{"id":"arxiv/2601.08184","url":"https://arxiv.org/abs/2601.08184","tags":["math.PR","cs.LG","stat.ML"],"score":0,"title":"Wasserstein-p Central Limit Theorem Rates: From Local Dependence to Markov Chains","domain":"math.PR","source":"arxiv","authors":["Yixuan Zhang","Qiaomin Xie"],"reviews":0,"subject":"math.PR","summary":"Non-asymptotic central limit theorem (CLT) rates play a central role in modern machine learning and operations research. In this paper, we study CLT rates for multivariate dependent data in Wasserstein-$p$ ($W_p$) distance, for general $p\\ge 1$. We focus on two fundamental dependence structures that commonly arise in practice: locally dependent sequences and geometrically ergodic Markov chains. In both settings, we establish the first optimal $\\mathcal O(n^{-1/2})$ rate in $W_1$, as well as the first $W_p$ ($p\\ge 2$) CLT rates under mild moment assumptions, substantially improving the best…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-01-13","reactions":{}}