Curriculum learning is an optimization strategy in neural networks which consists of feeding the iterative learning algorithm with progressively harder examples. The idea is inspired by the way curriculums are built for humans to learn: from easier/simpler to harder/complex topics.