ASYNCHRONOUS STOCHASTIC OPTIMIZATION FOR SEQUENCE TRAINING OF DEEP NEURAL NETWORKS

This paper explores asynchronous stochastic optimization for se-quence training of deep neural networks. Sequence training requires more computation than frame-level training using pre-computed frame data. This leads to several complications for stochastic op-timization, arising from significant asynchrony in model updates under massive parallelization, and limited data shuffling due to utterance-chunked processing. We analyze the impact of these two issues on the efficiency and performance of sequence training. In particular, we suggest a framework to formalize the reasoning about the asynchrony and present experimental results on both small and large scale Voice Search tasks to validate the effectiveness and efficiency of asynchronous stochastic optimization. Index Terms — speech recognition, acoustic modeling, neural networks, sequence training, asynchronous stochastic optimizatio