A shortcut that adds a layer's input back to its output, so the layer only has to learn a correction, not the whole answer.
Continue to AI University →