Amazon Technologies, Inc.

Storz Endoskop Produktions Gmbh

Sankaran Panchapagesan

Arindam Mandal

Bjorn Hoffmeister

Matteo Contolini

Ted H Applebaum

Shiv Naga Prasad Vitaladevuni

Ming Sun

Ariya Rastrow

Spyridon Matsoukas

Nikko Strom

Baiyang Liu

Ryan Paul Thomas

Shiva Kumar Sundaram

Aparna Khare

Minhua Wu

Kenichi Kumatani

Michael Reese Bastian

An approach to wakeword detection uses an explicit representation of non-wakeword speech in the form of subword (e.g., phonetic monophone) units that do not necessarily occur in the wakeword and that broadly represent general speech. These subword units are arranged in a 'background' model, which at runtime essentially competes with the wakeword model such that a wakeword is less likely to be declare as occurring when the input matches that background model well. An HMM may be used with the model to locate possible occurrences of the wakeword. Features are determined from portions of the input corresponding to subword units of the wakeword detected using the HMM. A secondary classifier is then used to process the features to yield a decision of whether the wakeword occurred.

Monophone-based background modeling for wakeword detection

An approach to speech recognition, and in particular trigger word detection, implements fixed feature extraction form waveform samples with a neural network (NN). For example, rather than computing Log Frequency Band Energies (LFBEs), a convolutional neural network is used. In some implementations, this NN waveform processing is combined with a trained secondary classification that makes use of phonetic segmentation of a possible trigger word occurrence.

Trigger word detection using neural network waveform processing

An approach to keyword spotting makes use of acoustic parameters that are trained on a keyword spotting task as well as on a second speech recognition task, for example, a large vocabulary continuous speech recognition task. The parameters may be optimized according to a weighted measure that weighs the keyword spotting task more highly than the other task, and that weighs utterances of a keyword more highly than utterances of other speech. In some applications, a keyword spotter configured with the acoustic parameters is used for trigger or wake word detection.

Keyword spotting using multi-task configuration

Features are disclosed for error tolerant model compression. Such features could be used to reduce the size of a deep neural network model including several hidden node layers. The size reduction in an error tolerant fashion ensures predictive applications relying on the model do not experience performance degradation due to model compression. Such predictive applications include automatic recognition of speech, image recognition, and recommendation engines. Partially quantized models are re-trained such that any degradation of accuracy is 'trained out' of the model providing improved error tolerance with compression.

Error tolerant neural network model compression

A neural network acoustic model is trained to be robust and produce accurate output when used to process speech signals having acoustic interference. The neural network acoustic model can be trained using a source-separation process by which, in addition to producing the main acoustic model output for a given input, the neural network generates predictions of the separate speech and interference portions of the input. The parameters of the neural network can be adjusted to jointly optimize all three outputs (e.g., the main acoustic model output, the speech signal prediction, and the interference signal prediction), rather than only optimizing the main acoustic model output. Once trained, output layers for the speech and interference signal predictions can be removed from the neural network or otherwise disabled.

Robust neural network acoustic model with side task prediction of reference signals

A voice controlled medical system with improved speech recognition includes a first microphone array, a second microphone array, a controller in communication with the first and second microphone arrays, and a medical device operable by the controller. The controller includes a beam module that generates a first beamed signal using signals from the first microphone array and a second beamed signal using signals from the second microphone array. The controller also includes a comparison module that compares the first and second beamed signals and determines a correlation between the first and second beamed signals. The controller also includes a voice interpreting module that identifies commands within the first and second beamed signals if the correlation is above a correlation threshold. The controller also includes an instrument control module that executes the commands to operate said medical device.

Voice control system with multiple microphone arrays

A voice controlled medical system with improved speech recognition includes a microphone in an operating environment and a medical device. The voice controlled medical system further includes a controller in communication with the microphone and the medical device that operates the medical device by executing audio commands received by the microphone. The voice controlled medical system further includes a calibration signal indicative of distortion in the operating environment which is generated by comparing an audio signal played in the operating environment with a sound signal received by the microphone. The controller uses the calibration signal to interpret audio commands received by the microphone.

System and method for calibrating a speech recognition system to an operating environment

A voice controlled surgical system including a wireless command microphone receiving audio input, a voice control module for generating commands from the audio input received by said wireless command microphone, a detection module for generating signals indicative of a proximity of said wireless command microphone, a switch module for disabling the commands in response to one or more of the signals, and an alarm module activated in response to the one or more of the signals.

Growing community of inventors

Mountain View, CA, United States of America

Sankaran Panchapagesan