Improving Keyword Spotting with a Tandem BLSTM-DBN Architecture

Abstract. We propose a novel architecture for keyword spotting which is composed of a Dynamic Bayesian Network (DBN) and a bidirectional Long Short-Term Memory (BLSTM) recurrent neural net. The DBN uses a hidden garbage variable as well as the concept of switching parents to discriminate between keywords and arbitrary speech. Contextual informa-tion is incorporated by a BLSTM network, providing a discrete phoneme prediction feature for the DBN. Together with continuous acoustic fea-tures, the discrete BLSTM output is processed by the DBN which detect