Abstract: A method of summarizing a document includes a step of extracting one or more sections of the document. The method also includes a step of separating at least one of the one or more extracted sections into one or more subsections based at least in part on a conjunctive structure of the section, wherein each subsection comprises one or more terms. The method also includes steps of determining whether one or more terms within a designated set of terms are present within at least one of the one or more subsections and, responsive to a determination that one or more terms within the designated set of terms are present within at least one of the one or more subsections, removing the one or more terms from the one or more subsections. The method also includes a step of aggregating at least a portion of the one or more sections into a summary of the document.

Abstract: A method for adaptive display of internet advertisements to look-alike users using a desired user profile dataset as a seed to machine learning modules. Upon availability of a desired user profile, that user profile is mapped other look-alike users (from a larger database of users). The method proceeds to normalize the desired user profile object, proceeds to normalize known user profile objects, then seeding a machine-learning training model with the normalized desired user profile object. A scoring engine uses the normalized user profiles for matching based on extracted features (i.e. extracted from the normalized user profile objects). Once look-alike users have been identified, the internet display system may serve advertisements to the look-alike users, and analyze look-alike users' behaviors for storing the predicted similar user profile objects into the desired user profile object dataset, thus adapting to changing user behavior.

Abstract: A method for adaptive display of internet advertisements to look-alike users using a desired user profile dataset as a seed to machine learning modules. Upon availability of a desired user profile, that user profile is mapped other look-alike users (from a larger database of users). The method proceeds to normalize the desired user profile object, proceeds to normalize known user profile objects, then seeding a machine-learning training model with the normalized desired user profile object. A scoring engine uses the normalized user profiles for matching based on extracted features (i.e. extracted from the normalized user profile objects). Once look-alike users have been identified, the internet display system may serve advertisements to the look-alike users, and analyze look-alike users' behaviors for storing the predicted similar user profile objects into the desired user profile object dataset, thus adapting to changing user behavior.

Abstract: A method of summarizing a document includes a step of extracting one or more sections of the document. The method also includes a step of separating at least one of the one or more extracted sections into one or more subsections based at least in part on a conjunctive structure of the section, wherein each subsection comprises one or more terms. The method also includes steps of determining whether one or more terms within a designated set of terms are present within at least one of the one or more subsections and, responsive to a determination that one or more terms within the designated set of terms are present within at least one of the one or more subsections, removing the one or more terms from the one or more subsections. The method also includes a step of aggregating at least a portion of the one or more sections into a summary of the document.

Abstract: A method and apparatus for transporting data for a data warehouse application is described. The data from an operational data store (the source database) is organized in non-overlapping data partitions. Separate execution threads read the data from the operational data store concurrently. This is followed by concurrent transformation of the data in multiple execution threads. Finally, the data is loaded into the target data warehouse concurrently using multiple execution threads. By using multiple execution threads, the data contention is reduced. Thereby the apparatus and method of the present invention achieves increased throughput.