Neural Network Perceptron

Computer Science

How It Works

An artificial perceptron is the fundamental computational building block of artificial neural networks, executing linear binary classification. Multiple numerical inputs are modulated by learnable synaptic connection weights, aggregated with a scalar bias offset, and projected through a nonlinear activation function like the Sigmoid or ReLU. By adjusting weights via gradient-based backpropagation, the perceptron synthesizes hyperplanes separating multidimensional data classes.

Governing Equation
y = σ(∑_{i=1}^n w_i x_i + b) , σ(z) = 1 / (1 + e^(-z))