Perceptron §

---

Simplified model of a biological neuron, introduced by Frank Rosenblatt in the late 1950s.

Inspiration from Biological Neurons §

---

The model was inspired by the way biological neurons work. They receive electrical signals (inputs) from dendrites, process these signals and produce an output signal through an axon.

Inputs and Weights §

---

It takes multiple binary inputs (0 or 1), where each input is associated with a weight (which represents the importance or strength of that input).

Processing §

---

The Perceptron calculates the weighted sum of its inputs, which is often called the activation or net input. It then applies an activation function to this weighted sum to produce an output.

A simple step function can be used as the activation function, where the output is 1 if the weighted sum is above a certain threshold and 0 otherwise.

Output §

---

It is a binary value, typically 0 or 1. It represents the decision or classification made based on its inputs and weights. In some cases, the output might be modified to be -1 and 1 instead of 0 and 1.

Learning §

---

One of the key features of the Perceptron is its ability to learn from data. It can adjust its weights based on the input data and a learning rule, allowing it to adapt and make decisions over time.

Applications §

---

They were initially used for simple binary classification tasks, one example being to distinguish between two classes of data points by finding an optimal decision boundary.

They serve as a basis for more advanced neural network architectures, like multi-layer perceptrons (MLPs) and deep neural networks (DNNs).