Distilling the Knowledge in a Neural Network
2015·Arxiv