公开(公告)号：US20210366459A1

公开(公告)日：2021-11-25

申请号：US17444557

申请日：2021-08-05

Applicant: Google LLC

Inventor： Matthew Sharifi ,  Aleksandar Kracun

IPC: G10L13/027 ,  G06K9/62 ,  G10L13/08 ,  G10L17/24 ,  G10L25/87

Abstract: A method includes receiving text input data for conversion into synthesized speech and determining, using a hotword-aware model trained to detect a presence of a hotword assigned to a user device, whether a pronunciation of the text input data includes the hotword. The hotword is configured to initiate a wake-up process on the user device for processing the hotword and/or one or more other terms following the hotword in the audio input data. When the pronunciation of the text input data includes the hotword, the method also includes generating an audio output signal from the text input data and providing the audio output signal to an audio output device to output the audio output signal. The audio output signal when captured by an audio capture device of the user device, configured to prevent initiation of the wake-up process on the user device.

公开(公告)号：US20210256379A1

公开(公告)日：2021-08-19

申请号：US17306934

申请日：2021-05-03

Applicant: Google LLC

Inventor： Dominik Roblek ,  Matthew Sharifi

IPC: G06N3/08 ,  G06N3/04 ,  G10L25/30 ,  G06F3/16

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for audio processing using neural networks. One of the systems includes multiple neural network layers, wherein the neural network system is configured to receive time domain features of an audio sample and to process the time domain features to generate a neural network output for the audio sample, the plurality of neural network layers comprising: a frequency-transform (F-T) layer that is configured to apply a transformation defined by a set of F-T layer parameters that transforms a window of time domain features into frequency domain features; and one or more other neural network layers having respective layer parameters, wherein the one or more neural network layers are configured to process frequency domain features to generate a neural network output.

公开(公告)号：US11089457B2

公开(公告)日：2021-08-10

申请号：US16241704

申请日：2019-01-07

Applicant: Google LLC

Inventor： Matthew Sharifi ,  Jorge Pereira ,  Dominik Roblek ,  Julian Odell ,  Cong Li ,  David Petrou

IPC: H04W4/18 ,  H04W4/60 ,  G06F16/248 ,  G06F16/9535 ,  G06F16/2457 ,  H04W4/029 ,  G06F16/907 ,  G06F16/587 ,  G06K9/32 ,  H04L29/08 ,  G06F16/23

Abstract: Systems and methods are provided for a personalized entity repository. For example, a computing device comprises a personalized entity repository having fixed sets of entities from an entity repository stored at a server, a processor, and memory storing instructions that cause the computing device to identify fixed sets of entities that are relevant to a user based on context associated with the computing device, rank the fixed sets by relevancy, and update the personalized entity repository using selected sets determined based on the rank and on set usage parameters applicable to the user. In another example, a method includes generating fixed sets of entities from an entity repository, including location-based sets and topic-based sets, and providing a subset of the fixed sets to a client, the client requesting the subset based on the client's location and on items identified in content generated for display on the client.

公开(公告)号：US20210104221A1

公开(公告)日：2021-04-08

申请号：US16609326

申请日：2018-06-25

Applicant: Google LLC

Inventor： Matthew Sharifi ,  Aleksander Krancun

IPC: G10L13/027 ,  G10L13/08 ,  G10L17/24 ,  G10L25/87 ,  G06K9/62

Abstract: A method includes receiving text input data for conversion into synthesized speech and determining, using a hotword-aware model trained to detect a presence of a hotword assigned to a user device, whether a pronunciation of the text input data includes the hotword. The hotword is configured to initiate a wake-up process on the user device for processing the hotword and/or one or more other terms following the hotword in the audio input data. When the pronunciation of the text input data includes the hotword, the method also includes generating an audio output signal from the text input data and providing the audio output signal to an audio output device to output the audio output signal. The audio output signal when captured by an audio capture device of the user device, configured to prevent initiation of the wake-up process on the user device.

公开(公告)号：US20210056133A1

公开(公告)日：2021-02-25

申请号：US17093551

申请日：2020-11-09

Applicant: Google LLC

Inventor： Matthew Sharifi

IPC: G06F16/487 ,  G06F16/245 ,  G06F16/432 ,  G06F16/435 ,  G06F16/48 ,  G06F16/683 ,  G06F16/955 ,  G06F16/2455 ,  G06F16/783 ,  G06F16/9535 ,  G06F16/2457 ,  G06Q30/06 ,  G06Q30/02

Abstract: Methods, systems, and apparatus for receiving a natural language query of a user, and environmental data, identifying a media item based on the environmental data, determining an entity type based on the natural language query, selecting an entity associated with the media item that matches the entity type, selecting, from a media consumption database that identifies media items that have been indicated as consumed by the user, one or more media items that have been indicated as consumed by the user and that are associated with the selected entity, and providing a response to the query based on selecting the one or more media items that have been indicated as consumed by the user and that are associated with the selected entity.

公开(公告)号：US10809968B2

公开(公告)日：2020-10-20

申请号：US16148338

申请日：2018-10-01

Applicant: Google LLC

Inventor： Dominik Roblek ,  Blaise Hilary Aguera-Arcas ,  Thomas W. Hume ,  Marvin Karl Ritter ,  Brandon Charles Barbello ,  Kevin I. Kilgour ,  Mihajlo Velimirovic ,  Christopher Thornton ,  Gabriel Oak Taubman ,  James David Lyon ,  Jan Heinrich Althaus ,  Katsiaryna Naliuka ,  Julian James Odell ,  Matthew Sharifi ,  Beat Gfeller

IPC: G06F17/00 ,  G06F3/16 ,  G06F16/635 ,  G06F16/683 ,  G06N3/08 ,  G06N20/00

Abstract: In general, the subject matter described in this disclosure can be embodied in methods, systems, and program products. A computing device stores reference song characterization data and receives digital audio data. The computing device determines whether the digital audio data represents music and then performs a different process to recognize that the digital audio data represents a particular reference song. The computing device then outputs an indication of the particular reference song.

公开(公告)号：US10761802B2

公开(公告)日：2020-09-01

申请号：US16148401

申请日：2018-10-01

Applicant: Google LLC

Inventor： Dominik Roblek ,  Blaise Hilary Aguera-Arcas ,  Thomas W. Hume ,  Marvin Karl Ritter ,  Brandon Charles Barbello ,  Kevin I. Kilgour ,  Mihajlo Velimirović ,  Christopher Thornton ,  Gabriel Oak Taubman ,  James David Lyon ,  Jan Heinrich Althaus ,  Katsiaryna Naliuka ,  Julian James Odell ,  Matthew Sharifi ,  Beat Gfeller

IPC: G06F17/00 ,  G06F3/16 ,  G06F16/635 ,  G06F16/683 ,  G06N3/08 ,  G06N20/00

Abstract: In general, the subject matter described in this disclosure can be embodied in methods, systems, and program products for indicating a reference song. A computing device stores reference song characterization data that identifies a plurality of audio characteristics for each reference song in a plurality of reference songs. The computing device receives digital audio data that represents audio recorded by a microphone, converts the digital audio data from time-domain format into frequency-domain format, and uses the digital audio data in the frequency-domain format in a music-characterization process. In response to determining that characterization values for the digital audio data are most relevant to characterization values for a particular reference song, the computing device outputs an indication of the particular reference song.

公开(公告)号：US10663313B2

公开(公告)日：2020-05-26

申请号：US15844006

申请日：2017-12-15

Applicant: Google LLC

Inventor： Matthew Sharifi ,  Jakob Foerster

IPC: G01C21/34 ,  G01C21/36

Abstract: Systems and methods for generating return journey notifications include obtaining a request for navigational directions to a target destination. An outbound journey route from an initial location to the target destination can be determined, wherein the outbound journey route includes an estimated outbound journey time. A return journey route from the target destination to a return destination can be determined, wherein the return journey route includes an estimated return journey time. The outbound journey route and/or return journey route can be determined at least in part from one or more of current traffic conditions or historical traffic conditions. One or more notifications regarding the return journey route can be generated when comparing the estimated outbound journey time to the estimated return journey time results in a determination that one or more predetermined criteria are met.

公开(公告)号：US20200154151A1

公开(公告)日：2020-05-14

申请号：US16741657

申请日：2020-01-13

Applicant: GOOGLE LLC

Inventor： Matthew Sharifi

IPC: H04N21/235 ,  H04N21/84 ,  H04N21/466 ,  H04N21/25 ,  H04N21/234

Abstract: Systems and methods for matching media content are disclosed, including: at a server, obtaining first media content from a client device, wherein the first media content item corresponds to a first portion of media content being played on the client device; obtaining second media content from a content source distinct from the server; comparing the first media content and the second media content; based on a determination that the second media content corresponds to a portion of the media content that is earlier than the first media content: obtaining third media content from the content source corresponding to a third portion of the media content subsequent to the second media content; comparing the first media content with the third media content; and based on a determination that the first and third media content are concurrent, identifying the first media content using identification information corresponding to the third media content.

公开(公告)号：US10593330B2

公开(公告)日：2020-03-17

申请号：US16171495

申请日：2018-10-26

Applicant: Google LLC

Inventor： Matthew Sharifi

IPC: G10L15/28 ,  G10L15/22 ,  G10L15/08 ,  G10L17/22 ,  G10L15/32 ,  G10L15/01 ,  G06F3/16

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for hotword detection on multiple devices are disclosed. In one aspect, a method includes the actions of receiving, by a first computing device, audio data that corresponds to an utterance. The actions further include determining a first value corresponding to a likelihood that the utterance includes a hotword. The actions further include receiving a second value corresponding to a likelihood that the utterance includes the hotword, the second value being determined by a second computing device. The actions further include comparing the first value and the second value. The actions further include based on comparing the first value to the second value, initiating speech recognition processing on the audio data.