Gradient descent
Optimisation method that repeatedly moves a model's parameters a small step in the direction that lowers the error most; the step size is set by the learning rate.
Gradient descent is an optimisation method that repeatedly moves a model’s parameters a small step in the direction where the error decreases most steeply. The learning rate controls how large each step is. The error surface of a deep neural network can contain local minima, so the method does not always find a global optimum.
Backpropagation computes the gradients with which gradient descent adjusts all weights during training of modern neural networks. You can try a hands-on version in Geek of the Week: Gradient descent.