AI RESEARCH

Truncated Kernel Stochastic Gradient Descent with General Losses and Spherical Radial Basis Functions

arXiv CS.LG

ArXi:2510.04237v5 Announce Type: replace In this paper, we propose a novel kernel stochastic gradient descent (SGD) algorithm for large-scale supervised learning with general losses. Compared to traditional kernel SGD, our algorithm improves efficiency and scalability through an innovative regularization strategy.