Not All Samples Are Created Equal: Deep Learning with Importance Sampling
[edit]
Proceedings of the 35th International Conference on Machine Learning, PMLR 80:25252534, 2018.
Abstract
Deep Neural Network training spends most of the computation on examples that are properly handled, and could be ignored. We propose to mitigate this phenomenon with a principled importance sampling scheme that focuses computation on "informative" examples, and reduces the variance of the stochastic gradients during training. Our contribution is twofold: first, we derive a tractable upper bound to the persample gradient norm, and second we derive an estimator of the variance reduction achieved with importance sampling, which enables us to switch it on when it will result in an actual speedup. The resulting scheme can be used by changing a few lines of code in a standard SGD procedure, and we demonstrate experimentally on image classification, CNN finetuning, and RNN training, that for a fixed wallclock time budget, it provides a reduction of the train losses of up to an order of magnitude and a relative improvement of test errors between 5% and 17%.
Related Material


