no code implementations • 12 Jan 2020 • Kunping Huang, Paul Siegel, Anxiao, Jiang
That is, by seeing the NeuralNet as a function of its input, the error correction scheme is function-oriented.
