RSO: A Gradient Free Sampling Based Approach For Training Deep Neural Networks
This addresses the problem of gradient-based training limitations for deep learning practitioners, offering a novel alternative but is incremental as it builds on existing sampling methods.
The authors tackled training deep neural networks without gradients by proposing RSO, a gradient-free MCMC-based method that perturbs weights and updates them if loss decreases, achieving 99.1% accuracy on MNIST and 81.8% on CIFAR-10 with 6-10 layer networks.
We propose RSO (random search optimization), a gradient free Markov Chain Monte Carlo search based approach for training deep neural networks. To this end, RSO adds a perturbation to a weight in a deep neural network and tests if it reduces the loss on a mini-batch. If this reduces the loss, the weight is updated, otherwise the existing weight is retained. Surprisingly, we find that repeating this process a few times for each weight is sufficient to train a deep neural network. The number of weight updates for RSO is an order of magnitude lesser when compared to backpropagation with SGD. RSO can make aggressive weight updates in each step as there is no concept of learning rate. The weight update step for individual layers is also not coupled with the magnitude of the loss. RSO is evaluated on classification tasks on MNIST and CIFAR-10 datasets with deep neural networks of 6 to 10 layers where it achieves an accuracy of 99.1% and 81.8% respectively. We also find that after updating the weights just 5 times, the algorithm obtains a classification accuracy of 98% on MNIST.