NEURON IN NEURAL NETWORK(artificial neuron)

Artificial Neural Networks (ANNs) are one of the main tools that take advantage of the concept of deep learning. They are an abstract representation of our nervous system, which contains a collection of neurons that communicate with each other through connections called axons

Similar to the biological one, the artificial neuron consists of the following:

  • > it has multiple inputs connection and each inputs connectionwith some weights
  • >it has multiple output connection (which helps to make further calculation
  • >it has activation function which helps for making calculation and it make calculation on the basis of input connection signals and gives ouput for further calculation

summation function : it helps for summing up all the input with respect to their weights make input for activation function for making decision or calculation on the basis of types of activation function

X =( A1W1+A2W2 + A3W3+…………. AnWn)

now this X is input for the activation function for making calculation or decision

WHY WE NEED ACTIVATION FUNCTION ?

neurons helps to perform task by machine by making decision and these decision making queries are solved by activation function (activation mainly used for decision making)

their are various type of activation function :( according to their different type of decision

  • Step function
  • Linear combination
  • Sigmoid function
  • Hyperbolic tangent function
  • Softmax function
  • Radial basis function
  • Conic section function
  • Relu activation function
  • Leaky ReLU
  • Tanh