Id: arxiv/2611.17727 Url: https://arxiv.org/abs/2611.17727 Tags: math.OC, math.NA Score: 0 Title: Proximal Adam algorithm for nonsmooth composite optimization and linear convergence under proximal PL geometry Domain: math.OC Source: arxiv Authors: Xiaopeng Zhu, Zhongyi Huang Reviews: 0 Subject: math.OC Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-11-13 Page: https://delphi.science/r/arxiv/2611.17727 We analyze Adam-type methods for nonsmooth composite optimization problems $\min_xΦ(x)=f(x)+h(x)$, where $f$ is smooth and $h$ is convex, proper, and lower semicontinuous. The nonsmooth term is treated by a metric-weighted proximal map, while the smooth term is preconditioned by the Adam second-moment matrix. We prove a composite descent lemma whose constants depend explicitly on the spectral bounds of the adaptive metric and a residual--step equivalence that converts proximal step decay into subgradient stationarity. These yield an $O(1/N)$ bound on the minimum proximal stationarity…