Differential Equation Scaling Limits of Shaped and Unshaped Neural Networks
Shaped neural networks with activation functions scaled as the network size grows, and unshaped neural networks with unchanged activation functions, both have differential equation-based asymptotic characterizations.