利索-全球专利检索

  1. 登录
  2. 注册

搜索结果

46 条记录 (耗时 0.022 秒)

    Method and apparatus for predicting word error rates from text
    3.
    发明授权
    Method and apparatus for predicting word error rates from text 有权
    用于从文本中预测字错误率的方法和装置

    公开(公告)号：US07117153B2

    公开(公告)日：2006-10-03

    申请号：US10365850

    申请日：2003-02-13

    申请人： Milind Mahajan ,  Yonggang Deng ,  Alejandro Acero ,  Asela J. R. Gunawardana ,  Ciprian Chelba

    发明人： Milind Mahajan ,  Yonggang Deng ,  Alejandro Acero ,  Asela J. R. Gunawardana ,  Ciprian Chelba

    IPC分类号： G10L15/06 ,  G10L15/14

    CPC分类号： G10L15/197 ,  G10L15/183

    摘要： A method of modeling a speech recognition system includes decoding a speech signal produced from a training text to produce a sequence of predicted speech units. The training text comprises a sequence of actual speech units that is used with the sequence of predicted speech units to form a confusion model. In further embodiments, the confusion model is used to decode a text to identify an error rate that would be expected if the speech recognition system decoded speech based on the text.

    摘要翻译： 对语音识别系统进行建模的方法包括对从训练文本产生的语音信号进行解码以产生预测语音单元的序列。 训练文本包括与预测语音单元的序列一起使用以形成混淆模型的实际语音单元的序列。 在另外的实施例中，混淆模型用于对文本进行解码以识别如果语音识别系统基于文本解码的语音将会预期的错误率。

    Hidden conditional random field models for phonetic classification and speech recognition
    5.
    发明申请
    Hidden conditional random field models for phonetic classification and speech recognition 有权
    用于语音分类和语音识别的隐藏条件随机场模型

    公开(公告)号：US20060085190A1

    公开(公告)日：2006-04-20

    申请号：US10966047

    申请日：2004-10-15

    申请人： Asela Gunawardana ,  Milind Mahajan ,  Alejandro Acero

    发明人： Asela Gunawardana ,  Milind Mahajan ,  Alejandro Acero

    IPC分类号： G10L15/14

    CPC分类号： G10L15/14

    摘要： A method and apparatus are provided for training and using a hidden conditional random field model for speech recognition and phonetic classification. The hidden conditional random field model uses features, at least one of which is based on a hidden state in a phonetic unit. Values for the features are determined from a segment of speech, and these values are used to identify a phonetic unit for the segment of speech.

    摘要翻译： 提供了一种用于训练和使用用于语音识别和语音分类的隐藏条件随机场模型的方法和装置。 隐藏的条件随机场模型使用特征，其中至少一个基于语音单元中的隐藏状态。 特征的值由语音段确定，并且这些值用于识别语音段的语音单元。

    Updating hidden conditional random field model parameters after processing individual training samples
    6.
    发明申请
    Updating hidden conditional random field model parameters after processing individual training samples 有权
    在处理个人培训样本后更新隐藏的条件随机场模型参数

    公开(公告)号：US20070067171A1

    公开(公告)日：2007-03-22

    申请号：US11233148

    申请日：2005-09-22

    申请人： Milind Mahajan ,  Alejandro Acero ,  Asela Gunawardana ,  John Platt

    发明人： Milind Mahajan ,  Alejandro Acero ,  Asela Gunawardana ,  John Platt

    IPC分类号： G10L15/00

    CPC分类号： G10L15/063

    摘要： A method and apparatus are provided for training parameters in a hidden conditional random field model for use in speech recognition and phonetic classification. The hidden conditional random field model uses parameterized features that are determined from a segment of speech, and those values are used to identify a phonetic unit for the segment of speech. The parameters are updated after processing of individual training samples.

    摘要翻译： 提供了一种用于训练用于语音识别和语音分类的隐藏条件随机场模型中的参数的方法和装置。 隐藏条件随机场模型使用从语音段确定的参数化特征，并且这些值用于识别语音段的语音单元。 参数在处理单个培训样本后更新。

    Discriminative training of language models for text and speech classification
    7.
    发明授权
    Discriminative training of language models for text and speech classification 有权
    文本和语言分类语言模型的歧视性训练

    公开(公告)号：US08306818B2

    公开(公告)日：2012-11-06

    申请号：US12103035

    申请日：2008-04-15

    申请人： Ciprian Chelba ,  Alejandro Acero ,  Milind Mahajan

    发明人： Ciprian Chelba ,  Alejandro Acero ,  Milind Mahajan

    IPC分类号： G10L15/00 ,  G06F17/27

    CPC分类号： G06F17/2715 ,  G06F17/2818 ,  G10L15/183 ,  G10L15/197

    摘要： Methods are disclosed for estimating language models such that the conditional likelihood of a class given a word string, which is very well correlated with classification accuracy, is maximized. The methods comprise tuning statistical language model parameters jointly for all classes such that a classifier discriminates between the correct class and the incorrect ones for a given training sentence or utterance. Specific embodiments of the present invention pertain to implementation of the rational function growth transform in the context of a discriminative training technique for n-gram classifiers.

    摘要翻译： 公开了用于估计语言模型的方法，使得给定字串的类的条件似然性与分类准确度非常良好地相关联。 这些方法包括对所有类共同调整统计语言模型参数，使得分类器在给定训练句或话语中区分正确类和不正确类之间的差异。 本发明的具体实施例涉及在n-gram分类器的鉴别训练技术的上下文中实现有理函数增长变换。

    Discriminative training of language models for text and speech classification
    8.
    发明授权
    Discriminative training of language models for text and speech classification 有权
    文本和语言分类语言模型的歧视性训练

    公开(公告)号：US07379867B2

    公开(公告)日：2008-05-27

    申请号：US10453349

    申请日：2003-06-03

    申请人： Ciprian Chelba ,  Alejandro Acero ,  Milind Mahajan

    发明人： Ciprian Chelba ,  Alejandro Acero ,  Milind Mahajan

    IPC分类号： G06F17/27 ,  G10L15/00

    CPC分类号： G06F17/2715 ,  G06F17/2818 ,  G10L15/183 ,  G10L15/197

    摘要： Methods are disclosed for estimating language models such that the conditional likelihood of a class given a word string, which is very well correlated with classification accuracy, is maximized. The methods comprise tuning statistical language model parameters jointly for all classes such that a classifier discriminates between the correct class and the incorrect ones for a given training sentence or utterance. Specific embodiments of the present invention pertain to implementation of the rational function growth transform in the context of a discriminative training technique for n-gram classifiers.

    摘要翻译： 公开了用于估计语言模型的方法，使得给定字串的类的条件似然性与分类准确度非常良好地相关联。 这些方法包括对所有类共同调整统计语言模型参数，使得分类器在给定训练句或话语中区分正确类和不正确类之间的差异。 本发明的具体实施例涉及在n-gram分类器的鉴别训练技术的上下文中实现有理函数增长变换。

    DISCRIMINATIVE TRAINING OF LANGUAGE MODELS FOR TEXT AND SPEECH CLASSIFICATION
    10.
    发明申请
    DISCRIMINATIVE TRAINING OF LANGUAGE MODELS FOR TEXT AND SPEECH CLASSIFICATION 有权
    用于文本和语音分类的语言模式的歧视性培训

    公开(公告)号：US20080215311A1

    公开(公告)日：2008-09-04

    申请号：US12103035

    申请日：2008-04-15

    申请人： Ciprian Chelba ,  Alejandro Acero ,  Milind Mahajan

    发明人： Ciprian Chelba ,  Alejandro Acero ,  Milind Mahajan

    IPC分类号： G06F17/27

    CPC分类号： G06F17/2715 ,  G06F17/2818 ,  G10L15/183 ,  G10L15/197

    摘要： Methods are disclosed for estimating language models such that the conditional likelihood of a class given a word string, which is very well correlated with classification accuracy, is maximized. The methods comprise tuning statistical language model parameters jointly for all classes such that a classifier discriminates between the correct class and the incorrect ones for a given training sentence or utterance. Specific embodiments of the present invention pertain to implementation of the rational function growth transform in the context of a discriminative training technique for n-gram classifiers.

    摘要翻译： 公开了用于估计语言模型的方法，使得给定字串的类的条件似然性与分类准确度非常良好地相关联。 这些方法包括对所有类共同调整统计语言模型参数，使得分类器在给定训练句或话语中区分正确类和不正确类之间的差异。 本发明的具体实施例涉及在n-gram分类器的鉴别训练技术的上下文中实现有理函数增长变换。