Optimal approximation of piecewise smooth functions using deep ReLU neural networks
Keywords: Deep neural networks; Piecewise smooth functions; Function approximation; Sparse connectivity; Metric entropy; Curse of dimension;