A parameter that directs the machine learning process. Selected by the AI engineer, hyperparameters are chosen before any training is done. Examples are architecture model, number of branches in a tree and number of training passes (epochs, iterations). See
AI,
neural network,
machine learning and
AI weights and biases.