The process of automatically tuning configuration parameters of a machine learning model to improve its performance.