The process of further training a pre-trained model on new data to adapt it for specific tasks or domains.