Gradient Descent with Random Initialization: Fast Global Convergence for Nonconvex Phase Retrieval

Open in new window