Biased Importance Sampling for Deep Neural Network Training
This addresses the problem of slow training in deep neural networks for researchers and practitioners, though it is incremental as it builds on existing importance sampling techniques.
The paper tackles the challenge of applying importance sampling to deep learning by proposing a method that uses loss values as an importance metric, approximated via a small parallel model, resulting in 30% faster training for a CNN on CIFAR10.
Importance sampling has been successfully used to accelerate stochastic optimization in many convex problems. However, the lack of an efficient way to calculate the importance still hinders its application to Deep Learning. In this paper, we show that the loss value can be used as an alternative importance metric, and propose a way to efficiently approximate it for a deep model, using a small model trained for that purpose in parallel. This method allows in particular to utilize a biased gradient estimate that implicitly optimizes a soft max-loss, and leads to better generalization performance. While such method suffers from a prohibitively high variance of the gradient estimate when using a standard stochastic optimizer, we show that when it is combined with our sampling mechanism, it results in a reliable procedure. We showcase the generality of our method by testing it on both image classification and language modeling tasks using deep convolutional and recurrent neural networks. In particular, our method results in 30% faster training of a CNN for CIFAR10 than when using uniform sampling.