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. 发明授权

US08589319B2 Adaptive pairwise preferences in recommenders 有权
标题翻译：推荐者的自适应成对偏好
公开(公告)号：US08589319B2
公开(公告)日：2013-11-19
申请号：US12958434
申请日：2010-12-02
申请人： Suhrid Balakrishnan , Sumit Chopra
发明人： Suhrid Balakrishnan , Sumit Chopra
IPC分类号： G06F15/18
CPC分类号： G06N5/027 , G06F17/18 , G06F17/30979 , G06N7/005 , G06N99/005 , G06Q30/0203 , G06Q30/0631
摘要： Methods, systems, and products adapt recommender systems with pairwise feedback. A pairwise question is posed to a user. A response is received that selects a preference for a pair of items in the pairwise question. A latent factor model is adapted to incorporate the response, and an item is recommended to the user based on the response.
摘要翻译：方法，系统和产品适应具有成对反馈的推荐系统。一个成对的问题是向用户提出的。收到一个响应，选择对成对问题中的一对项目的偏好。潜在因素模型适用于纳入响应，并且基于响应向用户推荐项目。

2. 发明授权

US09092425B2 System and method for feature-rich continuous space language models 有权
标题翻译：功能丰富的连续空间语言模型的系统和方法
公开(公告)号：US09092425B2
公开(公告)日：2015-07-28
申请号：US12963161
申请日：2010-12-08
申请人： Piotr Wojciech Mirowski , Srinivas Bangalore , Suhrid Balakrishnan , Sumit Chopra
发明人： Piotr Wojciech Mirowski , Srinivas Bangalore , Suhrid Balakrishnan , Sumit Chopra
IPC分类号： G06F17/27 , G06F17/28
CPC分类号： G06F17/28
摘要： Disclosed herein are systems, methods, and non-transitory computer-readable storage media for predicting probabilities of words for a language model. An exemplary system configured to practice the method receives a sequence of words and external data associated with the sequence of words and maps the sequence of words to an X-dimensional vector, corresponding to a vocabulary size. Then the system processes each X-dimensional vector, based on the external data, to generate respective Y-dimensional vectors, wherein each Y-dimensional vector represents a dense continuous space, and outputs at least one next word predicted to follow the sequence of words based on the respective Y-dimensional vectors. The X-dimensional vector, which is a binary sparse representation, can be higher dimensional than the Y-dimensional vector, which is a dense continuous space. The external data can include part-of-speech tags, topic information, word similarity, word relationships, a particular topic, and succeeding parts of speech in a given history.
摘要翻译：这里公开了用于预测语言模型的单词概率的系统，方法和非暂时的计算机可读存储介质。配置为实施该方法的示例性系统接收与该单词序列相关联的单词序列和外部数据序列，并将该单词序列映射到对应于词汇大小的X维向量。然后系统根据外部数据对每个X维向量进行处理，以产生各自的Y维向量，其中每个Y维向量表示密集的连续空间，并且输出至少一个预测的下一个单词以跟随单词序列基于相应的Y维向量。作为二进制稀疏表示的X维向量可以比作为密集连续空间的Y维向量更高的维度。外部数据可以包括在给定历史中的部分词汇标签，主题信息，单词相似性，单词关系，特定主题以及后续部分语音。

3. 发明申请

US20120143802A1 Adaptive Pairwise Preferences in Recommenders 有权
标题翻译：推荐者中的自适应成对偏好
公开(公告)号：US20120143802A1
公开(公告)日：2012-06-07
申请号：US12958434
申请日：2010-12-02
申请人： Suhrid Balakrishnan , Sumit Chopra
发明人： Suhrid Balakrishnan , Sumit Chopra
IPC分类号： G06F15/18 , G06N5/02
CPC分类号： G06N5/027 , G06F17/18 , G06F17/30979 , G06N7/005 , G06N99/005 , G06Q30/0203 , G06Q30/0631
摘要： Methods, systems, and products adapt recommender systems with pairwise feedback. A pairwise question is posed to a user. A response is received that selects a preference for a pair of items in the pairwise question. A latent factor model is adapted to incorporate the response, and an item is recommended to the user based on the response.
摘要翻译：方法，系统和产品适应具有成对反馈的推荐系统。一个成对的问题是向用户提出的。收到一个响应，选择对成对问题中的一对项目的偏好。潜在因素模型适用于纳入响应，并且基于响应向用户推荐项目。

4. 发明授权

US09653066B2 System and method for estimating the reliability of alternate speech recognition hypotheses in real time 有权
公开(公告)号：US09653066B2
公开(公告)日：2017-05-16
申请号：US12604650
申请日：2009-10-23
申请人： Jason Williams , Suhrid Balakrishnan
发明人： Jason Williams , Suhrid Balakrishnan
IPC分类号： G10L15/01 , G10L15/08
CPC分类号： G10L15/01 , G10L15/04 , G10L15/08 , G10L15/083 , G10L15/14 , G10L15/22 , G10L15/28
摘要： Disclosed herein are systems, methods, and computer-readable storage media for estimating reliability of alternate speech recognition hypotheses. A system configured to practice the method receives an N-best list of speech recognition hypotheses and features describing the N-best list, determines a first probability of correctness for each hypothesis in the N-best list based on the received features, determines a second probability that the N-best list does not contain a correct hypothesis, and uses the first probability and the second probability in a spoken dialog. The features can describe properties of at least one of a lattice, a word confusion network, and a garbage model. In one aspect, the N-best lists are not reordered according to reranking scores. The determination of the first probability of correctness can include a first stage of training a probabilistic model and a second stage of distributing mass over items in a tail of the N-best list.

5. 发明授权

US08433578B2 System and method for automatically generating a dialog manager 有权
标题翻译：自动生成对话管理器的系统和方法
公开(公告)号：US08433578B2
公开(公告)日：2013-04-30
申请号：US12627617
申请日：2009-11-30
申请人： Jason Williams , Suhrid Balakrishnan , Lihong Li
发明人： Jason Williams , Suhrid Balakrishnan , Lihong Li
IPC分类号： G10L21/00 , G10L11/00
CPC分类号： G10L15/063 , G06F3/167 , H04M3/4936 , H04M2203/355
摘要： Disclosed herein are systems, methods, and computer-readable storage media for automatically generating a dialog manager for use in a spoken dialog system. A system practicing the method receives a set of user interactions having features, identifies an initial policy, evaluates all of the features in a linear evaluation step of the algorithm to identify a set of most important features, performs a cubic policy improvement step on the identified set of most important features, repeats the previous two steps one or more times, and generates a dialog manager for use in a spoken dialog system based on the resulting policy and/or set of most important features. Evaluating all of the features can include estimating a weight for each feature which indicates how much each feature contributes to at least one of the identified policies. The system can ignore features not in the set of most important features.
摘要翻译：这里公开了用于自动生成用于在口头对话系统中使用的对话管理器的系统，方法和计算机可读存储介质。实施该方法的系统接收具有特征的一组用户交互，识别初始策略，评估算法的线性评估步骤中的所有特征以识别一组最重要的特征，对所识别的一个或多个特征进行立体策略改进步骤一组最重要的特征，重复前一个步骤一次或多次，并且基于所得到的策略和/或一组最重要的特征生成用于在口头对话系统中的对话管理器。评估所有特征可以包括估计每个特征的权重，其指示每个特征对至少一个所识别的策略贡献多少。系统可以忽略不属于最重要功能集的功能。

6. 发明申请

US20120150532A1 SYSTEM AND METHOD FOR FEATURE-RICH CONTINUOUS SPACE LANGUAGE MODELS 有权
标题翻译：特征丰富的连续空间语言模型的系统与方法
公开(公告)号：US20120150532A1
公开(公告)日：2012-06-14
申请号：US12963161
申请日：2010-12-08
申请人： Piotr Wojciech Mirowski , Srinivas Banglore , Suhrid Balakrishnan , Sumit Chopra
发明人： Piotr Wojciech Mirowski , Srinivas Banglore , Suhrid Balakrishnan , Sumit Chopra
IPC分类号： G06F17/27
CPC分类号： G06F17/28
摘要： Disclosed herein are systems, methods, and non-transitory computer-readable storage media for predicting probabilities of words for a language model. An exemplary system configured to practice the method receives a sequence of words and external data associated with the sequence of words and maps the sequence of words to an X-dimensional vector, corresponding to a vocabulary size. Then the system processes each X-dimensional vector, based on the external data, to generate respective Y-dimensional vectors, wherein each Y-dimensional vector represents a dense continuous space, and outputs at least one next word predicted to follow the sequence of words based on the respective Y-dimensional vectors. The X-dimensional vector, which is a binary sparse representation, can be higher dimensional than the Y-dimensional vector, which is a dense continuous space. The external data can include part-of-speech tags, topic information, word similarity, word relationships, a particular topic, and succeeding parts of speech in a given history.
摘要翻译：这里公开了用于预测语言模型的单词概率的系统，方法和非暂时的计算机可读存储介质。配置为实施该方法的示例性系统接收与该单词序列相关联的单词序列和外部数据序列，并将该单词序列映射到对应于词汇大小的X维向量。然后系统根据外部数据对每个X维向量进行处理，以产生各自的Y维向量，其中每个Y维向量表示密集的连续空间，并且输出至少一个预测的下一个单词以跟随单词序列基于相应的Y维向量。作为二进制稀疏表示的X维向量可以比作为密集连续空间的Y维向量更高的维度。外部数据可以包括在给定历史中的部分词汇标签，主题信息，单词相似性，单词关系，特定主题以及后续部分语音。

7. 发明申请

US20110131048A1 SYSTEM AND METHOD FOR AUTOMATICALLY GENERATING A DIALOG MANAGER 有权
标题翻译：用于自动生成对话管理器的系统和方法
公开(公告)号：US20110131048A1
公开(公告)日：2011-06-02
申请号：US12627617
申请日：2009-11-30
申请人： Jason Williams , Suhrid Balakrishnan , Lihong Li
发明人： Jason Williams , Suhrid Balakrishnan , Lihong Li
IPC分类号： G10L21/00
CPC分类号： G10L15/063 , G06F3/167 , H04M3/4936 , H04M2203/355
摘要： Disclosed herein are systems, methods, and computer-readable storage media for automatically generating a dialog manager for use in a spoken dialog system. A system practicing the method receives a set of user interactions having features, identifies an initial policy, evaluates all of the features in a linear evaluation step of the algorithm to identify a set of most important features, performs a cubic policy improvement step on the identified set of most important features, repeats the previous two steps one or more times, and generates a dialog manager for use in a spoken dialog system based on the resulting policy and/or set of most important features. Evaluating all of the features can include estimating a weight for each feature which indicates how much each feature contributes to at least one of the identified policies. The system can ignore features not in the set of most important features.
摘要翻译：这里公开了用于自动生成用于在口头对话系统中使用的对话管理器的系统，方法和计算机可读存储介质。实施该方法的系统接收具有特征的一组用户交互，识别初始策略，评估算法的线性评估步骤中的所有特征以识别一组最重要的特征，对所识别的一个或多个特征进行立体策略改进步骤一组最重要的特征，重复前一个步骤一次或多次，并且基于所得到的策略和/或一组最重要的特征生成用于在口头对话系统中的对话管理器。评估所有特征可以包括估计每个特征的权重，其指示每个特征对至少一个所识别的策略贡献多少。系统可以忽略不属于最重要功能集的功能。

8. 发明申请

US20110099012A1 SYSTEM AND METHOD FOR ESTIMATING THE RELIABILITY OF ALTERNATE SPEECH RECOGNITION HYPOTHESES IN REAL TIME 有权
标题翻译：实时评估替代语音识别假设的可靠性的系统和方法
公开(公告)号：US20110099012A1
公开(公告)日：2011-04-28
申请号：US12604650
申请日：2009-10-23
申请人： Jason WILLIAMS , Suhrid BALAKRISHNAN
发明人： Jason WILLIAMS , Suhrid BALAKRISHNAN
IPC分类号： G10L15/00
CPC分类号： G10L15/01 , G10L15/04 , G10L15/08 , G10L15/083 , G10L15/14 , G10L15/22 , G10L15/28
摘要： Disclosed herein are systems, methods, and computer-readable storage media for estimating reliability of alternate speech recognition hypotheses. A system configured to practice the method receives an N-best list of speech recognition hypotheses and features describing the N-best list, determines a first probability of correctness for each hypothesis in the N-best list based on the received features, determines a second probability that the N-best list does not contain a correct hypothesis, and uses the first probability and the second probability in a spoken dialog. The features can describe properties of at least one of a lattice, a word confusion network, and a garbage model. In one aspect, the N-best lists are not reordered according to reranking scores. The determination of the first probability of correctness can include a first stage of training a probabilistic model and a second stage of distributing mass over items in a tail of the N-best list.
摘要翻译：这里公开了用于估计替代语音识别假设的可靠性的系统，方法和计算机可读存储介质。配置为实施该方法的系统接收描述N最佳列表的语音识别假设和特征的N最佳列表，基于接收的特征确定N最佳列表中的每个假设的第一概率的正确性，确定第二 N最佳列表不包含正确假设的概率，并且在口头对话中使用第一概率和第二概率。特征可以描述网格，词混淆网络和垃圾模型中的至少一个的属性。在一方面，N个最佳列表根据排名得分不重新排列。确定正确性的第一个概率可以包括训练概率模型的第一阶段和在N最佳列表的尾部中的项目分发质量的第二阶段。

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