Nuance Communications, Inc.

Microsoft Technology Licensing, LLC

Puming Zhan

Paul Joseph Vozila

Daniel Willett

Yang Sun

Xinwei Li

Stijn Van Even

Xianju Du

Felix Weninger

Ralf Leibold

Neeraj Deshmukh

Marco Gaudesi

Li Li

An end-to-end automatic speech recognition (ASR) system includes: a first encoder configured for close-talk input captured by a close-talk input mechanism; a second encoder configured for far-talk input captured by a far-talk input mechanism; and an encoder selection layer configured to select at least one of the first and second encoders for use in producing ASR output. The selection is made based on at least one of short-time Fourier transform (STFT), Mel-frequency Cepstral Coefficient (MFCC) and filter bank derived from at least one of the close-talk input and the far-talk input. If signals from both the close-talk input mechanism and the far-talk input mechanism are present for a speech segment, the encoder selection layer dynamically selects between the close-talk encoder and the far-talk encoder to select the encoder that better recognizes the speech segment. An encoder-decoder model is used to produce the ASR output.

Multi-encoder end-to-end automatic speech recognition (ASR) for joint modeling of multiple input devices

A system, method and computer-readable storage device provides an improved speech processing approach in which hyper parameters used for speech recognition are modified dynamically or in batch mode rather than fixed statically. The method includes estimating, via a model trained on audio data and/or metadata, a set of parameters useful for performing automatic speech recognition, receiving speech at an automatic speech recognition system, applying, by the automatic speech recognition system, the set of parameters to processing the speech to yield text and outputting the text from the automatic speech recognition system.

System and method for performing automatic speech recognition system parameter adjustment via machine learning

Techniques for adapting a trained neural network acoustic model, comprising using at least one computer hardware processor to perform: generating initial speaker information values for a speaker; generating first speech content values from first speech data corresponding to a first utterance spoken by the speaker; processing the first speech content values and the initial speaker information values using the trained neural network acoustic model; recognizing, using automatic speech recognition, the first utterance based, at least in part on results of the processing; generating updated speaker information values using the first speech data and at least one of the initial speaker information values and/or information used to generate the initial speaker information values; and recognizing, based at least in part on the updated speaker information values, a second utterance spoken by the speaker.

System and methods for adapting neural network acoustic models

A method of speech recognition converts an unknown speech input into a stream of representative features. The feature stream is transformed based on speaker dependent adaptation of multi-class feature models. Then automatic speech recognition is used to compare the transformed feature stream to multi-class speaker independent acoustic models to generate an output representative of the unknown speech input.

Multi-class constrained maximum likelihood linear regression

A method is described that corrects incorrect text associated with recognition errors in computer-implemented speech recognition. The method includes the step of performing speech recognition on an utterance to produce a recognition result for the utterance. The command includes a word and a phrase. The method includes determining if a word closely corresponds to a portion of the phrase. A speech recognition result is produced if the word closely corresponds to a portion of the phrase.

Growing community of inventors

Acton, MA, United States of America

Puming Zhan