A training strategy that gradually teaches models from simple tasks to complex ones, mimicking human learning progression.