tions highlights a new feature of the system. Nor are these improvements speci c to the SwitchThis paper describes recent improvements made to board corpus. Even though the system was trained Dragon's speech recognition system which have imentirely on Switchboard data, we have demonstrated proved performance on Switchboard recognition by strong performance on a \blind" test of English conroughl...

In this paper we show how we have achieved the state-of-theart performance on the industry-standard NIST 2000 Hub5 English evaluation set. We explore densely connected LSTMs, inspired by the densely connected convolutional networks recently introduced for image classification tasks. We also propose an acoustic model adaptation scheme that simply averages the parameters of a seed neural network ...

We present a study of data simulated using acoustic models trained on Switchboard data, and then recognized using various Switchboard-trained acoustic models. When we recognize real Switchboard conversations, simple development models give a word error rate (WER) of about 47 percent. If instead we simulate the speech data using word transcriptions of the conversation, obtaining the pronunciatio...

Very deep CNNs with small 3 × 3 kernels have recently been shown to achieve very strong performance as acoustic models in hybrid NN-HMM speech recognition systems. In this paper, we demonstrate that the accuracy gains of these deep CNNs are retained both on larger scale data, and after sequence training. We show this by carrying out sequence training on both the 300h switchboard-1 and the 2000h...

Models of speech recognition (by both human and machine) have traditionally assumed the phoneme to serve as the fundamental unit of phonetic and phonological analysis. However, phoneme-centric models have failed to provide a convincing theoretical account of the process by which the brain extracts meaning from the speech signal and have fared poorly in automatic recognition of natural, informal...