公开(公告)号：US08498950B2

公开(公告)日：2013-07-30

申请号：US12905543

申请日：2010-10-15

申请人： Dragomir Yankov ,  Suju Rajan ,  Adwait Ratnaparkhi ,  Rajesh G. Parekh

发明人： Dragomir Yankov ,  Suju Rajan ,  Adwait Ratnaparkhi ,  Rajesh G. Parekh

IPC分类号： G06F15/18

CPC分类号： G06N99/005

摘要： A system for training classifiers in multiple categories through an active learning system, including a computer having a memory and a processor, the processor programmed to: train an initial set of m binary one-versus-all classifiers, one for each category in a taxonomy, on a labeled dataset of examples stored in a database coupled with the computer; uniformly sample up to a predetermined large number of examples from a second, larger dataset of unlabeled examples stored in a database coupled with the computer; order the sampled unlabeled examples in order of informativeness for each classifier; determine a minimum subset of the unlabeled examples that are most informative for a maximum number of the classifiers to form an active set for learning; and use editorially-labeled versions of the examples of the active set to re-train the classifiers, thereby improving the accuracy of at least some of the classifiers.

摘要翻译： 一种用于通过主动学习系统来训练分类器的系统，包括具有存储器和处理器的计算机，该处理器被编程为：训练一组初始的二进制一对全分类器，一个分类中的每个类别 在存储在与计算机耦合的数据库中的示例的标记数据集上; 从存储在与计算机耦合的数据库中的未标记示例的第二较大数据集中均匀地采样到预定的大量示例; 按照每个分类器的信息顺序对采样的未标记的示例进行排序; 确定对最大数量的分类器形成用于学习的活动集合的最有帮助的未标记示例的最小子集; 并使用编辑标签的版本的活动集的示例重新训练分类器，从而提高至少一些分类器的准确性。

公开(公告)号：US20120095943A1

公开(公告)日：2012-04-19

申请号：US12905543

申请日：2010-10-15

申请人： Dragomir Yankov ,  Suju Rajan ,  Adwait Ratnaparkhi ,  Rajesh G. Parekh

发明人： Dragomir Yankov ,  Suju Rajan ,  Adwait Ratnaparkhi ,  Rajesh G. Parekh

IPC分类号： G06F15/18

CPC分类号： G06N99/005

摘要： A system for training classifiers in multiple categories through an active learning system, including a computer having a memory and a processor, the processor programmed to: train an initial set of m binary one-versus-all classifiers, one for each category in a taxonomy, on a labeled dataset of examples stored in a database coupled with the computer; uniformly sample up to a predetermined large number of examples from a second, larger dataset of unlabeled examples stored in a database coupled with the computer; order the sampled unlabeled examples in order of informativeness for each classifier; determine a minimum subset of the unlabeled examples that are most informative for a maximum number of the classifiers to form an active set for learning; and use editorially-labeled versions of the examples of the active set to re-train the classifiers, thereby improving the accuracy of at least some of the classifiers.

摘要翻译： 一种用于通过主动学习系统来训练分类器的系统，包括具有存储器和处理器的计算机，该处理器被编程为：训练一组初始的二进制一对全分类器，一个分类中的每个类别 在存储在与计算机耦合的数据库中的示例的标记数据集上; 从存储在与计算机耦合的数据库中的未标记示例的第二较大数据集中均匀地采样到预定的大量示例; 按照每个分类器的信息顺序对采样的未标记的示例进行排序; 确定对最大数量的分类器形成用于学习的活动集合的最有帮助的未标记示例的最小子集; 并使用编辑标签的版本的活动集的示例重新训练分类器，从而提高至少一些分类器的准确性。

公开(公告)号：US20120054027A1

公开(公告)日：2012-03-01

申请号：US12871240

申请日：2010-08-30

申请人： Randolph Preston McAfee ,  Vijay Krishna Narayanan ,  Jayavel Shanmugasundaram ,  Rajesh G. Parekh

发明人： Randolph Preston McAfee ,  Vijay Krishna Narayanan ,  Jayavel Shanmugasundaram ,  Rajesh G. Parekh

IPC分类号： G06Q30/00

CPC分类号： G06Q30/0251

摘要： A network based advertisement system includes an optimizer configured to forecast a supply of opportunities, forecast a supply of guaranteed contracts, and forecast a supply of non-guaranteed contracts. Each opportunity represents a user visiting a webpage. Each guaranteed contract guarantees the matching of an advertisement to a number of opportunities. Each non-guaranteed contract guarantees a user event associated with an advertisement. The optimizer then generates a plan for matching contracts to opportunities based on the forecasted supply of opportunities, the forecasted supply of guaranteed contracts, the forecasted supply of non-guaranteed contracts, and an objective function that balances a group of parameters that define the representativeness of contracts, a cost associated with not serving non-guaranteed contracts, and performance objectives associated with contracts.

摘要翻译： 基于网络的广告系统包括优化器，其被配置为预测机会供应，预测保证合同的供应以及预测非保证合同的供应。 每个机会代表访问网页的用户。 每个保证合同保证广告与许多机会的匹配。 每个非保证合同保证与广告相关联的用户事件。 然后，优化者根据预测的机会供应，预期的保证合同供应，预期的无担保合同供应，以及平衡一组定义代表性的参数的目标函数，生成一个将契约与机会相匹配的计划。 合同，与服务非保证合同相关的成本以及与合同相关的业绩目标。