Reg-Nyström Optimization for Machine Learning
Hessian Sketching for Scalable ML Optimization is a second-order optimization framework that combines Nyström-based Hessian approximation with adaptive regularization to enable fast, stable training of large-scale machine learning models
