Skip to main content

🏢 School of Mathematics and Statistics, Xidian University

Globally Q-linear Gauss-Newton Method for Overparameterized Non-convex Matrix Sensing
·1454 words·7 mins· loading · loading
Machine Learning Optimization 🏢 School of Mathematics and Statistics, Xidian University
A globally Q-linearly converging Gauss-Newton method (AGN) is introduced for overparameterized non-convex low-rank matrix sensing, significantly improving convergence compared to existing gradient des…