Video Summary

The paper studies the problem of training neural networks with few data. The authors propose a method called data distillation, which trains a small student network by distilling the knowledge from a large teacher network. The method is evaluated on several image classification tasks, and it achieves state-of-the-art results on some of them.