MP neuron, Perceptron and Sigmoid Neuron

This article helps us differentiate between MP neuron, Perceptron and Sigmoid Neuron.

McCulloh-Pitts(MP) Neuron:

1. Input and outputs are boolean(i.e zero or one).

2. No weights

3. Classify only linearly separable boolean functions.

Drawbacks : Cannot deal with real valued inputs and weightage cannot be given to inputs.

MP Neuron

Perceptron:

1. Inputs are real values.

2. Outputs are Boolean values.

3. Deal with only linearly separable functions.

4. Weights can be assigned to each input.

Drawbacks : Thresholding is very harsh in perceptron.

Output function where b is the threshold (blue : perceptron and red : sigmoid neuron)

Sigmoid Neuron:

1. Inputs and outputs are real.

2. Weights can be assigned to inputs.

3. Thresholding is not harsh.

4. Continuous and differentiable function..