Abstract: Embodiments improve speech data quality through training a neural network for de-noising audio enhancement. One such embodiment creates simulated noisy speech data from high quality speech data. In turn, training, e.g., deep normalizing flow training, is performed on a neural network using the high quality speech data and the simulated noisy speech data to train the neural network to create de-noised speech data given noisy speech data. Performing the training includes minimizing errors in the neural network according to at least one of (i) a decoding error of an Automatic Speech Recognition (ASR) system processing current de-noised speech data results generated by the neural network during the training and (ii) spectral distance between the high quality speech data and the current de-noised speech data results generated by the neural network during the training.

Abstract: A method for removing private data from an acoustic model includes capturing speech from a large population of users, creating a text-to-speech voice from at least a portion of the large population of users, discarding speech data from a database of speech, creating text-to-speech waveforms from the text-to-speech voice and the new database of speech with the discarded speech data and generating an automatic speech recognition model using the text-to-speech waveforms.

Abstract: A method, computer program product, and computing system for receiving a plurality of signals from a plurality of microphones, thus defining a plurality of channels. A weighted multichannel representation of the plurality of channels may be generated. A plurality of weights for each channel of the plurality of channels may be generated based upon, at least in part, the weighted multichannel representation of the plurality of channels. A single channel representation of the plurality of channels may be generated based upon, at least in part, the weighted multichannel representation of the plurality of channels and the plurality of weights generated for each channel of the plurality of channels.

Abstract: Embodiments improve speech data quality through training a neural network for de-noising audio enhancement. One such embodiment creates simulated noisy speech data from high quality speech data. In turn, training, e.g., deep normalizing flow training, is performed on a neural network using the high quality speech data and the simulated noisy speech data to train the neural network to create de-noised speech data given noisy speech data. Performing the training includes minimizing errors in the neural network according to at least one of (i) a decoding error of an Automatic Speech Recognition (ASR) system processing current de-noised speech data results generated by the neural network during the training and (ii) spectral distance between the high quality speech data and the current de-noised speech data results generated by the neural network during the training.

Abstract: A method for removing private data from an acoustic model includes capturing speech from a large population of users, creating a text-to-speech voice from at least a portion of the large population of users, discarding speech data from a database of speech, creating text-to-speech waveforms from the text-to-speech voice and the new database of speech with the discarded speech data and generating an automatic speech recognition model using the text-to-speech waveforms.

Abstract: Methods described herein provide functionality for automatic speech recognition (ASR). One such embodiment performs speech recognition using received speech recognition result candidates, where the received candidates were generated by performing Statistical Language Model (SLM) based speech recognition on one or more frames of audio data. In turn, such an embodiment transmits results of the speech recognition, performed using the received speech recognition result candidates, to a user device via a communications network. Results of the speech recognition are available with lower latency than pure cloud based ASR solution.

Abstract: Methods described herein provide functionality for automatic speech recognition (ASR). One such embodiment performs speech recognition using received speech recognition result candidates, where the received candidates were generated by performing Statistical Language Model (SLM) based speech recognition on one or more frames of audio data. In turn, such an embodiment transmits results of the speech recognition, performed using the received speech recognition result candidates, to a user device via a communications network. Results of the speech recognition are available with lower latency than pure cloud based ASR solution.

Abstract: Method and apparatus for providing visual feedback on an electronic device in a client/server speech recognition system comprising the electronic device and a network device remotely located from the electronic device. The method comprises processing, by an embedded speech recognizer of the electronic device, at least a portion of input audio comprising speech to produce local recognized speech, sending at least a portion of the input audio to the network device for remote speech recognition, and displaying, on a user interface of the electronic device, visual feedback based on at least a portion of the local recognized speech prior to receiving streaming recognition results from the network device.

Abstract: Methods described herein provide functionality for automatic speech recognition (ASR). One such embodiment performs speech recognition using received speech recognition result candidates, where the received candidates were generated by performing Statistical Language Model (SLM) based speech recognition on one or more frames of audio data. In turn, such an embodiment transmits results of the speech recognition, performed using the received speech recognition result candidates, to a user device via a communications network. Results of the speech recognition are available with lower latency than pure cloud based ASR solutions.