知搜-全球专利检索

  1. Login
  2. Sign Up

Search Results

12 records (took 0.017 seconds)

    AUTOMATIC SPEAKER IDENTIFICATION USING SPEECH RECOGNITION FEATURES
    2.
    发明申请
    AUTOMATIC SPEAKER IDENTIFICATION USING SPEECH RECOGNITION FEATURES 审中-公开

    公开(公告)号：US20200349957A1

    公开(公告)日：2020-11-05

    申请号：US15929795

    申请日：2020-05-21

    Applicant: Amazon Technologies, Inc.

    Inventor： Hugh Evan Secker-Walker ,  Baiyang Liu ,  Frederick Victor Weber

    IPC: G10L17/06 ,  G10L17/12 ,  G10L17/02 ,  G10L17/16 ,  G10L15/18 ,  G10L17/22

    Abstract: Features are disclosed for automatically identifying a speaker. Artifacts of automatic speech recognition (“ASR”) and/or other automatically determined information may be processed against individual user profiles or models. Scores may be determined reflecting the likelihood that individual users made an utterance. The scores can be based on, e.g., individual components of Gaussian mixture models (“GMMs”) that score best for frames of audio data of an utterance. A user associated with the highest likelihood score for a particular utterance can be identified as the speaker of the utterance. Information regarding the identified user can be provided to components of a spoken language processing system, separate applications, etc.

    Error tolerant neural network model compression
    4.
    发明授权
    Error tolerant neural network model compression 有权

    公开(公告)号：US10229356B1

    公开(公告)日：2019-03-12

    申请号：US14581969

    申请日：2014-12-23

    Applicant: Amazon Technologies, Inc.

    Inventor： Baiyang Liu ,  Michael Reese Bastian ,  Bjorn Hoffmeister ,  Sankaran Panchapagesan ,  Ariya Rastrow

    IPC: G06N3/08

    Abstract: 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.

    AUTOMATIC SPEAKER IDENTIFICATION USING SPEECH RECOGNITION FEATURES
    8.
    发明申请
    AUTOMATIC SPEAKER IDENTIFICATION USING SPEECH RECOGNITION FEATURES 审中-公开

    公开(公告)号：US20190378517A1

    公开(公告)日：2019-12-12

    申请号：US16448788

    申请日：2019-06-21

    Applicant: Amazon Technologies, Inc.

    Inventor： Hugh Evan Secker-Walker ,  Baiyang Liu ,  Frederick Victor Weber

    IPC: G10L17/06 ,  G10L17/22 ,  G10L15/18 ,  G10L17/16 ,  G10L17/02 ,  G10L17/12

    Abstract: Features are disclosed for automatically identifying a speaker. Artifacts of automatic speech recognition (“ASR”) and/or other automatically determined information may be processed against individual user profiles or models. Scores may be determined reflecting the likelihood that individual users made an utterance. The scores can be based on, e.g., individual components of Gaussian mixture models (“GMMs”) that score best for frames of audio data of an utterance. A user associated with the highest likelihood score for a particular utterance can be identified as the speaker of the utterance. Information regarding the identified user can be provided to components of a spoken language processing system, separate applications, etc.

    Automatic speaker identification using speech recognition features
    10.
    发明授权
    Automatic speaker identification using speech recognition features 有权
    自动扬声器识别使用语音识别功能

    公开(公告)号：US09558749B1

    公开(公告)日：2017-01-31

    申请号：US13957257

    申请日：2013-08-01

    Applicant: Amazon Technologies, Inc.

    Inventor： Hugh Evan Secker-Walker ,  Baiyang Liu ,  Frederick Victor Weber

    IPC: G10L15/00 ,  G10L17/00 ,  G10L17/12

    CPC classification number: G10L17/06 ,  G10L15/18 ,  G10L15/20 ,  G10L15/26 ,  G10L17/02 ,  G10L17/12 ,  G10L17/16 ,  G10L17/22 ,  G10L2015/025 ,  G10L2015/088

    Abstract: Features are disclosed for automatically identifying a speaker. Artifacts of automatic speech recognition (“ASR”) and/or other automatically determined information may be processed against individual user profiles or models. Scores may be determined reflecting the likelihood that individual users made an utterance. The scores can be based on, e.g., individual components of Gaussian mixture models (“GMMs”) that score best for frames of audio data of an utterance. A user associated with the highest likelihood score for a particular utterance can be identified as the speaker of the utterance. Information regarding the identified user can be provided to components of a spoken language processing system, separate applications, etc.

    Abstract translation: 公开了用于自动识别扬声器的特征。 自动语音识别（“ASR”）和/或其他自动确定的信息的工件可以针对各个用户简档或模型进行处理。 可以确定反映个人用户发声的可能性的得分。 分数可以基于例如对于语音的音频数据的帧最佳得分的高斯混合模型（“GMM”）的各个组件。 与特定话语的最高似然分数相关联的用户可以被识别为话语的说话者。 关于识别的用户的信息可以被提供给口语处理系统的组件，单独的应用等。

Patent Agency Ranking