TY - JOUR
T1 - An Accelerated Stochastic Trust Region Method for Stochastic Optimization
AU - Qi , Rulei
AU - Xue , Dan
AU - Li , Jing
AU - Zhai , Yujia
JO - Journal of Computational Mathematics
VL - 5
SP - 1169
EP - 1193
PY - 2025
DA - 2025/09
SN - 43
DO - http://doi.org/10.4208/jcm.2504-m2023-0228
UR - https://global-sci.org/intro/article_detail/jcm/24476.html
KW - Stochastic optimization, Stochastic variance reduced gradient, Trust region, Gradient descent method, Machine learning.
AB - <p style="text-align: justify;">In this paper, we propose an accelerated stochastic variance reduction gradient method
with a trust-region-like framework, referred as the NMSVRG-TR method. Based on
NMSVRG, we incorporate a Katyusha-like acceleration step into the stochastic trust region
scheme, which improves the convergence rate of the SVRG methods. Under appropriate
assumptions, the linear convergence of the algorithm is provided for strongly convex objective functions. Numerical experiment results show that our algorithm is generally superior
to some existing stochastic gradient methods.</p>