{"id":"arxiv/2608.30579","url":"https://arxiv.org/abs/2608.30579","tags":["math.PR"],"score":0,"title":"A Finite-Entropy Criterion for the Entropic Conditional Central Limit Theorem","domain":"math.PR","source":"arxiv","authors":["Tong Ye","Liu-Quan Yao","Shuai Yuan","Guanghui Wang"],"reviews":0,"subject":"math.PR","summary":"We prove a finite-entropy criterion for the entropic conditional central limit theorem. Let $(ξ_i,η_i)_{i\\geq 1}$ be independent copies of a pair $(ξ,η)$, and set $W_n=n^{-1/2}\\sum_{i=1}^n ξ_i$ and $\\boldsymbolη_n=(η_1,\\ldots,η_n)$. Under the assumptions that $\\mathbb{E}\\operatorname{Var}(ξ\\midη) -\\infty$ for some $n_0$. The main technical ingredient is a continuity theorem for Fisher information under Gaussian smoothing, which allows us to replace the finite expected conditional Fisher-information assumption by a necessary and sufficient finite-entropy condition.","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-08-31","reactions":{}}