A backpropagation neural network is a way to train neural networks. It involves providing a neural network with a set of input values for which the correct output value is known beforehand. The network processes the input and produces an output value, which is compared to the correct value. If the result is incorrect, gradual changes are propagated backward through the network, and the calculation is made again. This process repeats until the desired output is calculated by the network.

Neural network, Programming terms

  • How to create a computer program.