Patents by Inventor Galen Andrew
Galen Andrew has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Patent number: 8655887Abstract: Technologies pertaining to compressing time-series signals are described herein. Groups of time-series signals are generated based upon similarities between time-series signals. Each group of time-series signals includes a respective base time-series signal. Ratio signals that are representative of time-series signals are computed, wherein the ratio signals are based upon the base time-series signal and other respective time-series signals in a group of time-series signals.Type: GrantFiled: June 8, 2012Date of Patent: February 18, 2014Assignee: Microsoft CorporationInventors: Jie Liu, Suman Kumar Nath, Feng Zhao, Galen Andrew Reeves, Sorabh Kumar Gandhi
-
Patent number: 8606786Abstract: A system described herein includes a receiver component that receives a dataset that is stored in a computer-readable medium of a computing device, wherein the dataset includes a plurality of queries issued by users to a search engine and a plurality of search results selected by the users upon issuing the plurality of queries. A distribution determiner component determines click distributions over the search results selected by the users with respect to the plurality of queries. A labeler component labels at least two queries in the plurality of queries as being substantially similar to one another based at least in part upon the click distributions over the search results selected by the users with respect to the plurality of queries.Type: GrantFiled: June 22, 2009Date of Patent: December 10, 2013Assignee: Microsoft CorporationInventors: Robert L. Rounthwaite, Galen Andrew, Emre Mehmet Kiciman, Xiaoxin Yin
-
Publication number: 20120246169Abstract: Technologies pertaining to compressing time-series signals are described herein. Groups of time-series signals are generated based upon similarities between time-series signals. Each group of time-series signals includes a respective base time-series signal. Ratio signals that are representative of time-series signals are computed, wherein the ratio signals are based upon the base time-series signal and other respective time-series signals in a group of time-series signals.Type: ApplicationFiled: June 8, 2012Publication date: September 27, 2012Applicant: MICROSOFT CORPORATIONInventors: Jie Liu, Suman Kumar Nath, Feng Zhao, Galen Andrew Reeves, Sorabh Kumar Gandhi
-
Patent number: 8219574Abstract: A system described herein includes a receiver component that receives a query that pertains to a raw time-series signal. A query executor component selectively executes the query over at least one of multiple available compressed representations of the raw time-series signal, wherein the query pertains to at least one of one of determining a trend pertaining to the raw time-series signal, generating a histogram pertaining to the raw time-series signal, or determining a correlation pertaining to the raw time-series signal.Type: GrantFiled: June 22, 2009Date of Patent: July 10, 2012Assignee: Microsoft CorporationInventors: Jie Liu, Suman Kumar Nath, Feng Zhao, Galen Andrew Reeves, Sorabh Kumar Gandhi
-
Patent number: 7984004Abstract: Described herein is a system that facilitates assigning indications of usefulness to query suggestions. The system includes a query suggestion generator component that receives a query and generates a query suggestion based at least in part upon the received query. A model component outputs an indication of usefulness with respect to the query suggestion, wherein the model component is a machine-learned model of user behavior with respect to query suggestions.Type: GrantFiled: January 17, 2008Date of Patent: July 19, 2011Inventors: Galen Andrew, Sooho Park, Robert L. Rounthwaite, Silviu-Petru Cucerzan, Jamie Paul Buckley, Joanna Chan
-
Patent number: 7933847Abstract: An algorithm that employs modified methods developed for optimizing differential functions but which can also handle the special non-differentiabilities that occur with the L1-regularization. The algorithm is a modification of the L-BFGS (limited-memory Broyden-Fletcher-Goldfarb-Shanno) quasi-Newton algorithm, but which can now handle the discontinuity of the gradient using a procedure that chooses a search direction at each iteration and modifies the line search procedure. The algorithm includes an iterative optimization procedure where each iteration approximately minimizes the objective over a constrained region of the space on which the objective is differentiable (in the case of L1-regularization, a given orthant), models the second-order behavior of the objective by considering the loss component alone, using a “line-search” at each iteration that projects search points back onto the chosen orthant, and determines when to stop the line search.Type: GrantFiled: October 17, 2007Date of Patent: April 26, 2011Assignee: Microsoft CorporationInventors: Galen Andrew, Jianfeng Gao
-
Publication number: 20100325132Abstract: A system described herein includes a receiver component that receives a query that pertains to a raw time-series signal. A query executor component selectively executes the query over at least one of multiple available compressed representations of the raw time-series signal, wherein the query pertains to at least one of one of determining a trend pertaining to the raw time-series signal, generating a histogram pertaining to the raw time-series signal, or determining a correlation pertaining to the raw time-series signal.Type: ApplicationFiled: June 22, 2009Publication date: December 23, 2010Applicant: Microsoft CorporationInventors: Jie Liu, Suman Kumar Nath, Feng Zhao, Galen Andrew Reeves, Sorabh Kumar Gandhi
-
Publication number: 20100325133Abstract: A system described herein includes a receiver component that receives a dataset that is stored in a computer-readable medium of a computing device, wherein the dataset includes a plurality of queries issued by users to a search engine and a plurality of search results selected by the users upon issuing the plurality of queries. A distribution determiner component determines click distributions over the search results selected by the users with respect to the plurality of queries. A labeler component labels at least two queries in the plurality of queries as being substantially similar to one another based at least in part upon the click distributions over the search results selected by the users with respect to the plurality of queries.Type: ApplicationFiled: June 22, 2009Publication date: December 23, 2010Applicant: Microsoft CorporationInventors: Robert L. Rounthwaite, Galen Andrew, Emre Mehmet Kiciman, Xiaoxin Yin
-
Patent number: 7844555Abstract: Systems and methods for selecting a ranker for statistical natural language processing are provided. One disclosed system includes a computer program configured to be executed on a computing device, the computer program comprising a data store including reference performance data for a plurality of candidate rankers, the reference performance data being calculated based on a processing of test data by each of the plurality of candidate rankers. The system may further include a ranker selector configured to receive a statistical natural language processing task and a performance target, and determine a selected ranker from the plurality of candidate rankers based on the statistical natural language processing task, the performance target, and the reference performance data.Type: GrantFiled: November 13, 2007Date of Patent: November 30, 2010Assignee: Microsoft CorporationInventors: Jianfeng Gao, Galen Andrew, Mark Johnson, Kristina Toutanova
-
Publication number: 20090187515Abstract: Described herein is a system that facilitates assigning indications of usefulness to query suggestions. The system includes a query suggestion generator component that receives a query and generates a query suggestion based at least in part upon the received query. A model component outputs an indication of usefulness with respect to the query suggestion, wherein the model component is a machine-learned model of user behavior with respect to query suggestions.Type: ApplicationFiled: January 17, 2008Publication date: July 23, 2009Applicant: MICROSOFT CORPORATIONInventors: Galen Andrew, Sooho Park, Robert L. Rounthwaite, Silviu-Petru Cucerzan, Jamie Paul Buckley, Joanna Chan
-
Publication number: 20090125501Abstract: Systems and methods for selecting a ranker for statistical natural language processing are provided. One disclosed system includes a computer program configured to be executed on a computing device, the computer program comprising a data store including reference performance data for a plurality of candidate rankers, the reference performance data being calculated based on a processing of test data by each of the plurality of candidate rankers. The system may further include a ranker selector configured to receive a statistical natural language processing task and a performance target, and determine a selected ranker from the plurality of candidate rankers based on the statistical natural language processing task, the performance target, and the reference performance data.Type: ApplicationFiled: November 13, 2007Publication date: May 14, 2009Applicant: MICROSOFT CORPORATIONInventors: Jianfeng Gao, Galen Andrew, Mark Johnson, Kristina Toutanova
-
Publication number: 20090106173Abstract: An algorithm that employs modified methods developed for optimizing differential functions but which can also handle the special non-differentiabilities that occur with the L1-regularization. The algorithm is a modification of the L-BFGS (limited-memory Broyden-Fletcher-Goldfarb-Shanno) quasi-Newton algorithm, but which can now handle the discontinuity of the gradient using a procedure that chooses a search direction at each iteration and modifies the line search procedure. The algorithm includes an iterative optimization procedure where each iteration approximately minimizes the objective over a constrained region of the space on which the objective is differentiable (in the case of L1-regularization, a given orthant), models the second-order behavior of the objective by considering the loss component alone, using a “line-search” at each iteration that projects search points back onto the chosen orthant, and determines when to stop the line search.Type: ApplicationFiled: October 17, 2007Publication date: April 23, 2009Applicant: MICROSOFT CORPORATIONInventors: Galen Andrew, Jianfeng Gao
-
Publication number: 20070083357Abstract: A weighted linear word alignment model linearly combines weighted features to score a word alignment for a bilingual, aligned pair of text fragments. The features are each weighted by a feature weight. One of the features is a word association metric, which may be generated from surface statistics.Type: ApplicationFiled: July 12, 2006Publication date: April 12, 2007Inventors: Robert Moore, Wen-tau Yih, Galen Andrew, Kristina Toutanova