Perceptron or Artificial Neuron

Let’s delve into the perceptron, which serves as a fundamental building block for more complex neural networks. Here’s a concise explanation along with a simple figure:

What Is a Perceptron?

• An artificial neuron (often referred to as a perceptron) is inspired by the biological neurons found in the human nervous system.
• It serves as the basic computational unit within neural networks as Building Block of Neural Networks.
• A perceptron is one of the simplest forms of artificial neural networks.
• It was introduced by Frank Rosenblatt in the late 1950s.
• The artificial neuron takes input signals, processes them, and produces an output based on a specific activation function.
• It plays a crucial role in tasks such as classification, regression, and pattern recognition.

Architecture:

The perceptron consists of:

• Input Features: These represent characteristics or attributes of the input data (x1, x2 ....)
• Weights: Each input feature has an associated weight, influencing the perceptron’s output (w1, w2 ....)
• Bias (b): Bias represent a constant offset or shift in neuron activations. An additional parameter learned during training. Without bias, neural networks might struggle to learn certain decision boundaries.
• Summation Function: It computes the weighted sum of inputs Weighted Sum (w1*x1 + w2*x2 + ... + wn*xn) + b
• Activation Function: Typically, the Heaviside step function is used to determine the output (0 or 1).

Here’s a simplified representation:

Input Features (x1, x2, ..., xn)

  ↓

Weighted Sum (w1*x1 + w2*x2 + ... + wn*xn)+ b

  ↓

Activation Function 

  ↓

Output (0 or 1)

Perceptron Algorithm

The Perceptron is inspired by the information processing of a single neural cell called a neuron.

A neuron accepts input signals via its dendrites, which pass the electrical signal down to the cell body.

In a similar way, the Perceptron receives input signals from examples of training data that we weight and combined in a linear equation called the activation.

activation = sum(weight_i * x_i) + bias

The activation is then transformed into an output value or prediction using a transfer function, such as the step transfer function.

prediction = 1.0 if activation >= 0.0 else 0.0

In this way, the Perceptron is a classification algorithm for problems with two classes (0 and 1) where a linear equation (like or hyperplane) can be used to separate the two classes.

It is closely related to linear regression and logistic regression that make predictions in a similar way (e.g. a weighted sum of inputs).

The weights of the Perceptron algorithm must be estimated from your training data using stochastic gradient descent.