Patents by Inventor DAVID G. GRANGIER
DAVID G. GRANGIER 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).
-
Publication number: 20240135098Abstract: A collection of data that is extremely large can be difficult to search and/or analyze. Relevance may be dramatically improved by automatically classifying queries and web pages in useful categories, and using these classification scores as relevance features. A thorough approach may require building a large number of classifiers, corresponding to the various types of information, activities, and products. Creation of classifiers and schematizers is provided on large data sets. Exercising the classifiers and schematizers on hundreds of millions of items may expose value that is inherent to the data by adding usable meta-data. Some aspects include active labeling exploration, automatic regularization and cold start, scaling with the number of items and the number of classifiers, active featuring, and segmentation and schematization.Type: ApplicationFiled: December 1, 2023Publication date: April 25, 2024Inventors: Patrice Y. SIMARD, David G. GRANGIER, Leon BOTTOU, Saleema A. AMERSHI
-
Patent number: 11023677Abstract: A collection of data that is extremely large can be difficult to search and/or analyze. Relevance may be dramatically improved by automatically classifying queries and web pages in useful categories, and using these classification scores as relevance features. A thorough approach may require building a large number of classifiers, corresponding to the various types of information, activities, and products. Creation of classifiers and schematizers is provided on large data sets. Exercising the classifiers and schematizers on hundreds of millions of items may expose value that is inherent to the data by adding usable meta-data. Some aspects include active labeling exploration, automatic regularization and cold start, scaling with the number of items and the number of classifiers, active featuring, and segmentation and schematization.Type: GrantFiled: July 13, 2016Date of Patent: June 1, 2021Assignee: MICROSOFT TECHNOLOGY LICENSING, LLCInventors: Patrice Y. Simard, David Max Chickering, David G. Grangier, Aparna Lakshmiratan, Saleema A. Amershi
-
Patent number: 10839790Abstract: Exemplary embodiments relate to improvements to neural networks for translation and other sequence-to-sequence tasks. A convolutional neural network may include multiple blocks, each having a convolution layer and gated linear units; gating may determine what information passes through to the next block level. Residual connections, which add the input of a block back to its output, may be applied around each block. Further, an attention may be applied to determine which word is most relevant to translate next. By applying repeated passes of the attention to multiple layers of the decoder, the decoder is able to work on the entire structure of a sentence at once (with no temporal dependency). In addition to better accuracy, this configuration is better at capturing long-range dependencies, better models the hierarchical syntax structure of a sentence, and is highly parallelizable and thus faster to run on hardware.Type: GrantFiled: December 20, 2017Date of Patent: November 17, 2020Assignee: FACEBOOK, INC.Inventors: Jonas Gehring, Michael Auli, Yann Nicolas Dauphin, David G. Grangier, Dzianis Yarats
-
Patent number: 10372815Abstract: A collection of data that is extremely large can be difficult to search and/or analyze. Relevance may be dramatically improved by automatically classifying queries and web pages in useful categories, and using these classification scores as relevance features. A thorough approach may require building a large number of classifiers, corresponding to the various types of information, activities, and products. Creation of classifiers and schematizers is provided on large data sets. Exercising the classifiers and schematizers on hundreds of millions of items may expose value that is inherent to the data by adding usable meta-data. Some aspects include active labeling exploration, automatic regularization and cold start, scaling with the number of items and the number of classifiers, active featuring, and segmentation and schematization.Type: GrantFiled: November 8, 2013Date of Patent: August 6, 2019Assignee: Microsoft Technology Licensing, LLCInventors: Patrice Y. Simard, David G. Grangier, Leon Bottou, Saleema A. Amershi
-
Publication number: 20190213252Abstract: A collection of data that is extremely large can be difficult to search and/or analyze. Relevance may be dramatically improved by automatically classifying queries and web pages in useful categories, and using these classification scores as relevance features. A thorough approach may require building a large number of classifiers, corresponding to the various types of information, activities, and products. Creation of classifiers and schematizers is provided on large data sets. Exercising the classifiers and schematizers on hundreds of millions of items may expose value that is inherent to the data by adding usable meta-data. Some aspects include active labeling exploration, automatic regularization and cold start, scaling with the number of items and the number of classifiers, active featuring, and segmentation and schematization.Type: ApplicationFiled: March 19, 2019Publication date: July 11, 2019Inventors: PATRICE Y. SIMARD, DAVID G. GRANGIER, LEON BOTTOU, SALEEMA A. AMERSHI
-
Publication number: 20180261214Abstract: Exemplary embodiments relate to improvements to neural networks for translation and other sequence-to-sequence tasks. A convolutional neural network may include multiple blocks, each having a convolution layer and gated linear units; gating may determine what information passes through to the next block level. Residual connections, which add the input of a block back to its output, may be applied around each block. Further, an attention may be applied to determine which word is most relevant to translate next. By applying repeated passes of the attention to multiple layers of the decoder, the decoder is able to work on the entire structure of a sentence at once (with no temporal dependency). In addition to better accuracy, this configuration is better at capturing long-range dependencies, better models the hierarchical syntax structure of a sentence, and is highly parallelizable and thus faster to run on hardware.Type: ApplicationFiled: December 20, 2017Publication date: September 13, 2018Inventors: Jonas Gehring, Michael Auli, Yann Nicolas Dauphin, David G. Grangier, Dzianis Yarats
-
Patent number: 9886669Abstract: Methods, computer systems, computer-storage media, and graphical user interfaces are provided for visualizing a performance of a machine-learned model. An interactive graphical user interface includes an item representation display area that displays a plurality of item representations corresponding to a plurality of items processed by the machine-learned model. The plurality of item representations are arranged according to scores assigned to the plurality of items by the machine-learned model. Further, each of the plurality of item representations is visually configured to represent a label assigned to a corresponding item.Type: GrantFiled: February 26, 2014Date of Patent: February 6, 2018Assignee: MICROSOFT TECHNOLOGY LICENSING, LLCInventors: Saleema A. Amershi, Steven M. Drucker, Bongshin Lee, Patrice Yvon Rene Simard, Aparna Lakshmiratan, Carlos Garcia Jurado Suarez, Denis X. Charles, David G. Grangier, David Maxwell Chickering
-
Patent number: 9779081Abstract: A collection of data that is extremely large can be difficult to search and/or analyze. Relevance may be dramatically improved by automatically classifying queries and web pages in useful categories, and using these classification scores as relevance features. A thorough approach may require building a large number of classifiers, corresponding to the various types of information, activities, and products. Creation of classifiers and schematizers is provided on large data sets. Exercising the classifiers and schematizers on hundreds of millions of items may expose value that is inherent to the data by adding usable meta-data. Some aspects include active labeling exploration, automatic regularization and cold start, scaling with the number of items and the number of classifiers, active featuring, and segmentation and schematization.Type: GrantFiled: April 21, 2016Date of Patent: October 3, 2017Assignee: Microsoft Technology Licensing, LLCInventors: Patrice Y. Simard, David Max Chickering, David G. Grangier, Denis X. Charles, Leon Bottou, Carlos Garcia Jurado Suarez
-
Publication number: 20170039486Abstract: A collection of data that is extremely large can be difficult to search and/or analyze. Relevance may be dramatically improved by automatically classifying queries and web pages in useful categories, and using these classification scores as relevance features. A thorough approach may require building a large number of classifiers, corresponding to the various types of information, activities, and products. Creation of classifiers and schematizers is provided on large data sets. Exercising the classifiers and schematizers on hundreds of millions of items may expose value that is inherent to the data by adding usable meta-data. Some aspects include active labeling exploration, automatic regularization and cold start, scaling with the number of items and the number of classifiers, active featuring, and segmentation and schematization.Type: ApplicationFiled: July 13, 2016Publication date: February 9, 2017Inventors: Patrice Y. SIMARD, David Max CHICKERING, David G. GRANGIER, Aparna LAKSHMIRATAN, Saleema A. AMERSHI
-
Patent number: 9489373Abstract: A collection of data that is extremely large can be difficult to search and/or analyze. Relevance may be dramatically improved by automatically classifying queries and web pages in useful categories, and using these classification scores as relevance features. A thorough approach may require building a large number of classifiers, corresponding to the various types of information, activities, and products. Creation of classifiers and schematizers is provided on large data sets. Exercising the classifiers and schematizers on hundreds of millions of items may expose value that is inherent to the data by adding usable meta-data. Some aspects include active labeling exploration, automatic regularization and cold start, scaling with the number of items and the number of classifiers, active featuring, and segmentation and schematization.Type: GrantFiled: November 8, 2013Date of Patent: November 8, 2016Assignee: Microsoft Technology Licensing, LLCInventors: Patrice Y. Simard, David Max Chickering, David G. Grangier, Denis X. Charles, Leon Bottou, Saleema A. Amershi, Aparna Lakshmiratan, Carlos Garcia Jurado Suarez
-
Patent number: 9430460Abstract: A collection of data that is extremely large can be difficult to search and/or analyze. Relevance may be dramatically improved by automatically classifying queries and web pages in useful categories, and using these classification scores as relevance features. A thorough approach may require building a large number of classifiers, corresponding to the various types of information, activities, and products. Creation of classifiers and schematizers is provided on large data sets. Exercising the classifiers and schematizers on hundreds of millions of items may expose value that is inherent to the data by adding usable meta-data. Some aspects include active labeling exploration, automatic regularization and cold start, scaling with the number of items and the number of classifiers, active featuring, and segmentation and schematization.Type: GrantFiled: November 8, 2013Date of Patent: August 30, 2016Assignee: Microsoft Technology Licensing, LLCInventors: Patrice Y. Simard, David Max Chickering, David G. Grangier, Aparna Lakshmiratan, Saleema A. Amershi
-
Publication number: 20160239761Abstract: A collection of data that is extremely large can be difficult to search and/or analyze. Relevance may be dramatically improved by automatically classifying queries and web pages in useful categories, and using these classification scores as relevance features. A thorough approach may require building a large number of classifiers, corresponding to the various types of information, activities, and products. Creation of classifiers and schematizers is provided on large data sets. Exercising the classifiers and schematizers on hundreds of millions of items may expose value that is inherent to the data by adding usable meta-data. Some aspects include active labeling exploration, automatic regularization and cold start, scaling with the number of items and the number of classifiers, active featuring, and segmentation and schematization.Type: ApplicationFiled: April 21, 2016Publication date: August 18, 2016Inventors: PATRICE Y. SIMARD, DAVID MAX CHICKERING, DAVID G. GRANGIER, DENIS X. CHARLES, LEON BOTTOU, CARLOS GARCIA JURADO SUAREZ
-
Patent number: 9355088Abstract: A collection of data that is extremely large can be difficult to search and/or analyze. Relevance may be dramatically improved by automatically classifying queries and web pages in useful categories, and using these classification scores as relevance features. A thorough approach may require building a large number of classifiers, corresponding to the various types of information, activities, and products. Creation of classifiers and schematizers is provided on large data sets. Exercising the classifiers and schematizers on hundreds of millions of items may expose value that is inherent to the data by adding usable meta-data. Some aspects include active labeling exploration, automatic regularization and cold start, scaling with the number of items and the number of classifiers, active featuring, and segmentation and schematization.Type: GrantFiled: November 8, 2013Date of Patent: May 31, 2016Assignee: Microsoft Technology Licensing, LLCInventors: Patrice Y. Simard, David Max Chickering, David G. Grangier, Denis X. Charles, Leon Bottou, Carlos Garcia Jurado Suarez
-
Publication number: 20150242761Abstract: Methods, computer systems, computer-storage media, and graphical user interfaces are provided for visualizing a performance of a machine-learned model. An interactive graphical user interface includes an item representation display area that displays a plurality of item representations corresponding to a plurality of items processed by the machine-learned model. The plurality of item representations are arranged according to scores assigned to the plurality of items by the machine-learned model. Further, each of the plurality of item representations is visually configured to represent a label assigned to a corresponding item.Type: ApplicationFiled: February 26, 2014Publication date: August 27, 2015Applicant: MICROSOFT CORPORATIONInventors: SALEEMA A. AMERSHI, STEVEN M. DRUCKER, BONGSHIN LEE, PATRICE YVON RENE SIMARD, APARNA LAKSHMIRATAN, CARLOS GARCIA JURADO SUAREZ, DENIS X. CHARLES, DAVID G. GRANGIER, DAVID MAXWELL CHICKERING
-
Publication number: 20150019461Abstract: A collection of data that is extremely large can be difficult to search and/or analyze. Relevance may be dramatically improved by automatically classifying queries and web pages in useful categories, and using these classification scores as relevance features. A thorough approach may require building a large number of classifiers, corresponding to the various types of information, activities, and products. Creation of classifiers and schematizers is provided on large data sets. Exercising the classifiers and schematizers on hundreds of millions of items may expose value that is inherent to the data by adding usable meta-data. Some aspects include active labeling exploration, automatic regularization and cold start, scaling with the number of items and the number of classifiers, active featuring, and segmentation and schematization.Type: ApplicationFiled: November 8, 2013Publication date: January 15, 2015Applicant: Microsoft CorporationInventors: PATRICE Y. SIMARD, DAVID MAX CHICKERING, DAVID G. GRANGIER, DENIS X. CHARLES, LEON BOTTOU, SALEEMA A. AMERSHI, APARNA LAKSHMIRATAN, CARLOS GARCIA JURADO SUAREZ
-
Publication number: 20150019211Abstract: A collection of data that is extremely large can be difficult to search and/or analyze. Relevance may be dramatically improved by automatically classifying queries and web pages in useful categories, and using these classification scores as relevance features. A thorough approach may require building a large number of classifiers, corresponding to the various types of information, activities, and products. Creation of classifiers and schematizers is provided on large data sets. Exercising the classifiers and schematizers on hundreds of millions of items may expose value that is inherent to the data by adding usable meta-data. Some aspects include active labeling exploration, automatic regularization and cold start, scaling with the number of items and the number of classifiers, active featuring, and segmentation and schematization.Type: ApplicationFiled: November 8, 2013Publication date: January 15, 2015Applicant: Microsoft CorportionInventors: PATRICE Y. SIMARD, DAVID G. GRANGIER, LEON BOTTOU, SALEEMA A. AMERSHI
-
Publication number: 20150019463Abstract: A collection of data that is extremely large can be difficult to search and/or analyze. Relevance may be dramatically improved by automatically classifying queries and web pages in useful categories, and using these classification scores as relevance features. A thorough approach may require building a large number of classifiers, corresponding to the various types of information, activities, and products. Creation of classifiers and schematizers is provided on large data sets. Exercising the classifiers and schematizers on hundreds of millions of items may expose value that is inherent to the data by adding usable meta-data. Some aspects include active labeling exploration, automatic regularization and cold start, scaling with the number of items and the number of classifiers, active featuring, and segmentation and schematization.Type: ApplicationFiled: November 8, 2013Publication date: January 15, 2015Applicant: Microsoft CorporationInventors: PATRICE Y. SIMARD, DAVID MAX CHICKERING, DAVID G. GRANGIER, APARNA LAKSHMIRATAN, SALEEMA A. AMERSHI
-
Publication number: 20150019204Abstract: A collection of data that is extremely large can be difficult to search and/or analyze. Relevance may be dramatically improved by automatically classifying queries and web pages in useful categories, and using these classification scores as relevance features. A thorough approach may require building a large number of classifiers, corresponding to the various types of information, activities, and products. Creation of classifiers and schematizers is provided on large data sets. Exercising the classifiers and schematizers on hundreds of millions of items may expose value that is inherent to the data by adding usable meta-data. Some aspects include active labeling exploration, automatic regularization and cold start, scaling with the number of items and the number of classifiers, active featuring, and segmentation and schematization.Type: ApplicationFiled: November 8, 2013Publication date: January 15, 2015Applicant: Microsoft CorporationInventors: PATRICE Y. SIMARD, DAVID MAX CHICKERING, DAVID G. GRANGIER, DENIS X. CHARLES, LEON BOTTOU, CARLOS GARCIA JURADO SUAREZ