The Process and Activation in a Neuron

A neuron receives multiple input values, and commonly sum up the input values with different weights (ratios for input values). An activation function is then applied to the sum (with bias usually). Formula (for programming, k is in range [0..n-1] instead):

Out_{ij} = f(\sum_{k=1}^{n}x_k \times w_k + b)

Where i is the index of a layer, j is the index of neuron in that layer, and k is the index of input value to that neuron. Value n is the number of inputs to the neuron, and b is bias value in that neuron. The last one, f is the activation function.


You'll only receive email when they publish something new.

More from 19411
All posts