Escaping Saddle Points in Constrained Optimization

Open in new window