Backpropagation
Backpropagation is the algorithm that calculates how much each weight in a neural network contributed to the error, so the weights can be corrected during training.
What is Backpropagation?
It applies the chain rule of calculus backwards from the output layer to the input layer. Popularised in the 1980s, it made training multi-layer networks practical and remains the foundation of deep learning.
How it works
A forward pass produces a prediction and a loss; a backward pass propagates error gradients layer by layer; gradient descent then updates the weights.
← All Artificial Intelligence & Superintelligence definitions · ← Back to the A–Z definitions