Sparse Networks from Scratch: Faster Training without Losing Performance
This addresses the computational bottleneck in training large neural networks for machine learning practitioners by enabling faster training without performance loss, though it is incremental as it builds on existing sparse training methods.
The paper tackles the problem of accelerating deep neural network training while maintaining performance by introducing sparse learning, which uses sparse momentum to keep weights sparse throughout training and achieves up to 5.61x faster training with relative error reductions of 8%, 15%, and 6% on MNIST, CIFAR-10, and ImageNet compared to other sparse algorithms.
We demonstrate the possibility of what we call sparse learning: accelerated training of deep neural networks that maintain sparse weights throughout training while achieving dense performance levels. We accomplish this by developing sparse momentum, an algorithm which uses exponentially smoothed gradients (momentum) to identify layers and weights which reduce the error efficiently. Sparse momentum redistributes pruned weights across layers according to the mean momentum magnitude of each layer. Within a layer, sparse momentum grows weights according to the momentum magnitude of zero-valued weights. We demonstrate state-of-the-art sparse performance on MNIST, CIFAR-10, and ImageNet, decreasing the mean error by a relative 8%, 15%, and 6% compared to other sparse algorithms. Furthermore, we show that sparse momentum reliably reproduces dense performance levels while providing up to 5.61x faster training. In our analysis, ablations show that the benefits of momentum redistribution and growth increase with the depth and size of the network. Additionally, we find that sparse momentum is insensitive to the choice of its hyperparameters suggesting that sparse momentum is robust and easy to use.