激活函数，并不是去激活什么，而是指如何把“激活的神经元的特征”通过研究

The activation function does not deactivate anything, but refers to how to retain and map the "features of activated neurons" through the research function (preserving cultural features and removing some redundancy in enterprise data), which is A neural network can solve the problem of nonlinear relationship and the key.

Activation functions, not to activate anything, but to how to preserve and map the characteristics of "activated neurons" through research functions (retaining cultural features, removing redundancy from some enterprise data), is a nervous system network that solves nonlinear relationship problems as well as the key.

Activation function is not to activate anything, but how to preserve and map the "characteristics of activated neurons" through research function (to preserve cultural characteristics and remove redundancy in some enterprise data). It is a neural system network that can solve the problem of non-linear relationship and the key.<br>