1: \begin{abstract}

2: We derive a nonlinear integro-differential transport equation describing  collective evolution of weights under gradient descent in large-width neural-network-like models. 

3: We characterize stationary points of the evolution and analyze several scenarios where the transport equation can be solved approximately. We test our general method  in the special case of linear free-knot splines, and find good agreement between theory and experiment in observations of global optima, stability of stationary points, and convergence rates. 

4: \end{abstract}

5: