Scalable minimum Bayes risk training of deep neural network acoustic models using distributed hessian-free optimization

Abstract
No abstract available