Fetching the paper…

Gradient Descent on Infinitely Wide Neural Networks: Global Convergence and Generalization · Around