E2CM: Early Exit via Class Means for Efficient Supervised and Unsupervised Learning
Abstract
State-of-the-art neural networks with early exit mechanisms often need considerable amount of training and fine tuning to achieve good performance with low computational cost. We propose a novel early exit technique, Early Exit Class Means (E2CM), based on class means of samples. Unlike most existing schemes, E2CM does not require gradient-based training of internal classifiers and it does not modify the base network by any means. This makes it particularly useful for neural network training in low-power devices, as in wireless edge networks. We evaluate the performance and overheads of E2CM over various base neural networks such as MobileNetV3, EfficientNet, ResNet, and datasets such as CIFAR-100, ImageNet, and KMNIST. Our results show that, given a fixed training time budget, E2CM achieves higher accuracy as compared to existing early exit mechanisms. Moreover, if there are no limitations on the training time budget, E2CM can be combined with an existing early exit scheme to boost the latter's performance, achieving a better trade-off between computational cost and network accuracy. We also show that E2CM can be used to decrease the computational cost in unsupervised learning tasks.