A value that directs the machine learning process. Selected by the neural network designer, hyperparameters are chosen before any training is done. Examples of a hyperparameter are the number of hidden layers in the network, the number of neurons per layer and epochs (the number of passes through the dataset). Hyperparameters are adjusted throughout the training phase. See
AI training vs. inference.
Parameters
Both hyperparameters and parameters (weights and biases) are set at the beginning; however, AI engineers may change the hyperparameters, but the weights and biases are constantly updated by the software during the training stages. See
neural network and
AI weights and biases.