知搜-全球专利检索

  1. 登录
  2. 注册

搜索结果

7 条记录 (耗时 0.013 秒)

    Apparatus and methods for machine learning hypotheses
    1.
    发明授权
    Apparatus and methods for machine learning hypotheses 失效
    机器学习假设的装置和方法

    公开(公告)号：US5819247A

    公开(公告)日：1998-10-06

    申请号：US902106

    申请日：1997-07-29

    申请人： Yoav Freund ,  Robert Elias Schapire

    发明人： Yoav Freund ,  Robert Elias Schapire

    IPC分类号： G06F15/18 ,  G06F19/00 ,  G06N3/08 ,  G06T7/00

    CPC分类号： G06K9/6256 ,  G06N3/08 ,  G06N99/005

    摘要： Apparatus and methods for machine learning the hypotheses used in the classifier component of pattern classification devices such as OCRs, other image analysis systems, and and text retrieval systems. The apparatus and methods employ machine learning techniques for generating weak hypotheses from a set of examples of the patterns to be recognized and then evaluate the resulting hypothesis against example patterns. The results of the evaluation are used to increase the probability that the examples used to generate the next weak hypothesis are ones which the previous weak hypothesis did not correctly classify. The results of the evaluation are also used to give a weight to each weak hypothesis. A strong hypothesis is then made by combining the weak hypotheses according to their weights.

    摘要翻译： 用于机器学习在诸如OCR，其​​他图像分析系统和文本检索系统之类的模式分类装置的分类器组件中使用的假设的装置和方法。 该装置和方法采用机器学习技术从用于识别的模式的一组示例中产生弱假设，然后根据示例模式评估所得到的假设。 评估结果用于增加用于产生下一个弱假设的实例是以前的弱假设没有正确分类的概率。 评估结果也用于给每个弱假设加权。 然后通过将弱假设根据其权重组合来进行强烈的假设。

    Combining active and semi-supervised learning for spoken language understanding
    2.
    发明授权
    Combining active and semi-supervised learning for spoken language understanding 有权
    结合积极和半监督的学习语言理解

    公开(公告)号：US08010357B2

    公开(公告)日：2011-08-30

    申请号：US11033902

    申请日：2005-01-12

    申请人： Dilek Z. Hakkani-Tur ,  Robert Elias Schapire ,  Gokhan Tur

    发明人： Dilek Z. Hakkani-Tur ,  Robert Elias Schapire ,  Gokhan Tur

    IPC分类号： G10L15/06

    CPC分类号： G10L15/1822 ,  G10L15/063 ,  G10L15/183

    摘要： Combined active and semi-supervised learning to reduce an amount of manual labeling when training a spoken language understanding model classifier. The classifier may be trained with human-labeled utterance data. Ones of a group of unselected utterance data may be selected for manual labeling via active learning. The classifier may be changed, via semi-supervised learning, based on the selected ones of the unselected utterance data.

    摘要翻译： 组合主动和半监督学习，在训练口语语言理解模型分类器时减少手动标注量。 分类器可以用人标记的话语数据进行训练。 可以通过主动学习选择一组未选择的话语数据进行手动标注。 分类器可以通过半监督学习，基于所选择的未被选择的话语数据来改变。

    Active learning for spoken language understanding
    3.
    发明授权
    Active learning for spoken language understanding 失效
    积极学习口语理解

    公开(公告)号：US07742918B1

    公开(公告)日：2010-06-22

    申请号：US11773681

    申请日：2007-07-05

    申请人： Dilek Z. Hakkani-Tur ,  Robert Elias Schapire ,  Gokhan Tur

    发明人： Dilek Z. Hakkani-Tur ,  Robert Elias Schapire ,  Gokhan Tur

    IPC分类号： G10L15/06

    CPC分类号： G10L15/063

    摘要： Disclosed is a system and method of training a spoken language understanding module. Such a module may be utilized in a spoken dialog system. The method of training a spoken language understanding module comprises training acoustic and language models using a small set of transcribed data St, recognizing utterances in a set Su that are candidates for transcription using the acoustic and language models, computing confidence scores of the utterances, selecting k utterances that have the smallest confidence scores from Su and transcribing them into a new set Si, redefining St as the union of St and Si, redefining Su as Su minus Si, and returning to the step of training acoustic and language models if word accuracy has not converged.

    摘要翻译： 公开了一种训练口语理解模块的系统和方法。 这样的模块可以在口语对话系统中使用。 训练口语理解模块的方法包括使用一小组转录数据St来训练声学和语言模型，使用声学和语言模型识别作为用于转录的候选者的集合Su中的话语，计算话语的置信度分数，选择 从苏的信心得分最小的k k and and Si Si Si Si Si，，ining ining of Si Si Si Si Si Si accuracy accuracy accuracy as as accuracy accuracy accuracy accuracy accuracy accuracy accuracy accuracy accuracy accuracy accuracy accuracy accuracy accuracy accuracy accuracy accuracy accuracy accuracy accuracy accuracy 没有收敛。

    Combining active and semi-supervised learning for spoken language understanding
    4.
    发明申请
    Combining active and semi-supervised learning for spoken language understanding 有权
    结合积极和半监督的学习语言理解

    公开(公告)号：US20090063145A1

    公开(公告)日：2009-03-05

    申请号：US11033902

    申请日：2005-01-12

    申请人： Dilek Z. Hakkani-Tur ,  Robert Elias Schapire ,  Gokham Tur

    发明人： Dilek Z. Hakkani-Tur ,  Robert Elias Schapire ,  Gokham Tur

    IPC分类号： G10L15/06

    CPC分类号： G10L15/1822 ,  G10L15/063 ,  G10L15/183

    摘要： Combined active and semi-supervised learning to reduce an amount of manual labeling when training a spoken language understanding model classifier. The classifier may be trained with human-labeled utterance data. Ones of a group of unselected utterance data may be selected for manual labeling via active learning. The classifier may be changed, via semi-supervised learning, based on the selected ones of the unselected utterance data.

    摘要翻译： 组合主动和半监督学习，在训练口语语言理解模型分类器时减少手动标注量。 分类器可以用人标记的话语数据进行训练。 可以通过主动学习选择一组未选择的话语数据进行手动标注。 分类器可以通过半监督学习，基于所选择的未被选择的话语数据来改变。

    Active learning for spoken language understanding
    5.
    发明授权
    Active learning for spoken language understanding 有权
    积极学习口语理解

    公开(公告)号：US07263486B1

    公开(公告)日：2007-08-28

    申请号：US10404699

    申请日：2003-04-01

    申请人： Dilek Z. Hakkani-Tur ,  Robert Elias Schapire ,  Gokhan Tur

    发明人： Dilek Z. Hakkani-Tur ,  Robert Elias Schapire ,  Gokhan Tur

    IPC分类号： G10L15/16

    CPC分类号： G10L15/063

    摘要： Disclosed is a system and method of training a spoken language understanding module. Such a module may be utilized in a spoken dialog system. The method of training a spoken language understanding module comprises training acoustic and language models using a small set of transcribed data ST, recognizing utterances in a set Su that are candidates for transcription using the acoustic and language models, computing confidence scores of the utterances, selecting k utterances that have the smallest confidence scores from Su and transcribing them into a new set Si, redefining St as the union of St and Si, redefining Su as Su minus Si, and returning to the step of training acoustic and language models if word accuracy has not converged.

    摘要翻译： 公开了一种训练口语理解模块的系统和方法。 这样的模块可以在口语对话系统中使用。 训练口语理解模块的方法包括使用一小组转录数据S IN来训练声学和语言模型，识别作为候选语言的候选语言的集合S < 使用声学和语言模型进行转录，计算话语的置信度分数，从S＆lt; U＆gt;中选择具有最小置信度分数的k个话语，并将它们转录成新的集合S < ，重新定义为S＆lt; t＆gt;和S＆lt; i＆lt; i＆lt; i＆gt;的并集，将S 重新定义为S＆lt; 如果字精度没有收敛，则返回到训练声学和语言模型的步骤。