{"id":"arxiv/2609.23994","url":"https://arxiv.org/abs/2609.23994","tags":["math.PR","cs.DS"],"score":0,"title":"High-Dimensional Ultra-Log-Concave Distributions","domain":"math.PR","source":"arxiv","authors":["Zongchen Chen","Sihan Wang"],"reviews":0,"subject":"math.PR","summary":"Ultra-log-concave distributions are ubiquitous in probability, combinatorics, and statistical mechanics and have been studied extensively. In this paper, we introduce a quantitative high-dimensional extension of this notion, called $δ$-ultra-log-concavity, for probability measures on $\\mathbb{N}^d$ with downward closed support. When $δ= 1$, this notion coincides with the class studied by Gurvits (2009) via strongly log-concave generating functions, and with the class defined by Anari, Oveis Gharan, and Vinzant (2021) via completely log-concave generating functions; in one dimension, it…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-09-21","reactions":{}}