d}$ for $d\geq 3$, its border and symmetric border ranks are equal. In this paper, we prove the conjecture for large classes of concise tensors in $(\mathbb{C}^n)^{\otimes d}$ of border rank $n$, i.e., tensors of minimal border rank. These families include all tame tensors and all tensors whenever $n\leq d+1$. Our technical tools are border apolarity and border varieties of sums of powers.

Algebraic Geometry,Computational Complexity
A Gap in the Subrank of Tensors

Matthias Christandl,Fulvio Gesmundo,Jeroen Zuiddam

DOI: https://doi.org/10.1137/22M1543276

2023-11-24

Abstract:The subrank of tensors is a measure of how much a tensor can be ''diagonalized''. This parameter was introduced by Strassen to study fast matrix multiplication algorithms in algebraic complexity theory and is closely related to many central tensor parameters (e.g. slice rank, partition rank, analytic rank, geometric rank, G-stable rank) and problems in combinatorics, computer science and quantum information theory. Strassen (J. Reine Angew. Math., 1988) proved that there is a gap in the subrank when taking large powers under the tensor product: either the subrank of all powers is at most one, or it grows as a power of a constant strictly larger than one. In this paper, we precisely determine this constant for tensors of any order. Additionally, for tensors of order three, we prove that there is a second gap in the possible rates of growth. Our results strengthen the recent work of Costa and Dalai (J. Comb. Theory, Ser. A, 2021), who proved a similar gap for the slice rank. Our theorem on the subrank has wider applications by implying such gaps not only for the slice rank, but for any ``normalized monotone''. In order to prove the main result, we characterize when a tensor has a very structured tensor (the W-tensor) in its orbit closure. Our methods include degenerations in Grassmanians, which may be of independent interest.

Algebraic Geometry,Computational Complexity,Combinatorics,Quantum Physics
Classification and degenerations of small minimal border rank tensors via modules

Jakub Jagiełła,Joachim Jelisiejew

2024-09-10

Abstract:We give a self-contained classification of $1_*$-generic minimal border rank tensors in $C^m \otimes C^m \otimes C^m$ for $m \leq 5$. Together with previous results, this gives a classification of all minimal border rank tensors in $C^m \otimes C^m \otimes C^m$ for $m \leq 5$: there are $37$ isomorphism classes. We fully describe possible degenerations among the tensors. We prove that there are no $1$-degenerate minimal border rank tensors in $C^m \otimes C^m \otimes C^m $ for $m \leq 4$.

Algebraic Geometry,Computational Complexity,Commutative Algebra
Subrank and optimal reduction of scalar multiplications to generic tensors

Harm Derksen,Visu Makam,Jeroen Zuiddam

DOI: https://doi.org/10.1112/jlms.12963

2024-07-14

Journal of the London Mathematical Society

Abstract:Abstract The subrank of a tensor measures how much a tensor can be diagonalized. We determine this parameter precisely for essentially all (i.e., generic) tensors. Namely, we prove for generic tensors in with that the subrank is . Our result significantly improves on the previous upper bound from the work of Strassen (1991) and Bürgisser (1990) which was . Our result is tight up to an additive constant. Our full result covers not only 3‐tensors but also ‐tensors, for which we find that the generic subrank is . Moreover, as an application, we prove that the subrank is not additive under the direct sum. As a consequence of our result, we obtain several large separations between the subrank and tensor methods that have received much interest recently, notably the slice rank (Tao, 2016), analytic rank (Gowers–Wolf, 2011; Lovett, 2018; Bhrushundi–Harsha–Hatami–Kopparty–Kumar, 2020), geometric rank (Kopparty–Moshkovitz–Zuiddam, 2020), and G‐stable rank (Derksen, 2020). Our proofs of the lower bounds rely on a new technical result about an optimal decomposition of tensor space into structured subspaces, which we think may be of independent interest.

mathematics
On tensors of border rank $l$ in $\C^{m\times n\times l}$

Shmuel Friedland

DOI: https://doi.org/10.48550/arXiv.1003.1968

2010-03-09

Algebraic Geometry

Abstract:We study tensors in $\C^{m\times n\times l}$ whose border rank is $l$. We characterize the tensors in $\C^{3\times 3\times 4}$ and in $\C^{4\times 4\times 4}$ of border rank 4 at most.
A note on the gap between rank and border rank

Jeroen Zuiddam

DOI: https://doi.org/10.1016/j.laa.2017.03.015

2017-03-23

Abstract:We study the tensor rank of the tensor corresponding to the algebra of n-variate complex polynomials modulo the dth power of each variable. As a result we find a sequence of tensors with a large gap between rank and border rank, and thus a counterexample to a conjecture of Rhodes. At the same time we obtain a new lower bound on the tensor rank of tensor powers of the generalised W-state tensor. In addition, we exactly determine the tensor rank of the tensor cube of the three-party W-state tensor, thus answering a question of Chen et al.

Commutative Algebra,Discrete Mathematics,Quantum Physics
Geometric rank of tensors and subrank of matrix multiplication

Swastik Kopparty,Guy Moshkovitz,Jeroen Zuiddam

DOI: https://doi.org/10.48550/arXiv.2002.09472

2020-02-21

Computational Complexity

Abstract:Motivated by problems in algebraic complexity theory (e.g., matrix multiplication) and extremal combinatorics (e.g., the cap set problem and the sunflower problem), we introduce the geometric rank as a new tool in the study of tensors and hypergraphs. We prove that the geometric rank is an upper bound on the subrank of tensors and the independence number of hypergraphs. We prove that the geometric rank is smaller than the slice rank of Tao, and relate geometric rank to the analytic rank of Gowers and Wolf in an asymptotic fashion. As a first application, we use geometric rank to prove a tight upper bound on the (border) subrank of the matrix multiplication tensors, matching Strassen's well-known lower bound from 1987.
Affine subspaces of matrices with rank in a range

Elena Rubei

DOI: https://doi.org/10.13001/ela.2024.8817

2024-10-05

Electronic Journal of Linear Algebra

Abstract:The problem of finding the maximal dimension of linear or affine subspaces of matrices whose rank is constant, or bounded below, or bounded above, has attracted many mathematicians from the sixties to the present day. The problem has caught also the attention of algebraic geometers, since vector spaces of matrices of constant rank give rise to vector bundle maps whose images are vector bundles of rank . Moreover, there is a link with the so-called "rank metric codes," since a constant rank subspace of can be viewed as a constant weight rank metric code; it can be interesting to study also the maximal dimension of the subspaces of whose elements have rank in a range , since such subspaces obviously give rank metric codes with weights in . In this paper, with the main purpose to get an organic result including the ones on spaces of matrices with constant rank, the ones on spaces of matrices with rank bounded below and the ones on spaces of matrices with rank bounded above and to generalize a previous result on real matrices with constant rank to matrices on a more general field, we study the maximal dimension of affine subspaces of matrices whose rank is between two numbers under mild assumptions on the field. We get also a result on antisymmetric matrices and on matrices in row echelon form.

mathematics
General tail bounds for random tensors summation: Majorization approach

Shih Yu Chang,Yimin Wei

DOI: https://doi.org/10.1016/j.cam.2022.114533

IF: 2.872

2022-01-01

Journal of Computational and Applied Mathematics

Abstract:In recent years, tensors have been applied to different applications in science and engineering fields. In order to establish theory about tail bounds of the tensors summation behavior, this research extends previous work by considering the tensors summation tail behavior of the top k-largest singular values of a function of the tensors summation, instead of the largest/smallest singular value of the tensors summation directly (identity function) explored in Chang (2020). Majorization and antisymmetric tensor product tools are main techniques utilized to establish inequalities for unitarily invariant norms of multivariate tensors. The Laplace transform method is integrated with these inequalities for unitarily invariant norms of multivariate tensors to give us tail bounds estimation for the Ky Fan k-norm for a function of the tensors summation. By restricting different random tensor conditions, we obtain generalized tensor Chernoff and Bernstein inequalities.
Maximal invariant subspaces for a class of operators

Kunyu Guo,Wei He,Shengzhao Hou

DOI: https://doi.org/10.1007/s11512-009-0109-1

2009-01-01

Arkiv för matematik

Abstract:In this note, we characterize maximal invariant subspaces for a class of operators. Let T be a Fredholm operator and 1-TT^*∈𝒮_p for some p ≥1. It is shown that if M is an invariant subspace for T such that dim M ⊖ TM <∞, then every maximal invariant subspace of M is of codimension 1 in M . As an immediate consequence, we obtain that if M is a shift invariant subspace of the Bergman space and dim M ⊖ zM <∞, then every maximal invariant subspace of M is of codimension 1 in M . We also apply the result to translation operators and their invariant subspaces.
Probabilistic bounds on best rank-1 approximation ratio

Khazhgali Kozhasov,Josué Tonelli-Cueto

DOI: https://doi.org/10.1080/03081087.2024.2304146

2024-03-05

Linear and Multilinear Algebra

Abstract:We provide new upper and lower bounds on the minimum possible ratio of the spectral and Frobenius norms of a (partially) symmetric tensor. In the particular case of general tensors, our result recovers a known upper bound. For symmetric tensors, our upper bound unveils that the ratio of norms has the same order of magnitude as the trivial lower bound 1/nd−12 , when the order of a tensor d is fixed and the dimension of the underlying vector space n tends to infinity. However, when n is fixed and d tends to infinity, our lower bound is better than 1/nd−12 .

mathematics
The Maximum Principle for Minimal Varieties of Arbitrary Codimension

Brian White

DOI: https://doi.org/10.48550/arXiv.0906.0189

2010-11-12

Abstract:We prove that an m-dimensional minimal variety in a Riemannian manifold cannot touch the boundary at a point where the sum of the smallest m principal curvatures is greater than 0. We also prove an analogous result for varieties with bounded mean curvature.

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

Henrik Eisenmann,André Uschmajew

DOI: https://doi.org/10.48550/arXiv.2111.12611

2022-09-26

Abstract:It is shown that the relative distance in Frobenius norm of a real symmetric order-$d$ tensor of rank two to its best rank-one approximation is upper bounded by $\sqrt{1-(1-1/d)^{d-1}}$. This is achieved by determining the minimal possible ratio between spectral and Frobenius norm for symmetric tensors of border rank two, which equals $\left(1-{1}/{d}\right)^{(d-1)/{2}}$. These bounds are also verified for arbitrary real rank-two tensors by reducing to the symmetric case.

Algebraic Geometry,Optimization and Control
Border apolarity and varieties of sums of powers

Tomasz Mańdziuk,Emanuele Ventura

DOI: https://doi.org/10.48550/arXiv.2310.19625

2024-02-14

Abstract:We study border varieties of sums of powers ($\underline{\mathrm{VSP}}$'s for short), recently introduced by Buczyńska and Buczyński, parameterizing border rank decompositions of a point (e.g. of a tensor or a homogeneous polynomial) with respect to a smooth projective toric variety and living in the Haiman-Sturmfels multigraded Hilbert scheme. Their importance stems from the role of border tensor rank in theoretical computer science, especially in the estimation of the exponent of matrix multiplication, a fundamental and still unknown quantity in the theory of computation. We compare $\underline{\mathrm{VSP}}$'s to other well-known loci in the Hilbert scheme, parameterizing scheme-theoretic versions of decompositions. The latter ones are crucial in that they naturally explain the existing severe barriers to giving good lower bounds on ranks. We introduce the notion of border identifiability and provide sufficient criteria for its appearance, which rely on the multigraded regularity of Maclagan and Smith. We link border identifiability to wildness of points. Finally, we determine $\underline{\mathrm{VSP}}$'s in several instances and regimes, in the contexts of tensors and homogeneous polynomials. These include concise $3$-tensors of minimal border rank and in particular of border rank three.

Algebraic Geometry
On the ranks and border ranks of symmetric tensors

J.M. Landsberg,Zach Teitler

DOI: https://doi.org/10.48550/arXiv.0901.0487

2009-09-29

Abstract:Motivated by questions arising in signal processing, computational complexity, and other areas, we study the ranks and border ranks of symmetric tensors using geometric methods. We provide improved lower bounds for the rank of a symmetric tensor (i.e., a homogeneous polynomial) obtained by considering the singularities of the hypersurface defined by the polynomial. We obtain normal forms for polynomials of border rank up to five, and compute or bound the ranks of several classes of polynomials, including monomials, the determinant, and the permanent.

Algebraic Geometry
Exact values of generic subrank

Paweł Pielasa,Matouš Šafránek,Anatoli Shatsila

2024-08-14

Abstract:In this article we prove the subrank of a generic tensor in $\mathbb{C}^{n,n,n}$ to be $Q(n) = \lfloor\sqrt{3n - 2}\rfloor$ by providing a lower bound to the known upper bound. More generally, we find the generic subrank of tensors of all orders and dimensions. This answers two open questions posed in <a class="link-https" data-arxiv-id="2205.15168v2" href="https://arxiv.org/abs/2205.15168v2">arXiv:2205.15168v2</a>. Finally, we compute dimensions of varieties of tensors of subrank at least $r$.

Algebraic Geometry,Computational Complexity
Discreteness of asymptotic tensor ranks

Jop Briët,Matthias Christandl,Itai Leigh,Amir Shpilka,Jeroen Zuiddam

2024-09-24

Abstract:Tensor parameters that are amortized or regularized over large tensor powers, often called "asymptotic" tensor parameters, play a central role in several areas including algebraic complexity theory (constructing fast matrix multiplication algorithms), quantum information (entanglement cost and distillable entanglement), and additive combinatorics (bounds on cap sets, sunflower-free sets, etc.). Examples are the asymptotic tensor rank, asymptotic slice rank and asymptotic subrank. Recent works (Costa-Dalai, Blatter-Draisma-Rupniewski, Christandl-Gesmundo-Zuiddam) have investigated notions of discreteness (no accumulation points) or "gaps" in the values of such tensor parameters. We prove a general discreteness theorem for asymptotic tensor parameters of order-three tensors and use this to prove that (1) over any finite field (and in fact any finite set of coefficients in any field), the asymptotic subrank and the asymptotic slice rank have no accumulation points, and (2) over the complex numbers, the asymptotic slice rank has no accumulation points. Central to our approach are two new general lower bounds on the asymptotic subrank of tensors, which measures how much a tensor can be diagonalized. The first lower bound says that the asymptotic subrank of any concise three-tensor is at least the cube-root of the smallest dimension. The second lower bound says that any concise three-tensor that is "narrow enough" (has one dimension much smaller than the other two) has maximal asymptotic subrank. Our proofs rely on new lower bounds on the maximum rank in matrix subspaces that are obtained by slicing a three-tensor in the three different directions. We prove that for any concise tensor, the product of any two such maximum ranks must be large, and as a consequence there are always two distinct directions with large max-rank.

Computational Complexity,Combinatorics,Quantum Physics

Maximal border subrank tensors

Border subrank via a generalised Hilbert-Mumford criterion

Border rank bounds for $GL(V)$-invariant tensors arising from matrices of constant rank

Concise tensors of minimal border rank

Symmetrization maps and minimal border rank Comon's conjecture

A Gap in the Subrank of Tensors

Classification and degenerations of small minimal border rank tensors via modules

Subrank and optimal reduction of scalar multiplications to generic tensors

On tensors of border rank $l$ in $\C^{m\times n\times l}$

A note on the gap between rank and border rank

Geometric rank of tensors and subrank of matrix multiplication

Affine subspaces of matrices with rank in a range

General tail bounds for random tensors summation: Majorization approach

Maximal invariant subspaces for a class of operators

Probabilistic bounds on best rank-1 approximation ratio

The Maximum Principle for Minimal Varieties of Arbitrary Codimension

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

Border apolarity and varieties of sums of powers

On the ranks and border ranks of symmetric tensors

Exact values of generic subrank

Discreteness of asymptotic tensor ranks