Reinforcement Learning with TensorFlow
上QQ阅读APP看书,第一时间看更新

Activation functions for deep learning

Activation functions are the integral units of artificial neural networks. They decide whether a particular neuron is activated or not, that is, whether the information received by the neuron is relevant or not. The activation function performs nonlinear transformation on the receiving signal (data).

We will discuss some of the popular activation functions in the following sections.