Id: arxiv/2601.19379 Url: https://arxiv.org/abs/2601.19379 Tags: math.OC, cs.LG Score: 0 Title: Optimal Asynchronous Stochastic Nonconvex Optimization under Heavy-Tailed Noise Domain: math.OC Source: arxiv Authors: Yidong Wu, Luo Luo Reviews: 0 Subject: math.OC Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-01-27 Page: https://delphi.science/r/arxiv/2601.19379 This paper considers the problem of asynchronous stochastic nonconvex optimization with heavy-tailed gradient noise and arbitrarily heterogeneous computation times across workers. We propose an asynchronous normalized stochastic gradient descent algorithm with momentum. The analysis show that our method achieves the optimal time complexity under the assumption of bounded $p$th-order central moment with $p\in(1,2]$. We also provide numerical experiments to show the effectiveness of proposed method.