Return to Article Details Improving Neural Network Efficiency Using Piecewise Linear Approximation of Activation Functions Download Download PDF