Id: arxiv/2602.05460 Url: https://arxiv.org/abs/2602.05460 Tags: math.ST Score: 0 Title: Complexity reduction in online stochastic Newton methods with potential O(N d) total cost Domain: math.ST Source: arxiv Authors: Antoine Godichon-Baggioni, Bruno Portier, Guillaume Sallé Reviews: 0 Subject: math.ST Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-02-05 Page: https://delphi.science/r/arxiv/2602.05460 Optimizing smooth convex functions in stochastic settings, where only noisy estimates of gradients and Hessians are available, is a fundamental problem in optimization. While first-order methods possess a low per-iteration cost, their convergence is slow for ill-conditioned problems. Stochastic Newton methods utilize second-order information to correct for local curvature, but the O(d 3 ) per-iteration cost of computing and inverting a full Hessian, where d is the problem dimension, is prohibitive in high dimensions. This paper introduces an online mini-batch stochastic Newton algorithm. The…