Deep Layers as Stochastic Solvers

Deep Layers as Stochastic Solvers

Adel Bibi, Bernard Ghanem, Vladlen Koltun, Rene Ranftl,
"Deep Layers as Stochastic Solvers",​ 
International Conference on Learning Representation (ICLR 2019)​.
Adel Bibi, Bernard Ghanem, Vladlen Koltun, Rene Ranftl
deep networks, optimization
2019
​​We provide a novel perspective on the forward pass through a block of layers in a deep network. In particular, we show that a forward pass through a standard dropout layer followed by a linear layer and a non-linear activation is equivalent to optimizing a convex optimization objective with a single iteration of a au-nice Proximal Stochastic Gradient method. We further show that replacing standard Bernoulli dropout with additive dropout is equivalent to optimizing the same convex objective with a variance-reduced proximal method. By expressing both fully-connected and convolutional layers as special cases of a high-order tensor product, we unify the underlying convex optimization problem in the tensor setting and derive a formula for the Lipschitz constant L used to determine the optimal step size of the above proximal methods. We conduct experiments with standard convolutional networks applied to the CIFAR-10 and CIFAR-100 datasets, and show that replacing a block of layers with multiple iterations of the corresponding solver, with step size set via L, consistently improves classification accuracy.