Secant varieties of Segre-Veronese varieties ℙ m × ℙ n \mathbb{P}^{m}\times\mathbb{P}^{n} embedded by 𝒪 ⁡ ( 1 , 2 ) \mathcal{O}(1,2) are non-defective for n ≫ m 3 n\gg m^{3} , m ≥ 3 m\geq 3

Matěj Doležálek, Nikhil Ken · 2025

Read Secant varieties of Segre-Veronese varieties ℙ m × ℙ n \mathbb{P}^{m}\times\mathbb{P}^{n} embedded by 𝒪 ⁡ ( 1 , 2 ) \mathcal{O}(1,2) are non-defective for n ≫ m 3 n\gg m^{3} , m ≥ 3 m\geq 3 by Matěj Doležálek, Nikhil Ken (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-21972.txt.

Introduction

Matěj Doležálek and Nikhil Ken present a 2025 mathematics paper on the dimensions of secant varieties. A Segre-Veronese variety packages pairs of projective spaces with a chosen embedding. Its secant varieties join points of the original in straight families, and they can fall short of their expected dimension, a defect.

The authors prove that for the embedding given by the bundle O(1,2), with the first factor at least three-dimensional and the second vastly larger, no secant variety is defective. Earlier work covered the smaller cases; this paper handles the superabundant range. Their tool, called the inductant, reduces the claim to finitely many base cases. A computer-assisted check then verifies those cases.

Catalog record

CREATORMatěj Doležálek, Nikhil Ken
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE45 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-21972

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Secant varieties of Segre-Veronese varieties ℙ m × ℙ n \mathbb{P}^{m}\times\mathbb{P}^{n} embedded by 𝒪 ⁡ ( 1 , 2 ) \mathcal{O}(1,2) are non-defective for n ≫ m 3 n\gg m^{3} , m ≥ 3 m\geq 3 Thanks: This work has been supported by European Union’s HORIZON–MSCA-2023-DN-JD programme under the Horizon Europe (HORIZON) Marie Skłodowska-Curie Actions, grant agreement 101120296 (TENORS) Matěj Doležálek, Nikhil Ken Abstract . We prove that for any m ≥ 3 m\geq 3 , n ≫ m 3 n\gg m^{3} , all secant varieties of the Segre-Veronese variety ℙ m × ℙ n \mathbb{P}^{m}\times\mathbb{P}^{n} have the expected dimension. This was already proved by Abo and Brambilla in the subabundant case, hence we focus on the superabundant case. We generalize an approach due to Brambilla and Ottaviani into a construction we call the inductant . With a combinatorial investigation of these constructions, the proof of non-defectivity reduces to checking a finite collection of base cases, which we verify using a computer-assisted proof. 2020 Mathematics Subject Classification Primary 14N07; Secondary 14N20, 14Q15, 15A69 1. Introduction The study of secant varieties is a classical problem of algebraic geometry. When restricted to varieties such as the Segre or Veronese varieties, it also presents a geometric formulation of questions concerning border rank or symmetric border rank in tensor spaces. The famous theorem of Alexander and Hirschowitz [ alexander-hirschowitz95 , alexander-hirschowitz97 ] solved completely the problem of determining the dimensions of secant varieties of Veronese varieties. It says that this dimension is always the so-called expected dimension , except for an explicitly enumerated and relatively small set of defective cases. Brambilla and Ottaviani [ brambilla-ottaviani ] later gave an alternative presentation, some of whose techniques motivate this article. For Segre varieties or for Segre-Veronese varieties, complete solutions in the spirit of the Alexander-Hirschowitz theorem are currently not known, although many partial results are available. Strassen [ strassen69 , strassen83 ] originally studied the problem of tensor rank (and border rank) in connection with complexity of matrix multiplication. Lickteig [ lickteig ] provided further results on three-factor Segre varieties, notably in the case of \mathbb{P}^{n}\times\mathbb{P}^{n}\times\mathbb{P}^{n} . Abo, Ottaviani and Peterson [ abo-ottaviani-peterson ] studied (\mathbb{P}^{n})^{k} in greater generality using a general inductive technique; they also classified defective s -th secant varieties of Segre varieties for small s . Blomenhofer and Casarotti [ blomenhofer-casarotti ] gave a unified treatment for Segre-Veronese varieties, Grassmannians and other varieties and proved non-defectivity of their s -th secant varieties for s outside of a certain small interval. In the case of Segre-Veronese varieties, the problem of defective secant varieties has also received much attention in recent years, especially in the two-factor case. Bernardi, Carlini and Catalisano [ bernardi-carlini-catalisano ] studied s -th secant varieties of \mathbb{P}^{n}\times\mathbb{P}^{n} embedded by \mathcal{O}(1,d) , d\geq 3 and proved non-defectivity for s outside of an interval bounded by multiples of n+1 nearest to \frac{(n+1)\binom{m+d}{d}}{m+n+1} . In [ ballico-bernardi-catalisano ] , Ballico, Bernardi and Catalisano completely characterized dimensions of secant varieties to \mathbb{P}^{n}\times\mathbb{P}^{1} embedded by any \mathcal{O}(a,b) . In general, the trend with Segre-Veronese varieties seems to be that factors of high degree make the problem easier. Most notably in this vein, Abo, Brambilla, Galuppi and Oneto [ abo-brambilla-galuppi-oneto ] proved recently that Segre-Veronese varieties \mathbb{P}^{n_{1}}\times\cdots\times\mathbb{P}^{n_{k}} embedded by \mathcal{O}(d_{1},\dots,d_{k}) with k\geq 3 and all d_{i}\geq 3 never have defective secant varieties. This result built on earlier works by Galuppi and Oneto [ galuppi-oneto ] and by Ballico [ ballico ] . Ballico [ ballico2024 ] proved further improvements on the results of Abo, Brambilla, Galuppi and Oneto which allow for some of the factors to be of degree d_{i}=2 . All of these results rely on first proving non-defectivity in some concrete multidegrees – such as \mathcal{O}(3,3) , \mathcal{O}(3,4) and \mathcal{O}(4,4) in [ galuppi-oneto ] – and subsequently performing inductions on degrees and on the number of factors. These inductive techniques leave the cases with small degrees as seemingly the most interesting. The most notable of these are \mathcal{O}(1,1,1) and \mathcal{O}(1,2) , the latter of which is the focus of this article. In this domain, Cartwright, Erman and Oeding [ cartwright-erman-oeding ] provided equations defining s -th secant varieties of \mathbb{P}^{2}\times\mathbb{P}^{n} embedded by \mathcal{O}(1,2) for small s . The case of \mathbb{P}^{2}\times\mathbb{P}^{n} is also notable because it contains an infinite family of examples of defective secant varieties [ abo-brambilla , Theorem 5.3] . Abo and Brambilla worked with a general \mathbb{P}^{m}\times\mathbb{P}^{n} , proving the following: Theorem ( [ abo-brambilla , Theorem 3.13] ) . Let m , n be positive integers with n\geq m-2 . Then for any s\leq\underline{s}(m,n):=\left\lfloor\frac{(m+1)(n-m+2)}{2}\right\rfloor , the Segre-Veronese variety \mathbb{P}^{m}\times\mathbb{P}^{n} embedded by \mathcal{O}(1,2) has a non-defective s -th secant variety. For s<\frac{(m+1)\binom{n+2}{2}}{n+m+1} , the secant variety cannot fill the ambient space \mathbb{P}^{(m+1)\binom{n+2}{2}-1} and hence the cases s\leq\left\lfloor\frac{(m+1)\binom{n+2}{2}}{n+m+1}\right\rfloor are called subabundant . Since the bound \left\lfloor\frac{(m+1)(n-m+2)}{2}\right\rfloor in the result of Abo and Brambilla equals \left\lfloor\frac{(m+1)\binom{n+2}{2}}{n+m+1}\right\rfloor for n large enough with respect to m , this settles all subabundant cases for sufficiently large n . Abo [ abo ] also investigated \mathbb{P}^{n}\times\mathbb{P}^{n} and \mathbb{P}^{n+1}\times\mathbb{P}^{n} with \mathcal{O}(1,2) , proving that these are non-defective except for \mathbb{P}^{4}\times\mathbb{P}^{3} . Our main result in this paper is the following: Theorem 1.1 . If m\geq 3 and n\geq\begin{cases}m^{3}-m-3,&\text{if $m$ is even and $n$ is odd},\\ \left\lceil\frac{m^{3}-2m-5}{2}\right\rceil,&\text{otherwise},\end{cases} then all secant varieties of the Segre-Veronese variety \mathbb{P}^{m}\times\mathbb{P}^{n} embedded by \mathcal{O}(1,2) are non-defective. We will prove this through Theorems and . Our focus will be on the so-called superabundant case, when s\geq\left\lceil\frac{(m+1)\binom{n+2}{2}}{n+m+1}\right\rceil , in order to complement the subabundant result due to Abo and Brambilla. In particular, we will prove that under the bounds of Theorem 1.1 , the s -th secant variety for s=\overline{s}(m,n):=\underline{s}(m,n)+1 is superabundant and non-defective. This will then immediately imply the same also for s\geq\overline{s}(m,n) , hence obtaining Theorem 1.1 when combined with the result of Abo and Brambilla. Previously, there remained even for large n an interval of values of s for which the non-defectivity of the secant variety was not known. We are able to close this gap (for large n ) by systematically handling the combinatorics of configurations that arise from specializing points to subspaces. Note that cases m=1 and m=2 have already been completely solved (see [ abo-brambilla , Corollary 1.4] ), and leaving them out allows us to sidestep the defective cases with m=2 and n odd. Let us also note that our result is consistent with the list of defective cases which was conjectured to be exhaustive in [ abo-brambilla , Section 5] . To close out this introduction, we outline the structure of the paper. In Section 2 we overview the basic setting of secant varieties and prepare some arithmetic related to secant varieties of Segre-Veronese varieties embedded by \mathcal{O}(1,2) . In Section 3 , we introduce the notions of coordinate configurations and the inductant construction. In Section , we apply these to prove Theorem 1.1 . Finally, in Section , we describe the computations which were needed to supply the base cases of inductive proofs in Section ; the source code and resulting certificates for these computations are available on GitHub at [ gitrepo ] . 2. Preliminaries 2.1. Segre-Veronese varieties and their secant varieties Throughout this paper, let U

Read the complete text