公开(公告)号：US07991615B2

公开(公告)日：2011-08-02

申请号：US11952267

申请日：2007-12-07

申请人： Xiao Li ,  Asela J. R. Gunawardana ,  Alejandro Acero

发明人： Xiao Li ,  Asela J. R. Gunawardana ,  Alejandro Acero

IPC分类号： G10L15/04

CPC分类号： G10L13/08 ,  G10L15/063 ,  G10L15/187

摘要： Described is the use of acoustic data to improve grapheme-to-phoneme conversion for speech recognition, such as to more accurately recognize spoken names in a voice-dialing system. A joint model of acoustics and graphonemes (acoustic data, phonemes sequences, grapheme sequences and an alignment between phoneme sequences and grapheme sequences) is described, as is retraining by maximum likelihood training and discriminative training in adapting graphoneme model parameters using acoustic data. Also described is the unsupervised collection of grapheme labels for received acoustic data, thereby automatically obtaining a substantial number of actual samples that may be used in retraining. Speech input that does not meet a confidence threshold may be filtered out so as to not be used by the retrained model.

摘要翻译： 描述了使用声学数据来改进用于语音识别的字形到音素转换，例如更准确地识别语音拨号系统中的语音名称。 描述了声学和图形（声学数据，音素序列，字形序列以及音素序列和图形序列之间的对齐）的联合模型，正如通过使用声学数据适应图形模型参数的最大似然训练和鉴别训练来重新训练。 还描述了用于接收的声学数据的无监督的字母标签集合，从而自动获得可用于再培训的大量实际样本。 不满足置信阈值的语音输入可以被滤除，以便不被再培训的模型使用。

公开(公告)号：US20100268725A1

公开(公告)日：2010-10-21

申请号：US12426370

申请日：2009-04-20

申请人： Ye-Yi Wang ,  Xiao Li ,  Raphael D. Hoffmann

发明人： Ye-Yi Wang ,  Xiao Li ,  Raphael D. Hoffmann

IPC分类号： G06F17/30 ,  G06F17/27

CPC分类号： G06F17/30672 ,  G06F17/2735 ,  G06F17/2785

摘要： A user's search experience may be enhanced by providing additional content based upon an understanding of the user's intent. Query tagging, the assigning of semantic labels to terms within a query, is one technique that may be utilized to determine the context of a user's search query. Accordingly, as provided herein, a query tagging model may be updated using one or more stratified lexicons. A list data structure (e.g., lists of phrases obtained from web pages) and seed distribution data (e.g., pre-labeled probability data) may be used by a graph learning technique to obtain an expanded set of phrases and their respective probabilities of corresponding with particular lexicons (e.g., semantic class lexicons). The expanded set of phrases may be used to group phrases into stratified lexicons. The stratified lexicons may be used as features for updating and/or executing the query tagging model.

摘要翻译： 可以通过基于对用户意图的理解来提供附加内容来增强用户的搜索体验。 查询标记，将语义标签分配给查询中的术语，是可用于确定用户搜索查询的上下文的一种技术。 因此，如本文所提供的，可以使用一个或多个分层词典来更新查询标签模型。 列表数据结构（例如，从网页获得的短语的列表）和种子分布数据（例如，预先标记的概率数据）可以通过图形学习技术来使用以获得扩展的短语集合及其相应的概率 特定词典（例如语义类词典）。 扩展的短语组可用于将短语分组成分层词典。 分层词汇可以用作更新和/或执行查询标签模型的特征。

公开(公告)号：US20100256969A1

公开(公告)日：2010-10-07

申请号：US12419336

申请日：2009-04-07

申请人： Xiao Li ,  Ye-Yi Wang

发明人： Xiao Li ,  Ye-Yi Wang

IPC分类号： G06F17/50

CPC分类号： G06K9/6296

摘要： A training module is described for training a conditional random field (CRF) tagging model. The training module trains the tagging model based on an explicitly-labeled training set and an implicitly-labeled training set. The explicitly-labeled training set includes explicit labels that are manually selected via human annotation, while the implicitly-labeled training set includes implicit labels that are generated in an unsupervised manner. In one approach, the training module can train the tagging model by treating the implicit labels as non-binding evidence that has a bearing on values of hidden state sequence variables. In another approach, the training module can treat the implicit labels as binding or hard evidence. A labeling system is also described for providing the implicit labels.

摘要翻译： 描述训练模块用于训练条件随机场（CRF）标记模型。 培训模块基于明确标记的训练集和隐式标记的训练集来训练标记模型。 明确标记的训练集包括通过人工注释手动选择的显式标签，而隐含标记的训练集包括以无监督方式生成的隐式标签。 在一种方法中，训练模块可以通过将隐式标签视为与隐含状态序列变量的值有关的非约束证据来训练标记模型。 在另一种方法中，培训模块可以将隐含的标签视为具有约束力或硬性证据。 还描述了用于提供隐式标签的标签系统。

公开(公告)号：US20100076765A1

公开(公告)日：2010-03-25

申请号：US12233826

申请日：2008-09-19

申请人： Geoffrey G. Zweig ,  Xiao Li ,  Dan Bohus ,  Alejandro Acero ,  Eric J. Horvitz

发明人： Geoffrey G. Zweig ,  Xiao Li ,  Dan Bohus ,  Alejandro Acero ,  Eric J. Horvitz

IPC分类号： G10L15/00

CPC分类号： G10L15/1822

摘要： Described is a technology by which a structured model of repetition is used to determine the words spoken by a user, and/or a corresponding database entry, based in part on a prior utterance. For a repeated utterance, a joint probability analysis is performed on (at least some of) the corresponding word sequences as recognized by one or more recognizers) and associated acoustic data. For example, a generative probabilistic model, or a maximum entropy model may be used in the analysis. The second utterance may be a repetition of the first utterance using the exact words, or another structural transformation thereof relative to the first utterance, such as an extension that adds one or more words, a truncation that removes one or more words, or a whole or partial spelling of one or more words.

摘要翻译： 描述了一种技术，通过该技术，部分地基于先前的话语，使用结构化重复模型来确定用户说出的单词和/或相应的数据库条目。 对于重复的话语，对由一个或多个识别器识别的相应字序列（和至少一些）和相关联的声学数据进行联合概率分析。 例如，可以在分析中使用生成概率模型或最大熵模型。 第二个发音可以是使用精确的单词或相对于第一个发音的其他结构变换的第一个发音的重复，例如添加一个或多个单词的扩展，删除一个或多个单词的截断或整个 或一个或多个单词的部分拼写。

公开(公告)号：US07183181B2

公开(公告)日：2007-02-27

申请号：US10950839

申请日：2004-09-27

申请人： Xiao Li ,  Roger Y. B. Young ,  Bruce J. Whitefield

发明人： Xiao Li ,  Roger Y. B. Young ,  Bruce J. Whitefield

IPC分类号： H01L21/20 ,  H01L21/36

CPC分类号： H01L21/02087 ,  H01L21/6708

摘要： A method of removing an edge bead of a coated material on a substrate. The substrate is rotated, and a fluid that solvates the coated material is delivered. The delivery of the fluid is directed radially inward on the substrate at a rate of between about three millimeters per second and about twenty millimeters per second until a desired innermost fluid delivery position on the substrate is attained. Immediately upon attaining the desired innermost fluid delivery position on the substrate, the delivery of the fluid is directed radially outward off the substrate at a rate of more than zero millimeters per second and less than about four millimeters per second. The rotation of the substrate is ceased.

摘要翻译： 一种去除衬底上涂覆材料的边缘珠的方法。 旋转衬底，并且输送溶剂化涂覆材料的流体。 流体的输送以约3毫米每秒和约20毫米/秒的速率径向向内指向基底上，直到达到基底上所需的最内流体输送位置。 在基板上达到理想的最内流体输送位置之后，流体的输送立即以大于零毫米/秒的速度从基板径向向外指向，并且小于约4毫米/秒。 衬底的旋转停止。

公开(公告)号：US20050110417A1

公开(公告)日：2005-05-26

申请号：US10926525

申请日：2004-08-26

申请人： Xiao Li ,  Stephen Shapiro ,  James Bonnema ,  Wed Wang

发明人： Xiao Li ,  Stephen Shapiro ,  James Bonnema ,  Wed Wang

IPC分类号： F21S13/12 ,  F21S19/00 ,  F21V23/04 ,  F21V35/00 ,  H05B39/04

CPC分类号： F21S6/001 ,  F21S13/12 ,  F21S19/00 ,  F21V23/0442 ,  F21V35/00 ,  F21Y2113/20 ,  F21Y2115/10

摘要： A candle with a decorative light element is provided. A light emitting device is disposed within the main body of the candle, and a sensor connected to the light emitting device detects the presence of a flame on the wick. When a flame is detected by the sensor, the light emitting device is activated. The portion of the candle from which the wick projects is preferably substantially opaque, and the portion of the candle in which the light emitting device is disposed is preferably at least partially light transmissible or translucent, so that the glow of the flame does not interfere with the glow of the light emitting device within the candle.

摘要翻译： 提供带有装饰灯元素的蜡烛。 发光装置设置在蜡烛的主体内，连接到发光装置的传感器检测到灯芯上存在火焰。 当传感器检测到火焰时，发光装置被激活。 蜡烛投射的蜡烛的部分优选地基本上是不透明的，并且其中设置有发光装置的蜡烛的部分优选地至少部分地是透光的或半透明的，使得火焰的辉光不会干扰 蜡烛内的发光装置的辉光。

公开(公告)号：US09998909B2

公开(公告)日：2018-06-12

申请号：US14126471

申请日：2011-06-22

申请人： Lasse Olsson ,  Juying Gan ,  Xiao Li

发明人： Lasse Olsson ,  Juying Gan ,  Xiao Li

IPC分类号： H04W4/00 ,  H04W8/08 ,  H04W76/02 ,  H04W8/18

CPC分类号： H04W8/08 ,  H04W8/18 ,  H04W76/12

摘要： Example embodiments presented herein are directed towards a system, and corresponding method, for determining 3rd Generation Direct Tunnel (3GDT) usage based on the payload usage of a user.

公开(公告)号：US20170191110A1

公开(公告)日：2017-07-06

申请号：US15313683

申请日：2015-05-26

申请人： Xiao LI ,  Yan ZHANG ,  Jianwei LIU ,  BECTON DICKINSON HOLDINGS PTE LTD.

发明人： Xiao Li ,  Yan Zhang ,  Jianwei Liu

IPC分类号： C12Q1/04 ,  G01N33/02 ,  G01N33/00 ,  G01N33/04

CPC分类号： C12Q1/04 ,  G01N33/0036 ,  G01N33/004 ,  G01N33/02 ,  G01N33/04

摘要： A system that indicates the presence or absence of microorganisms in fluid food products. The system has a bottle for receiving sample to be tested. The bottle has a sensor that will monitor and detect changes in at least one sample parameter, but no additives that contain nutrients that support microbial growth. The bottle is placed in an incubator and the sensor in the bottle is monitored for changes. The incubator is programed so that, if the sensor detects that the value of the monitored parameter has reached a certain value, then the sample is determined to be positive for microbial growth.

公开(公告)号：US09336299B2

公开(公告)日：2016-05-10

申请号：US12426370

申请日：2009-04-20

申请人： Ye-Yi Wang ,  Xiao Li ,  Raphael Dominik Hoffmann

发明人： Ye-Yi Wang ,  Xiao Li ,  Raphael Dominik Hoffmann

IPC分类号： G06F17/30 ,  G06F17/27

CPC分类号： G06F17/30672 ,  G06F17/2735 ,  G06F17/2785

摘要： A user's search experience may be enhanced by providing additional content based upon an understanding of the user's intent. Query tagging, the assigning of semantic labels to terms within a query, is one technique that may be utilized to determine the context of a user's search query. Accordingly, as provided herein, a query tagging model may be updated using one or more stratified lexicons. A list data structure (e.g., lists of phrases obtained from web pages) and seed distribution data (e.g., pre-labeled probability data) may be used by a graph learning technique to obtain an expanded set of phrases and their respective probabilities of corresponding with particular lexicons (e.g., semantic class lexicons). The expanded set of phrases may be used to group phrases into stratified lexicons. The stratified lexicons may be used as features for updating and/or executing the query tagging model.

摘要翻译： 可以通过基于对用户意图的理解来提供附加内容来增强用户的搜索体验。 查询标记，将语义标签分配给查询中的术语，是可用于确定用户搜索查询的上下文的一种技术。 因此，如本文所提供的，可以使用一个或多个分层词典来更新查询标签模型。 列表数据结构（例如，从网页获得的短语的列表）和种子分布数据（例如，预先标记的概率数据）可以通过图形学习技术来使用以获得扩展的短语集合及其相应的概率 特定词典（例如语义类词典）。 扩展的短语组可用于将短语分组成分层词典。 分层词汇可以用作更新和/或执行查询标签模型的特征。

公开(公告)号：US09269646B2

公开(公告)日：2016-02-23

申请号：US13613235

申请日：2012-09-13

申请人： Shijian Luo ,  Xiao Li ,  Jian Li

发明人： Shijian Luo ,  Xiao Li ,  Jian Li

IPC分类号： H01L23/52 ,  H01L23/36 ,  H01L23/367 ,  H01L23/373 ,  H01L23/42 ,  H01L25/065 ,  H01L25/18

CPC分类号： H01L23/36 ,  H01L23/3675 ,  H01L23/3677 ,  H01L23/3736 ,  H01L23/42 ,  H01L25/0657 ,  H01L25/18 ,  H01L25/50 ,  H01L2224/16145 ,  H01L2224/16227 ,  H01L2224/17181 ,  H01L2224/73204 ,  H01L2225/06513 ,  H01L2225/06544 ,  H01L2225/06589 ,  H01L2924/15311

摘要： A semiconductor die assembly comprises a plurality of semiconductor dice in a stack. Another semiconductor die is adjacent to the stack and has a region, which may comprise a relatively higher power density region, extends peripherally beyond the stack. Conductive elements extend between and electrically interconnect integrated circuits of semiconductor dice in the stack and of the other semiconductor die. Thermal pillars are interposed between semiconductor dice of the stack, and a heat dissipation structure, such as a lid, is in contact with an uppermost die of the stack and the higher power density region of the other semiconductor die. Other die assemblies, semiconductor devices and methods of managing heat transfer within a semiconductor die assembly are also disclosed.

摘要翻译： 半导体管芯组件包括堆叠中的多个半导体管芯。 另一个半导体管芯与堆叠相邻并且具有可以包括相对较高功率密度区域的区域，其周边延伸超过堆叠。 导电元件在堆叠和另一个半导体管芯中的半导体管芯的集成电路之间延伸并电互连。 散热柱插入在堆叠的半导体管芯之间，散热结构（例如盖子）与堆叠的最上面的管芯和另一个半导体管芯的较高功率密度区域接触。 还公开了其他管芯组件，半导体器件和管理半导体管芯组件内的热传递的方法。