利索-全球专利检索

  1. 登录
  2. 注册

搜索结果

39 条记录 (耗时 0.029 秒)

    Efficient gradient computation for conditional Gaussian graphical models
    2.
    发明申请
    Efficient gradient computation for conditional Gaussian graphical models 有权
    条件高斯图形模型的有效梯度计算

    公开(公告)号：US20080010043A1

    公开(公告)日：2008-01-10

    申请号：US11005148

    申请日：2004-12-06

    申请人： Bo Thiesson ,  Christopher Meek

    发明人： Bo Thiesson ,  Christopher Meek

    IPC分类号： G06F17/10

    CPC分类号： G06K9/6296

    摘要： The subject invention leverages standard probabilistic inference techniques to determine a log-likelihood for a conditional Gaussian graphical model of a data set with at least one continuous variable and with data not observed for at least one of the variables. This provides an efficient means to compute gradients for CG models with continuous variables and incomplete data observations. The subject invention allows gradient-based optimization processes to employ gradients to iteratively adapt parameters of models in order to improve incomplete data log-likelihoods and identify maximum likelihood estimates (MLE) and/or local maxima of the incomplete data log-likelihoods. Conditional Gaussian local gradients along with conditional multinomial local gradients determined by the subject invention can be utilized to facilitate in providing parameter gradients for full conditional Gaussian models.

    摘要翻译： 本发明利用标准概率推理技术来确定具有至少一个连续变量的数据集的条件高斯图形模型的对数似然，并且对于至少一个变量未观察到数据。 这为用于连续变量和不完整数据观察的CG模型计算梯度提供了有效手段。 本发明允许基于梯度的优化过程使用梯度来迭代地适应模型的参数，以便改进不完整的数据对数似然性并且识别不完全数据对数似然性的最大似然估计（MLE）和/或局部最大值。 条件高斯局部梯度以及由本发明确定的条件多项式局部梯度可以用于促进为全条件高斯模型提供参数梯度。

    Publisher unions
    3.
    发明申请
    Publisher unions 有权
    出版社联合会

    公开(公告)号：US20070260617A1

    公开(公告)日：2007-11-08

    申请号：US11418899

    申请日：2006-05-05

    申请人： Ewa Dominowska ,  Christopher Meek ,  David Chickering ,  Jody Biggs ,  Brian Burdick

    发明人： Ewa Dominowska ,  Christopher Meek ,  David Chickering ,  Jody Biggs ,  Brian Burdick

    IPC分类号： G06F7/00

    CPC分类号： G06Q30/02 ,  G06Q10/00 ,  G06Q10/10 ,  G06Q30/08

    摘要： A publisher union comprises a plurality of publishers, a channel, and a publisher union administrator. The publisher union is administered by receiving a channel proposal, determining whether the channel proposal is acceptable, forming a channel, and presenting the channel for monetization. User information is gathered by the publisher union by establishing a domain, collecting user information, aggregating the user information, and providing the aggregated user information to publisher union members.

    摘要翻译： 发行商联盟由多个发布商，渠道和发布商联盟管理员组成。 发行商联盟通过接收渠道提议，确定渠道提案是否可以接受，形成渠道，以及呈现营利渠道来进行管理。 用户信息由发行商联合会通过建立域，收集用户信息，聚合用户信息以及向发布商联盟成员提供聚合的用户信息来收集。

    Trees of classifiers for detecting email spam
    4.
    发明申请
    Trees of classifiers for detecting email spam 有权
    用于检测电子邮件垃圾邮件的分类树

    公开(公告)号：US20070038705A1

    公开(公告)日：2007-02-15

    申请号：US11193691

    申请日：2005-07-29

    申请人： David Chickering ,  Geoffrey Hulten ,  Robert Rounthwaite ,  Christopher Meek ,  David Heckerman ,  Joshua Goodman

    发明人： David Chickering ,  Geoffrey Hulten ,  Robert Rounthwaite ,  Christopher Meek ,  David Heckerman ,  Joshua Goodman

    IPC分类号： G06F15/16

    CPC分类号： H04L51/12

    摘要： Decision trees populated with classifier models are leveraged to provide enhanced spam detection utilizing separate email classifiers for each feature of an email. This provides a higher probability of spam detection through tailoring of each classifier model to facilitate in more accurately determining spam on a feature-by-feature basis. Classifiers can be constructed based on linear models such as, for example, logistic-regression models and/or support vector machines (SVM) and the like. The classifiers can also be constructed based on decision trees. “Compound features” based on internal and/or external nodes of a decision tree can be utilized to provide linear classifier models as well. Smoothing of the spam detection results can be achieved by utilizing classifier models from other nodes within the decision tree if training data is sparse. This forms a base model for branches of a decision tree that may not have received substantial training data.

    摘要翻译： 利用分类器模型填充的决策树利用电子邮件的每个功能使用单独的电子邮件分类器来提供增强的垃圾邮件检测。 这通过定制每个分类器模型提供了更高的垃圾邮件检测的概率，以便于在逐个特征的基础上更准确地确定垃圾邮件。 分类器可以基于诸如逻辑回归模型和/或支持向量机（SVM）等线性模型来构建。 分类器也可以基于决策树构建。 基于决策树的内部和/或外部节点的“复合特征”也可以用于提供线性分类器模型。 垃圾邮件检测结果的平滑可以通过使用来自决策树内的其他节点的分类器模型来实现，如果训练数据是稀疏的。 这形成了可能没有接收到大量训练数据的决策树的分支的基本模型。

    Viral advertising for interactive services
    5.
    发明申请
    Viral advertising for interactive services 有权
    用于互动服务的病毒广告

    公开(公告)号：US20060218577A1

    公开(公告)日：2006-09-28

    申请号：US11078555

    申请日：2005-03-11

    申请人： Joshua Goodman ,  Christopher Meek ,  Carl Kadie

    发明人： Joshua Goodman ,  Christopher Meek ,  Carl Kadie

    IPC分类号： H04N7/025 ,  H04N7/10

    CPC分类号： H04N7/173 ,  G06Q30/02 ,  H04N21/2547 ,  H04N21/4143 ,  H04N21/4788 ,  H04N21/812

    摘要： The subject invention provides a unique system and method that facilitates propagating selected advertisements among users of interactive services. Interactive service users can be targeted for specific types of advertisements for particular products or services. When a user selects at least one advertisement for more detailed viewing, the advertisement can be distributed to or shared with one or more other users. These other users may be part of the original user's social network. Thus user-selected advertisements can be shared among users who are familiar with each other's current or future interests. In some cases, user-selected advertisements can replace system-selected advertisements. As a result, advertisers can benefit from increased exposure of and interest in their advertisements.

    摘要翻译： 本发明提供了一种有助于在交互式服务的用户之间传播所选广告的独特系统和方法。 互动服务用户可以针对特定产品或服务的特定类型的广告。 当用户选择至少一个广告以进行更详细的观看时，该广告可以被分发给一个或多个其他用户或与一个或多个其他用户共享。 这些其他用户可能是原始用户社交网络的一部分。 因此，用户选择的广告可以在熟悉彼此的当前或未来兴趣的用户之间共享。 在某些情况下，用户选择的广告可以替代系统选择的广告。 因此，广告商可以从广告的曝光和兴趣中获益。

    Determining near-optimal block size for incremental-type expectation maximization (EM) algrorithms
    7.
    发明申请
    Determining near-optimal block size for incremental-type expectation maximization (EM) algrorithms 有权
    确定增量型期望最大化（EM）算法的近似最优块大小

    公开(公告)号：US20050267717A1

    公开(公告)日：2005-12-01

    申请号：US11177734

    申请日：2005-07-08

    申请人： Bo Thiesson ,  Christopher Meek ,  David Heckerman

    发明人： Bo Thiesson ,  Christopher Meek ,  David Heckerman

    IPC分类号： G06F17/10 ,  G06F17/18

    CPC分类号： G06K9/6226 ,  G06F17/10 ,  G06F17/18 ,  Y10S707/99942

    摘要： Determining the near-optimal block size for incremental-type expectation maximization (EM) algorithms is disclosed. Block size is determined based on the novel insight that the speed increase resulting from using an incremental-type EM algorithm as opposed to the standard EM algorithm is roughly the same for a given range of block sizes. Furthermore, this block size can be determined by an initial version of the EM algorithm that does not reach convergence. For a current block size, the speed increase is determined, and if the speed increase is the greatest determined so far, the current block size is set as the target block size. This process is repeated for new block sizes, until no new block sizes can be determined.

    摘要翻译： 公开了确定增量型期望最大化（EM）算法的近似最小块大小。 基于新的认识来确定块大小，即对于给定的块大小范围，使用增量型EM算法而不是标准EM算法导致的速度增加大致相同。 此外，该块大小可以由未达到收敛的EM算法的初始版本来确定。 对于当前块大小，确定速度增加，并且如果到目前为止确定的速度增加最大，则将当前块大小设置为目标块大小。 对于新的块大小重复此过程，直到不能确定新的块大小。

    Decision-theoretic web-crawling and predicting web-page change
    8.
    发明申请
    Decision-theoretic web-crawling and predicting web-page change 失效
    决策理论网络爬行和预测网页更改

    公开(公告)号：US20050192936A1

    公开(公告)日：2005-09-01

    申请号：US10777365

    申请日：2004-02-12

    申请人： Christopher Meek ,  Carl Kadie

    发明人： Christopher Meek ,  Carl Kadie

    IPC分类号： G06F13/00 ,  G06F17/30 ,  G06F7/00

    CPC分类号： G06F17/30864 ,  Y10S707/99932 ,  Y10S707/99933

    摘要： Systems and methods are described that facilitate predictive web-crawling in a computer environment. Aspects of the invention provide for predictive, utility-based, and decision theoretic probability assessments of changes in subsets of web pages, enhancing web-crawling ability and ensuring that web page information is maintained in a fresh state. Additionally, the invention facilitates selective crawling of pages with a high probability of change.

    摘要翻译： 描述了在计算机环境中促进预测性网络爬行的系统和方法。 本发明的方面提供了对网页子集的改变的预测，实用性和决策理论概率评估，增强了网络爬行能力，并确保网页信息保持在新鲜状态。 此外，本发明有助于选择性地爬行具有高变化概率的页面。

    Automatically generating content for presenting in a preview pane for ADS
    9.
    发明申请
    Automatically generating content for presenting in a preview pane for ADS 审中-公开
    自动生成用于在ADS的预览窗格中显示的内容

    公开(公告)号：US20070050253A1

    公开(公告)日：2007-03-01

    申请号：US11214485

    申请日：2005-08-29

    申请人： Jody Biggs ,  Christian Borgs ,  Jennifer Chayes ,  Uriel Feige ,  Kamal Jain ,  Ying Li ,  Christopher Meek ,  Tarek Najm ,  Joshua Goodman

    发明人： Jody Biggs ,  Christian Borgs ,  Jennifer Chayes ,  Uriel Feige ,  Kamal Jain ,  Ying Li ,  Christopher Meek ,  Tarek Najm ,  Joshua Goodman

    IPC分类号： G06Q30/00

    CPC分类号： G06Q30/02 ,  G06Q30/0271

    摘要： The user interfaces, methods and systems described herein facilitate user interaction with an ad space by conveying additional advertising content via a preview pane and facilitate automatically generating the content of the preview pane. By way of example, an electronic advertisement is conveyed to a user in an ad space provided by a third party, and a secondary advertisement generating component automatically generates at least part of the content of a secondary advertisement. The secondary advertisement provides content associated with the electronic advertisement and occurs upon receiving a user indication. A context acquiring component also may provide context information to the secondary advertisement generating component to automatically generate at least part of the content of the secondary advertisement. By way of another example, a user is provided with one or more ads from a plurality of different advertisers in a first ad space maintained by an ad space supplier. A user input identifying at least one of the ads from the plurality of different advertisers is received. A second ad space for a supplemental ad having supplemental advertising information relating to the at least one ad identified by the user input is provided. At least part of the supplemental advertising information supplied in the supplemental ad is automatically produced. Contextual information also may be employed to automatically produce at least part of the supplemental advertising information.

    摘要翻译： 这里描述的用户界面，方法和系统通过经由预览窗格传送额外的广告内容来促进用户与广告空间的交互，并促进自动生成预览窗格的内容。 作为示例，在由第三方提供的广告空间中向用户传送电子广告，并且辅助广告生成部件自动生成次要广告的内容的至少一部分。 次要广告提供与电子广告相关联的内容，并且在接收到用户指示时发生。 上下文获取组件还可以向次要广告生成组件提供上下文信息以自动生成辅助广告的至少部分内容。 作为另一示例，在由广告空间供应商维护的第一广告空间中向用户提供来自多个不同广告商的一个或多个广告。 接收识别来自多个不同广告商的广告中的至少一个的用户输入。 提供了补充广告的第二广告空间，其具有与由用户输入标识的至少一个广告相关的补充广告信息。 补充广告中提供的补充广告信息的至少一部分将自动生成。 还可以使用上下文信息来自动产生补充广告信息的至少一部分。