Frontend
On the Kissing Number of Craig's Lattice and Spherical List Decoding
arXiv math.IT0 views
arXiv:2611.22441v1 Announce Type: new Abstract: Craig's lattices $A_{p-1}^k$ are a family of lattices with large kissing numbers, i.e., relative to their dimension $n = p-1$, they contain a large number of shortest nonzero vectors. Specifically, for small $k = o(\log p / \log \log p)$, a lower bound of Bacher (2015) shows that $A_{p-1}^k$ has a super-polynomial kissing number of $\omega(n^k)$. Craig's lattices are therefore attractive candidates for the construction of spherical codes. In this work, we introduce a polynomial-time list decoder for spherical codes obtained from Craig's lattices for angular distance below $45^\circ$. Our second main result shows that Bacher's lower bound is asymptotically tight for $k = o(\log p / \log \log p)$. Finally, we introduce an efficient algorithm for computing the kissing number of $A_{p-1}^k$. Using this algorithm, we compute their kissing number in dimensions up to 521. Based on this data, we pose the open question whether the kissing number of $A_{p-1}^k$ with $k \approx \sqrt{p}$ grows like $e^{\Theta(\sqrt{n})}$ in dimension $n=p-1$. Concurrent work by Laarhoven and Kominers as well as Bennett, Hhan, Li and Stephens-Davidowitz proves such a bound for very closely related lattices.
Source:https://arxiv.org/abs/2611.22441PreviousNoise Characterization and Robust Signal Detection in Yeast Pheromone Molecular CommunicationNextConstructions of reversible linear and DNA codes through $T_{D,g}^n-$set and $(n_1,c_1|n_2,c_2)-$right polycyclic codes
Related posts
Our import worker was killed for memory and the last log line we had was thirty thousand rows old
0dev.to Docker
Chinese firm unveils first integrated analog chip for one-step matrix solving
0Hacker News New
TLAPS-Bench: Formally Proving Correctness Of Real-World System Specs With TLA+
0Hacker News New
Show HN: csshm – cluster SSH written in Swift for macOS
0Hacker News New
Alaska Senate race called: Democrats win
0Hacker News New
Don't ask the model to be brief. Cut it in code, and count what you cut.
0dev.to