{"id":"arxiv/2605.26260","url":"https://arxiv.org/abs/2605.26260","tags":["math.OC"],"score":0,"title":"Prox-NAG-GS: A Semi-Implicit Proximal Method for Composite Optimization","domain":"math.OC","source":"arxiv","authors":["Sikeh Gisele Wiykiynyuy","Kelvin Asu Ekuri","Valentin Leplat"],"reviews":0,"subject":"math.OC","summary":"Composite optimization problems, where a smooth loss is combined with a nonsmooth regularizer, are common in machine learning and inverse problems. In this work, we study a proximal extension of NAG-GS, a semi-implicit accelerated method obtained from a Gauss-Seidel discretization of an inertial dynamics. The proposed method, called Prox-NAG-GS, keeps the coupled structure of NAG-GS for the smooth part and replaces the second update by a proximal step. It therefore applies to objectives of the form $F=f+r$, where $f$ is smooth and $r$ is convex and proximable. We derive deterministic…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-05-25","reactions":{}}