中科微点-全球专利检索

  1. Login
  2. Sign Up

Search Results

17 records (took 0.026 seconds)

    PRIVACY-PRESERVING TRAINING CORPUS SELECTION
    1.
    发明申请
    PRIVACY-PRESERVING TRAINING CORPUS SELECTION 有权
    隐私保护培训选择

    公开(公告)号：US20160379639A1

    公开(公告)日：2016-12-29

    申请号：US14753912

    申请日：2015-06-29

    Applicant: Google Inc.

    Inventor： Eugene Weinstein ,  Pedro J. Moreno Mengibar

    IPC: G10L15/26 ,  G10L15/06

    CPC classification number: G10L15/26 ,  G06F21/6254 ,  G06F2221/2111 ,  G10L15/06 ,  G10L15/063 ,  G10L15/28 ,  G10L2015/0636 ,  H04W4/02 ,  H04W4/12

    Abstract: The present disclosure relates to training a speech recognition system. A system that includes an automated speech recognizer and receives data from a client device. The system determines that at least a portion of the received data is likely sensitive data. Before the at least a portion of the received data is deleted, the system provides the at least a portion of the received data to a model training engine that trains recognition models for the automated speech recognizer. After the at least a portion of the received data is provided, the system deletes the at least a portion of the received data.

    Abstract translation: 本公开涉及训练语音识别系统。 一种包括自动语音识别器并从客户端设备接收数据的系统。 系统确定接收到的数据的至少一部分可能是敏感数据。 在所接收的数据的至少一部分被删除之前，系统将接收的数据的至少一部分提供给训练用于自动语音识别器的识别模型的模型训练引擎。 在提供所接收的数据的至少一部分之后，系统删除所接收的数据的至少一部分。

    ACOUSTIC MODEL TRAINING CORPUS SELECTION
    4.
    发明申请
    ACOUSTIC MODEL TRAINING CORPUS SELECTION 有权

    公开(公告)号：US20160267903A1

    公开(公告)日：2016-09-15

    申请号：US15164263

    申请日：2016-05-25

    Applicant: Google Inc.

    Inventor： Olga Kapralova ,  John Paul Alex ,  Eugene Weinstein ,  Pedro J. Moreno Mengibar ,  Olivier Siohan ,  Ignacio Lopez Moreno

    IPC: G10L15/06 ,  G10L15/01 ,  G10L15/16 ,  G10L15/30

    CPC classification number: G10L15/063 ,  G10L15/01 ,  G10L15/16 ,  G10L15/187 ,  G10L15/30 ,  G10L15/32 ,  G10L2015/0633

    Abstract: The present disclosure relates to training a speech recognition system. One example method includes receiving a collection of speech data items, wherein each speech data item corresponds to an utterance that was previously submitted for transcription by a production speech recognizer. The production speech recognizer uses initial production speech recognizer components in generating transcriptions of speech data items. A transcription for each speech data item is generated using an offline speech recognizer, and the offline speech recognizer components are configured to improve speech recognition accuracy in comparison with the initial production speech recognizer components. The updated production speech recognizer components are trained for the production speech recognizer using a selected subset of the transcriptions of the speech data items generated by the offline speech recognizer. An updated production speech recognizer component is provided to the production speech recognizer for use in transcribing subsequently received speech data items.

    ACOUSTIC MODEL TRAINING CORPUS SELECTION
    5.
    发明申请
    ACOUSTIC MODEL TRAINING CORPUS SELECTION 有权
    ACOUSTIC MODEL TRAINING CORPUS选择

    公开(公告)号：US20160093294A1

    公开(公告)日：2016-03-31

    申请号：US14693268

    申请日：2015-04-22

    Applicant: Google Inc.

    Inventor： Olga Kapralova ,  John Paul Alex ,  Eugene Weinstein ,  Pedro J. Moreno Mengibar ,  Olivier Siohan ,  Ignacio Lopez Moreno

    IPC: G10L15/06 ,  G10L15/187 ,  G10L15/26 ,  G10L25/30

    CPC classification number: G10L15/063 ,  G10L15/01 ,  G10L15/16 ,  G10L15/187 ,  G10L15/30 ,  G10L15/32 ,  G10L2015/0633

    Abstract: The present disclosure relates to training a speech recognition system. One example method includes receiving a collection of speech data items, wherein each speech data item corresponds to an utterance that was previously submitted for transcription by a production speech recognizer. The production speech recognizer uses initial production speech recognizer components in generating transcriptions of speech data items. A transcription for each speech data item is generated using an offline speech recognizer, and the offline speech recognizer components are configured to improve speech recognition accuracy in comparison with the initial production speech recognizer components. The updated production speech recognizer components are trained for the production speech recognizer using a selected subset of the transcriptions of the speech data items generated by the offline speech recognizer. An updated production speech recognizer component is provided to the production speech recognizer for use in transcribing subsequently received speech data items.

    Abstract translation: 本公开涉及训练语音识别系统。 一个示例性方法包括接收语音数据项集合，其中每个语音数据项对应于先前由生产语音识别器提交用于转录的话语。 生产语音识别器使用初始生产语音识别器组件来产生语音数据项的转录。 使用离线语音识别器生成每个语音数据项的转录，并且将离线语音识别器组件配置为与初始制作语音识别器组件相比提高语音识别精度。 使用由离线语音识别器生成的语音数据项的转录的所选择的子集来对生产语音识别器进行更新的制作语音识别器组件的训练。 更新的生产语音识别器组件被提供给生产语音识别器，用于转录随后接收的语音数据项。

    CACHING SPEECH RECOGNITION SCORES
    6.
    发明申请
    CACHING SPEECH RECOGNITION SCORES 有权
    缓存语音识别码

    公开(公告)号：US20150371631A1

    公开(公告)日：2015-12-24

    申请号：US14311557

    申请日：2014-06-23

    Applicant: Google Inc.

    Inventor： Eugene Weinstein ,  Sanjiv Kumar ,  Ignacio L. Moreno ,  Andrew W. Senior ,  Nikhil Prasad Bhat

    IPC: G10L15/14 ,  G10L19/038

    CPC classification number: G10L15/08 ,  G10L15/285

    Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for caching speech recognition scores. In some implementations, one or more values comprising data about an utterance are received. An index value is determined for the one or more values. An acoustic model score for the one or more received values is selected, from a cache of acoustic model scores that were computed before receiving the one or more values, based on the index value. A transcription for the utterance is determined using the selected acoustic model score.

    Abstract translation: 方法，系统和装置，包括编码在计算机存储介质上的用于缓存语音识别分数的计算机程序。 在一些实现中，接收包括关于话语的数据的一个或多个值。 确定一个或多个值的索引值。 基于索引值，从接收到一个或多个值之前计算的声学模型分数的高速缓存中选择一个或多个接收值的声学模型分数。 使用所选择的声学模型得分确定发音的转录。

    Two-pass decoding for speech recognition of search and action requests
    7.
    发明授权
    Two-pass decoding for speech recognition of search and action requests 有权
    用于搜索和动作请求的语音识别的双程解码

    公开(公告)号：US08645138B1

    公开(公告)日：2014-02-04

    申请号：US13723191

    申请日：2012-12-20

    Applicant: Google Inc.

    Inventor： Eugene Weinstein ,  Austin Waters

    IPC: G10L15/00 ,  G10L21/00

    CPC classification number: G10L15/19

    Abstract: Disclosed are apparatus and methods for processing spoken speech. Input speech can be received at a computing system. During a first pass of speech recognition, a plurality of language model outputs can be determined by: providing the input speech to each of a plurality of language models and responsively receiving a language model output from each language model. A language model of the plurality of language models can be selected using a classifier operating on the plurality of language model outputs. During a second pass of speech recognition, a revised language model output can be determined by: providing the input speech and the language model output from the selected language model to the selected language model and responsively receiving the revised language model output from the selected language model. The computing system can generate a result based on the revised language model output.

    Abstract translation: 公开了用于处理口头语音的装置和方法。 可以在计算系统处接收输入语音。 在语音识别的第一次通过期间，可以通过以下方式来确定多个语言模型输出：将输入语音提供给多个语言模型中的每一个并且响应地接收来自每个语言模型的语言模型输出。 可以使用在多个语言模型输出上操作的分类器来选择多个语言模型的语言模型。 在语音识别的第二次通过期间，可以通过以下方式来确定经修改的语言模型输出：将所选语言模型的输入语音和语言模型输出提供给所选择的语言模型，并响应于接收来自所选语言模型的修订语言模型输出 。 计算系统可以根据修订后的语言模型输出生成一个结果。

    Caching speech recognition scores
    8.
    发明授权
    Caching speech recognition scores 有权

    公开(公告)号：US09858922B2

    公开(公告)日：2018-01-02

    申请号：US14311557

    申请日：2014-06-23

    Applicant: Google Inc.

    Inventor： Eugene Weinstein ,  Sanjiv Kumar ,  Ignacio L. Moreno ,  Andrew W. Senior ,  Nikhil Prasad Bhat

    IPC: G10L15/08 ,  G10L15/28

    CPC classification number: G10L15/08 ,  G10L15/285

    Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for caching speech recognition scores. In some implementations, one or more values comprising data about an utterance are received. An index value is determined for the one or more values. An acoustic model score for the one or more received values is selected, from a cache of acoustic model scores that were computed before receiving the one or more values, based on the index value. A transcription for the utterance is determined using the selected acoustic model score.

    Acoustic model training corpus selection
    9.
    发明授权
    Acoustic model training corpus selection 有权
    声学模型训练语料库选择

    公开(公告)号：US09378731B2

    公开(公告)日：2016-06-28

    申请号：US14693268

    申请日：2015-04-22

    Applicant: Google Inc.

    Inventor： Olga Kapralova ,  John Paul Alex ,  Eugene Weinstein ,  Pedro J. Moreno Mengibar ,  Olivier Siohan ,  Ignacio Lopez Moreno

    IPC: G10L15/00 ,  G10L15/06 ,  G10L25/30 ,  G10L15/187 ,  G10L15/26

    CPC classification number: G10L15/063 ,  G10L15/01 ,  G10L15/16 ,  G10L15/187 ,  G10L15/30 ,  G10L15/32 ,  G10L2015/0633

    Abstract: The present disclosure relates to training a speech recognition system. One example method includes receiving a collection of speech data items, wherein each speech data item corresponds to an utterance that was previously submitted for transcription by a production speech recognizer. The production speech recognizer uses initial production speech recognizer components in generating transcriptions of speech data items. A transcription for each speech data item is generated using an offline speech recognizer, and the offline speech recognizer components are configured to improve speech recognition accuracy in comparison with the initial production speech recognizer components. The updated production speech recognizer components are trained for the production speech recognizer using a selected subset of the transcriptions of the speech data items generated by the offline speech recognizer. An updated production speech recognizer component is provided to the production speech recognizer for use in transcribing subsequently received speech data items.

    Abstract translation: 本公开涉及训练语音识别系统。 一个示例性方法包括接收语音数据项集合，其中每个语音数据项对应于先前由生产语音识别器提交用于转录的话语。 生产语音识别器使用初始生产语音识别器组件来产生语音数据项的转录。 使用离线语音识别器生成每个语音数据项的转录，并且将离线语音识别器组件配置为与初始制作语音识别器组件相比提高语音识别精度。 使用由离线语音识别器生成的语音数据项的转录的所选择的子集来对生产语音识别器进行更新的制作语音识别器组件的训练。 更新的生产语音识别器组件被提供给生产语音识别器，用于转录随后接收的语音数据项。

Patent Agency Ranking