Entropy approximations of algebraic matroids over finite fields

September 18, 2025 ยท The Ethereal ยท ๐Ÿ› arXiv.org

๐Ÿ”ฎ THE ETHEREAL: The Ethereal
Pure theory โ€” exists on a plane beyond code

"No code URL or promise found in abstract"

Evidence collected by the PWNC Scanner

Authors Guillermo Matera arXiv ID 2509.15348 Category math.CO: Combinatorics Cross-listed cs.IT, math.AG, math.NT Citations 0 Venue arXiv.org Last Checked 3 months ago
Abstract
We investigate the asymptotic behavior of entropy polymatroids associated with algebraic matroids over finite fields. Given an algebraic matroid ${\sf M}:=(\mathcal{E},r)$ and the irreducible variety $V$ associated with ${\sf M}$, we consider the polymatroid $h_{\mathbb{F}}$ induced by the entropies of the projections of $V(\mathbb{F})$, where $\mathbb{F}$ is a finite extension of $\mathbb{G}$. Revisiting a construction of Matรบ$\check{\mathrm{s}}$, we show that the polymatroid $h_{\mathbb{F}}$ converges to the rank function $r$ of ${\sf M}$ as $q:=|\mathbb{F}|$ tends to infinity. Our main contribution is to make this convergence quantitative: we derive explicit uniform error bounds for the deviation $|h_\mathbb{F}-r|$, expressed in terms of the degree of $V$, the ground set size $|\mathcal{E}|$, the rank $r(\mathcal{E})$, and $q$. The proofs combine tools of algebraic geometry (effective Lang-Weil estimates and intrinsic degree bounds for annihilating polynomials of circuits) with information-theoretic arguments (submodularity of entropy and conditional entropy estimates). These results provide the first effective and uniform approximation bounds for algebraic matroids by entropy polymatroids, clarifying the quantitative link between algebraic independence (captured by matroid rank) and information-theoretic independence (captured by entropy).
Community shame:
Not yet rated
Community Contributions

Found the code? Know the venue? Think something is wrong? Let us know!

๐Ÿ“œ Similar Papers

In the same crypt โ€” Combinatorics

๐Ÿ”ฎ ๐Ÿ”ฎ The Ethereal

Tables of subspace codes

Daniel Heinlein, Michael Kiermaier, ... (+2 more)

math.CO ๐Ÿ› arXiv ๐Ÿ“š 94 cites 10 years ago