Search arXiv⌕ Search

arXiv · 2506.16132

Bounds for Geometric rank in Terms of Subrank

Abstract

For tensors of fixed order, we establish three types of upper bounds for the geometric rank in terms of the subrank. Firstly, we prove that, under a mild condition on the characteristic of the base field, the geometric rank of a tensor is bounded by a function in its subrank in some field extension of bounded degree. Secondly, we show that, over any algebraically closed field, the geometric rank of a tensor is bounded by a function in its subrank. Lastly, we prove that, for any order three tensor over an arbitrary field, its geometric rank is bounded by a quadratic polynomial in its subrank. Our results have several immediate but interesting implications: (1) We answer an open question posed by Kopparty, Moshkovitz and Zuiddam concerning the relation between the subrank and the geometric rank; (2) For order three tensors, we generalize the Biaggi-Chang- Draisma-Rupniewski (resp. Derksen-Makam-Zuiddam) theorem on the growth rate of the border subrank (resp. subrank), in an optimal way; (3) For order three tensors, we generalize the Biaggi- Draisma-Eggleston theorem on the stability of the subrank, from the real field to an arbitrary field; (4) We confirm the open problem raised by Derksen, Makam and Zuiddam on the maximality of the gap between the subrank of the direct sum and the sum of subranks; (5) We derive, for the first time, a de-bordering result for the border subrank and upper bounds for the partition rank and analytic rank in terms of the subrank; (6) We reprove a gap result for the subrank.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Qiyuan Chen, Ke Ye. 2025-06-19. Bounds for Geometric rank in Terms of Subrank. https://arxiv.org/abs/2506.16132

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

On nut graphs with two vertex and three edge orbits

Nut graphs are graphs whose adjacency matrix is singular with one-dimensional null space spanned by a vector with no zero entries. In a recent paper, Bašić, Fowler and Pisanski proved that the automorphism group of a nut graph has more orbits on the edge set than on the vertex set. They classified all orders for which a vertex-transitive nut graph with precisely two edge orbits exists, and conjectured that a nut graph with two vertex and three edge orbits exists for each non-prime order $n \ge 9$. Motivated by this conjecture, we introduce a very general construction that provides graphs with the desired symmetry properties, and we determine some sufficient spectral and structural conditions under which they are nut graphs. The construction yields infinite families of examples and confirms the above conjecture for all odd non-prime orders up to $2\,500$ and for at least $99.8$ percent of all odd non-prime orders up to a million. Finally, we present some additional interesting examples of nut graphs with two vertex and three edge orbits that do not arise from this construction.

math.CO↗

On vertex-minimal simplicial maps to the sphere

For positive integers $n,d$, let $λ(n,d)$ be the minimal number of vertices of a triangulation of the $n$-sphere which admits a degree $d$ simplicial map onto the boundary of the $(n+1)$-simplex. We show that for $h=\lfloor\frac{n+1}2\rfloor$, the function $λ(n,d)^h$ has linear order of growth in $d$, answering a question of O. Musin. All triangulations we obtained are isomorphic to boundaries of convex polytopes in $\mathbb{R}^{n+1}$.

math.CO↗