2 Preliminaries Weconsidertheproblemoftrainingofatwo-layerReLUneuralnetworkwith mscalarinputsanda singlescalaroutputusingtheleast-squaresloss: min

Open in new window