Principal Component Networks: Parameter Reduction Early in Training

In this paper, we show that hidden layer activations in overparameterized neural networks for image classification exist primarily in subspaces smaller than the actual model width. We further show that these subspaces can be identified early in training. Based on these observations, we show how to efficiently find small networks that exhibit similar accuracy to their overparameterized counterparts after only a few training epochs. We term these network architectures Principal Component Networks (PCNs). We evaluate PCNs on CIFAR-10 and ImageNet for VGG and ResNet style architectures and find that PCNs consistently reduce parameter counts with little accuracy loss, thus providing the potential to reduce the computational costs of deep neural network training.