What is Backpropagation?
Backpropagation is a key algorithm used to train neural networks by adjusting their weights through gradient descent. It works by calculating how much each weight in the network contributes to the error (or loss), which helps the model learn from its mistakes and improve over time. In simple terms, the network uses this information (the gradients) to update the weights, reducing the error for future predictions and making the model more accurate with each adjustment.