On parallelizability of stochastic gradient descent for speech DNNS.

Frank Seide, Hao Fu, Jasha Droppo, Gang Li, Dong Yu

Browse the full ICASSP paper archive.