Backpropagation

Backpropagation works by first calculating the difference between a neural network’s predicted output and the correct output. This error is then propagated backward through the network, from the output layer toward the input layer. The algorithm calculates how much each weight contributed to the error, allowing an optimizer to update those weights and reduce future prediction errors.

Backpropagation makes it possible for neural networks to learn efficiently from training data. By repeatedly identifying errors and adjusting weights, models can gradually improve their predictions. It is a foundational technique behind the training of many modern deep learning systems.