Combining stochastic average gradient and Hessian-free optimization for sequence training of deep neural networks.

Pierre L. Dognin, Vaibhava Goel

Browse the full ASRU paper archive.