{"id":"arxiv/2609.30105","url":"https://arxiv.org/abs/2609.30105","tags":["cs.LG","cs.CC","math.PR"],"score":0,"title":"On the SoS Certifiability of Log-Concave Distributions","domain":"cs.LG","source":"arxiv","authors":["Aleksandr Storozhenko"],"reviews":0,"subject":"cs.LG","summary":"We prove that for every isotropic log-concave distribution $P$ on $\\mathbb{R}^d$ and every even $m\\ge2$, the polynomial $(Cm)^m\\|v\\|_2^m - \\mathbb{E}_{X\\sim P}\\langle X,v\\rangle^m$ is a sum of squares, where $C>0$ is a universal constant. This improves on the Poincaré-dependent bounds (Kothari and Steinhardt, 2017), recovering the optimal moment bounds for log-concave distributions. As an immediate corollary, we obtain computationally efficient algorithms with dimension-free error guarantees for a wide range of statistical estimation problems. Our proof proceeds by using stochastic…","category":"research","comments":0,"disputed":false,"division":"cs","mentions":0,"discussed":0,"published":"2026-09-24","reactions":{}}