Unconstrained Learning of Networked Nonlinear Systems via Free Parametrization of Stable Interconnected Operators

Free parametrization enables unconstrained optimization for large-scale system identification and control.