专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

1. 发明公开

EP1526504A1 Multiple models integration for multi-environment speech recognition 失效
标题翻译：多种模式的整合在不同的环境下的语音识别
公开(公告)号：EP1526504A1
公开(公告)日：2005-04-27
申请号：EP05001473.7
申请日：1998-05-14
申请人： AT&T Corp.
发明人： Rahim, Mazin G.
IPC分类号： G10L15/20
CPC分类号： G10L15/20 , G10L15/144 , G10L21/0216 , G10L2015/228
摘要： The present invention relates to a method for recognizing an unknown speech signal comprising: storing, for each of a plurality of acoustic environments, information that defines a set of recognition models for that acoustic environment; receiving a signal representing unknown speech; identifying as the acoustic environment of the unknown speech a particular one of the plurality of acoustic environments; and recognizing the unknown speech signal using the set of recognition models for the identified acoustic environment, wherein the information that defines the set of recognition models for each acoustic environment includes a base set of recognition models; and information defining differences between the values of particular parameters of the base set of recognition models and the values of the corresponding parameters of the set of recognition models for each acoustic environment. Further, the invention also pertains to a signal processing system for recognizing a test utterance and method for developing the set of speech recognition models.

2. 发明公开

EP0881625A3 Multiple models integration for multi-environment speech recognition 失效
标题翻译：多种模式的整合在不同的环境下的语音识别
公开(公告)号：EP0881625A3
公开(公告)日：1999-07-28
申请号：EP98108805.7
申请日：1998-05-14
申请人： AT&T Corp.
发明人： Rahim, Mazin G.
IPC分类号： G10L3/00
CPC分类号： G10L15/20 , G10L15/144 , G10L21/0216 , G10L2015/228
摘要： A speech recognition system which effectively recognizes unknown speech from multiple acoustic environments includes a set of secondary models, each associated with one or more particular acoustic environments, integrated with a base set of recognition models. The speech recognition system is trained by making a set of secondary models in a first stage of training, and integrating the set of secondary models with a base set of recognition models in a second stage of training.

3. 发明授权

EP1526504B1 Multiple models integration for multi-environment speech recognition 失效
标题翻译：多种模式的整合在不同的环境下的语音识别
公开(公告)号：EP1526504B1
公开(公告)日：2007-08-01
申请号：EP05001473.7
申请日：1998-05-14
申请人： AT&T Corp.
发明人： Rahim, Mazin G.
IPC分类号： G10L15/20
CPC分类号： G10L15/20 , G10L15/144 , G10L21/0216 , G10L2015/228
摘要： A speech recognition system which effectively recognizes unknown speech from multiple acoustic environments includes a set of secondary models, each associated with one or more particular acoustic environments, integrated with a base set of recognition models. The speech recognition system is trained by making a set of secondary models in a first stage of training, and integrating the set of secondary models with a base set of recognition models in a second stage of training.

4. 发明公开

EP1679693A1 A system and method of providing an automated data-collection in spoken dialog systems 审中-公开
标题翻译：系统和Verfahren zur Bereitstellung einer automatisierten Datensammlung在Sprachdialogsystemen
公开(公告)号：EP1679693A1
公开(公告)日：2006-07-12
申请号：EP06100060.0
申请日：2006-01-04
申请人： AT&T Corp.
发明人： Di Fabbrizio, Giuseppe , Hakkani-Tur, Dilek, Z. , Rahim, Mazin G. , Renger, Bernard S. , Tur, Gokhan
IPC分类号： G10L15/22 , G10L15/18
CPC分类号： G10L15/063 , G10L15/183 , G10L15/22
摘要： The invention relates to a system and method for gathering data for use in a spoken dialog system. An aspect of the invention is generally referred to as an automated hidden human that performs data collection automatically at the beginning of a conversation with a user in a spoken dialog system. The method comprises presenting an initial prompt to a user, recognizing a received user utterance using an automatic speech recognition engine and classifying the recognized user utterance using a spoken language understanding module. If the recognized user utterance is not understood or classifiable to a predetermined acceptance threshold, then the method re-prompts the user. If the recognized user utterance is not classifiable to a predetermined rejection threshold, then the method transfers the user to a human as this may imply a task-specific utterance. The received and classified user utterance is then used for training the spoken dialog system.
摘要翻译：本发明涉及一种用于收集在口头对话系统中使用的数据的系统和方法。本发明的一个方面通常被称为在与对话系统中的用户的对话开始时自动执行数据收集的自动隐藏人物。该方法包括向用户呈现初始提示，使用自动语音识别引擎识别所接收的用户话语，并使用口语理解模块对所识别的用户话语进行分类。如果识别的用户话语不能被理解或可被分类到预定的接受阈值，则该方法重新提示用户。如果识别的用户话语不能被分类为预定的拒绝阈值，则该方法将用户转移给人，因为这可能意味着特定于任务的话语。接收和分类的用户话语随后用于训练口语对话系统。

5. 发明公开

EP1280136A1 Spoken language understanding that incorporates prior knowledge into boosting 审中-公开
标题翻译： Sprachverständnismit Vorwissen zurErhöhung
公开(公告)号：EP1280136A1
公开(公告)日：2003-01-29
申请号：EP02254994.3
申请日：2002-07-16
申请人： AT&T Corp.
发明人： Alshawi, Hiyan , Di Fabrizio, Giuseppe , Gupta, Nagendra K. , Rahim, Mazin G. , Schapire, Robert Elias , Yoram, Singer
IPC分类号： G10L15/06
CPC分类号： G10L15/063
摘要： A system for understanding entries, such as speech, develops a classifier by employing prior knowledge with which a given corpus of training entries is enlarged threefold. The prior knowledge is embodied in a rule, combined from separate rules created for each label outputted by the classifier, each of which includes a weight measure p ( x ). A first a set of created entries for increasing the corpus of training entries is created by attaching all labels to each entry of the original corpus of training entries, with a weight h p ( x ), or h(1- p ( x )), in association with each label that meets, or fails to meet, the condition specified for the label, h being a preselected positive number. The second set of is created by not attaching any of the labels to each of the original corpus of training entries, with a weight of h(1- p ( x )), or h p ( x ), in association with each label that meets, or fails to meet, the condition specified for the label.
摘要翻译：用于理解条目（例如语音）的系统通过使用预先知识来开发分类器，给定的训练条目语料库被放大三倍。现有知识体现在一个规则中，从为分类器输出的每个标签创建的单独规则组合起来，每个标签都包括权重度p（x）。通过将所有标签附加到训练条目的原始语料库的每个条目，具有权重hp（x）或h（1-p（x）），创建用于增加训练条目语料库的第一组创建条目。与符合或不符合标签规定的条件的每个标签相关联，h是预选的正数。第二组是通过不将任何标签附加到每个训练条目的原始语料库中，其重量与h（1-p（x））或hp（x）的重量相关联，并与每个符合的标签相关联，或者不符合标签规定的条件。

6. 发明公开

EP0920173A2 Enhanced telecommunications network 审中-公开
标题翻译： Verbointes Telekommunikationsnetzwerk
公开(公告)号：EP0920173A2
公开(公告)日：1999-06-02
申请号：EP98309153.9
申请日：1998-11-09
申请人： AT&T Corp.
发明人： Narayanan, Shrikanth S. , Potamianos, Alexandros , Rahim, Mazin G. , Wilpon, Jay G. , Zeljkovic, Ilija
IPC分类号： H04M3/42 , H04M3/50 , H04M3/40 , G10L3/00
CPC分类号： G10L15/065 , G10L15/142 , G10L15/20 , G10L21/02 , H04M1/271 , H04M3/40 , H04M3/42059 , H04M3/42204 , H04M2201/40 , H04M2242/22
摘要： A system deployed in a telecommunications network includes a terminal device (12), a telephone network switch (10), a network service, and an optional network database. The terminal device (12) measures, stores, and transmits to the network service data characterizing the terminal device (12), an acoustic environment, and optionally data characterizing the speech of a local population. These data can be stored in the network database facilitating long-term adaptation of the network service.
摘要翻译：部署在电信网络中的系统包括终端设备（12），电话网络交换机（10），网络服务和可选的网络数据库。终端设备（12）测量，存储和发送表征终端设备（12）的网络服务数据，声学环境以及表征当地群体的语音的可选数据。这些数据可以存储在网络数据库中，有助于长期适应网络服务。

7. 发明公开

EP1696421A2 Learning in automatic speech recognition 有权转让
标题翻译： Lernen zur Spracherkennung
公开(公告)号：EP1696421A2
公开(公告)日：2006-08-30
申请号：EP06110328.9
申请日：2006-02-23
申请人： AT&T Corp.
发明人： Hakkani-Tur, Dilek Z. , Rahim, Mazin G. , Tur, Gokhan , Riccardi, Giuseppe
IPC分类号： G10L15/06
CPC分类号： G10L15/063 , G10L15/07 , G10L15/18 , G10L15/26 , G10L2015/0638
摘要： Utterance data that includes at least a small amount of manually transcribed data is provided. Automatic speech recognition is performed on ones of the utterance data not having a corresponding manual transcription to produce automatically transcribed utterances. A model is trained using all of the manually transcribed data and the automatically transcribed utterances. A predetermined number of utterances not having a corresponding manual transcription are intelligently selected and manually transcribed. Ones of the automatically transcribed data as well as ones having a corresponding manual transcription are labeled. In another aspect of the invention, audio data is mined from at least one source, and a language model is trained for call classification from the mined audio data to produce a language model.
摘要翻译：提供了包括至少少量手动转录数据的语音数据。对没有相应的手动转录的话语数据中的一个执行自动语音识别以产生自动转录的话语。使用所有手动转录数据和自动转录的话语训练模型。智能地选择并手动转录预定数量的不具有相应手动转录的话语。自动转录的数据以及具有相应手动转录的数据的标签。在本发明的另一方面，音频数据从至少一个源开始，并且语言模型被训练用于从所开采的音频数据进行呼叫分类以产生语言模型。

8. 发明授权

EP0881625B1 Multiple models integration for multi-environment speech recognition 失效
标题翻译：用于多环境语音识别的多模型集成
公开(公告)号：EP0881625B1
公开(公告)日：2005-08-10
申请号：EP98108805.7
申请日：1998-05-14
申请人： AT&T Corp.
发明人： Rahim, Mazin G.
IPC分类号： G10L15/20
CPC分类号： G10L15/20 , G10L15/144 , G10L21/0216 , G10L2015/228
摘要： A speech recognition system which effectively recognizes unknown speech from multiple acoustic environments includes a set of secondary models, each associated with one or more particular acoustic environments, integrated with a base set of recognition models. The speech recognition system is trained by making a set of secondary models in a first stage of training, and integrating the set of secondary models with a base set of recognition models in a second stage of training.

9. 发明公开

EP0920173A3 Enhanced telecommunications network 审中-公开
标题翻译：提高电信网络
公开(公告)号：EP0920173A3
公开(公告)日：2002-05-08
申请号：EP98309153.9
申请日：1998-11-09
申请人： AT&T Corp.
发明人： Narayanan, Shrikanth S. , Potamianos, Alexandros , Rahim, Mazin G. , Wilpon, Jay G. , Zeljkovic, Ilija
IPC分类号： H04M3/42 , H04M3/50 , H04M3/40 , G10L3/00
CPC分类号： G10L15/065 , G10L15/142 , G10L15/20 , G10L21/02 , H04M1/271 , H04M3/40 , H04M3/42059 , H04M3/42204 , H04M2201/40 , H04M2242/22
摘要： A system deployed in a telecommunications network includes a terminal device (12), a telephone network switch (10), a network service, and an optional network database. The terminal device (12) measures, stores, and transmits to the network service data characterizing the terminal device (12), an acoustic environment, and optionally data characterizing the speech of a local population. These data can be stored in the network database facilitating long-term adaptation of the network service.

10. 发明授权

EP1696421B1 Learning in automatic speech recognition 有权转让
标题翻译：学习语音识别
公开(公告)号：EP1696421B1
公开(公告)日：2008-08-20
申请号：EP06110328.9
申请日：2006-02-23
申请人： AT&T Corp.
发明人： Hakkani-Tur, Dilek Z. , Rahim, Mazin G. , Tur, Gokhan , Riccardi, Giuseppe
IPC分类号： G10L15/06
CPC分类号： G10L15/063 , G10L15/07 , G10L15/18 , G10L15/26 , G10L2015/0638

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式