Improving the Expressive Power of Deep Neural Networks through Integral Activation Transform

Open in new window