{"id":"arxiv/2601.17375","url":"https://arxiv.org/abs/2601.17375","tags":["math.NA"],"score":0,"title":"Operator splitting based diffusion samplers and improved convergence analysis","domain":"math.NA","source":"arxiv","authors":["Peiyi Liu","Zhaoqiang Liu","Yiqi Gu"],"reviews":0,"subject":"math.NA","summary":"In this paper, we develop a class of samplers for the diffusion model using the operator-splitting technique. The linear drift term and the nonlinear score-driven drift of the probability flow ordinary differential equation are split and applied by flow maps alternatively. Moreover, we conduct detailed analyses for the second-order sampler, establishing a non-asymptotic total variation distance error bound of order $O(d/T^2+\\sqrt{d}\\varepsilon_{\\mathrm{score}}+d\\varepsilon_{\\mathrm{Jac}})$, where $d$ is the data dimension; $T$ is the number of sampling steps; $\\varepsilon_{\\mathrm{score}}$…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-01-24","reactions":{}}