1. A machine learning algorithm used to train artificial neural networks by calculating gradients of the loss function with respect to the network weights, enabling efficient optimization through backward error propagation
The backpropagation algorithm is fundamental to deep learning and allows neural networks to learn from data by adjusting weights based on error measurements.
O algoritmo de retropropagação é fundamental para o aprendizado profundo e permite que redes neurais aprendam com dados ajustando pesos baseado em medições de erro.
2. A method for computing derivatives in neural networks by propagating errors backward from the output layer to the input layer
Using the backpropagation algorithm, we can efficiently train large neural networks in reasonable time.
Usando o algoritmo de retropropagação, podemos treinar redes neurais grandes de forma eficiente em tempo razoável.