Patents by Inventor Yoram Singer

Yoram Singer 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: 10769191
    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for classifying data objects. One of the methods includes obtaining data that associates each term in a vocabulary of terms with a respective high-dimensional representation of the term; obtaining classification data for a data object, wherein the classification data includes a respective score for each of a plurality of categories, and wherein each of the categories is associated with a respective category label; computing an aggregate high-dimensional representation for the data object from high-dimensional representations for the category labels associated with the categories and the respective scores; identifying a first term in the vocabulary of terms having a high-dimensional representation that is closest to the aggregate high-dimensional representation; and selecting the first term as a category label for the data object.
    Type: Grant
    Filed: December 19, 2014
    Date of Patent: September 8, 2020
    Assignee: Google LLC
    Inventors: Gregory Sean Corrado, Tomas Mikolov, Samy Bengio, Yoram Singer, Jonathon Shlens, Andrea L. Frome, Jeffrey Adgate Dean, Mohammad Norouzi
  • Patent number: 10713585
    Abstract: Systems and techniques are provided for template exploration in a large-scale machine learning system. A method may include obtaining multiple base templates, each base template comprising multiple features. A template performance score may be obtained for each base template and a first base template may be selected from among the multiple base templates based on the template performance score of the first base template. Multiple cross-templates may be constructed by generating a cross-template of the selected first base template and each of the multiple base templates. Performance of a machine learning model may be tested based on each cross-template to generate a cross-template performance score for each of the cross-templates. A first cross-template may be selected from among the multiple cross-templates based on the cross-template performance score of the cross-template. Accordingly, the first cross-template may be added to the machine learning model.
    Type: Grant
    Filed: December 16, 2013
    Date of Patent: July 14, 2020
    Assignee: Google LLC
    Inventors: Tal Shaked, Tushar Deepak Chandra, James Vincent McFadden, Yoram Singer, Tze Way Eugene Ie
  • Publication number: 20200151614
    Abstract: Systems and techniques are provided for template exploration in a large-scale machine learning system. A method may include obtaining multiple base templates, each base template comprising multiple features. A template performance score may be obtained for each base template and a first base template may be selected from among the multiple base templates based on the template performance score of the first base template. Multiple cross-templates may be constructed by generating a cross-template of the selected first base template and each of the multiple base templates. Performance of a machine learning model may be tested based on each cross-template to generate a cross-template performance score for each of the cross-templates. A first cross-template may be selected from among the multiple cross-templates based on the cross-template performance score of the cross-template. Accordingly, the first cross-template may be added to the machine learning model.
    Type: Application
    Filed: December 16, 2013
    Publication date: May 14, 2020
    Applicant: Google Inc.
    Inventors: Tal Shaked, Tushar Deepak Chandra, James Vincent McFadden, Yoram Singer, Tze Way Eugene Ie
  • Publication number: 20200012905
    Abstract: Systems and techniques are disclosed for labeling objects within an image. The objects may be labeled by selecting an option from a plurality of options such that each option is a potential label for the object. An option may have an option score associated with. Additionally, a relation score may be calculated for a first option and a second option corresponding to a second object in an image. The relation score may be based on a frequency, probability, or observance corresponding to the co-occurrence of text associated with the first option and the second option in a text corpus such as the World Wide Web. An option may be selected as a label for an object based on a global score calculated based at least on an option score and relation score associated with the option.
    Type: Application
    Filed: September 19, 2019
    Publication date: January 9, 2020
    Inventors: Samuel Bengio, Jeffrey Adgate Dean, Quoc V. Le, Jonathon Shlens, Yoram Singer
  • Patent number: 10509772
    Abstract: The present disclosure provides systems and techniques for efficient locking of datasets in a database when updates to a dataset may be delayed. A method may include accumulating a plurality of updates to a first set of one or more values associated with one or more features. The first set of one or more values may be stored within a first database column. Next, it may be determined that a first database column update aggregation rule is satisfied. A lock assigned to at least a portion of at least a first database column may be acquired. Accordingly, one or more values in the first set within the first database column may be updated based on the plurality of updates. In an implementation, the first set of one or more values may be associated with the first lock.
    Type: Grant
    Filed: December 28, 2016
    Date of Patent: December 17, 2019
    Assignee: Google LLC
    Inventors: Tushar Deepak Chandra, Tal Shaked, Yoram Singer, Tze Way Eugene le, Joshua Redstone
  • Patent number: 10445623
    Abstract: Systems and techniques are disclosed for labeling objects within an image. The objects may be labeled by selecting an option from a plurality of options such that each option is a potential label for the object. An option may have an option score associated with. Additionally, a relation score may be calculated for a first option and a second option corresponding to a second object in an image. The relation score may be based on a frequency, probability, or observance corresponding to the co-occurrence of text associated with the first option and the second option in a text corpus such as the World Wide Web. An option may be selected as a label for an object based on a global score calculated based at least on an option score and relation score associated with the option.
    Type: Grant
    Filed: April 14, 2017
    Date of Patent: October 15, 2019
    Assignee: Google LLC
    Inventors: Samy Bengio, Jeffrey Adgate Dean, Quoc V. Le, Jonathon Shlens, Yoram Singer
  • Patent number: 10438129
    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training machine learning systems. One of the methods includes receiving a plurality of training examples; and training a machine learning system on each of the plurality of training examples to determine trained values for weights of a machine learning model, wherein training the machine learning system comprises: assigning an initial value for a regularization penalty for a particular weight for a particular feature; and adjusting the initial value for the regularization penalty for the particular weight for the particular feature during the training of the machine learning system.
    Type: Grant
    Filed: December 30, 2014
    Date of Patent: October 8, 2019
    Assignee: Google LLC
    Inventors: Yoram Singer, Tal Shaked, Tushar Deepak Chandra, Tze Way Eugene Ie
  • Patent number: 10062035
    Abstract: The present disclosure provides methods and systems for using variable length representations of machine learning statistics. A method may include storing an n-bit representation of a first statistic at a first n-bit storage cell. A first update to the first statistic may be received, and it may be determined that the first update causes a first loss of precision of the first statistic as stored in the first n-bit storage cell. Accordingly, an m-bit representation of the first statistic may be stored at a first m-bit storage cell based on the determination. The first m-bit storage cell may be associated with the first n-bit storage cell. As a result, upon receiving an instruction to use the first statistic in a calculation, a combination of the n-bit representation and the m-bit representation may be used to perform the calculation.
    Type: Grant
    Filed: December 12, 2013
    Date of Patent: August 28, 2018
    Assignee: Google LLC
    Inventors: Tal Shaked, Tushar Deepak Chandra, Yoram Singer, Tze Way Eugene Ie, Joshua Redstone
  • Patent number: 9805312
    Abstract: Methods and systems for replacing feature values of features in training data with integer values selected based on a ranking of the feature values. The methods and systems are suitable for preprocessing large-scale machine learning training data.
    Type: Grant
    Filed: December 13, 2013
    Date of Patent: October 31, 2017
    Assignee: Google Inc.
    Inventors: Tal Shaked, Tushar Deepak Chandra, Yoram Singer, Tze Way Eugene Ie, Joshua Redstone
  • Publication number: 20170220906
    Abstract: Systems and techniques are disclosed for labeling objects within an image. The objects may be labeled by selecting an option from a plurality of options such that each option is a potential label for the object. An option may have an option score associated with. Additionally, a relation score may be calculated for a first option and a second option corresponding to a second object in an image. The relation score may be based on a frequency, probability, or observance corresponding to the co-occurrence of text associated with the first option and the second option in a text corpus such as the World Wide Web. An option may be selected as a label for an object based on a global score calculated based at least on an option score and relation score associated with the option.
    Type: Application
    Filed: April 14, 2017
    Publication date: August 3, 2017
    Inventors: Samy Bengio, Jeffrey Adgate Dean, Quoc V. Le, Jonathon Shlens, Yoram Singer
  • Patent number: 9652695
    Abstract: Systems and techniques for labeling objects within an image. The objects may be labeled by selecting an option from a plurality of options such that each option is a potential label for the object. An option may have an option score associated with. Additionally, a relation score may be calculated for a first option and a second option corresponding to a second object in an image. The relation score may be based on a frequency, probability, or observance corresponding to the co-occurrence of text associated with the first option and the second option in a text corpus such as the World Wide Web. An option may be selected as a label for an object based on a global score calculated based at least on an option score and relation score associated with the option.
    Type: Grant
    Filed: December 20, 2013
    Date of Patent: May 16, 2017
    Assignee: Google Inc.
    Inventors: Samy Bengio, Jeffrey Adgate Dean, Quoc Le, Jonathon Shlens, Yoram Singer
  • Patent number: 9569481
    Abstract: The present disclosure provides systems and techniques for efficient locking of datasets in a database when updates to a dataset may be delayed. A method may include accumulating a plurality of updates to a first set of one or more values associated with one or more features. The first set of one or more values may be stored within a first database column. Next, it may be determined that a first database column update aggregation rule is satisfied. A lock assigned to at least a portion of at least a first database column may be acquired. Accordingly, one or more values in the first set within the first database column may be updated based on the plurality of updates. In an implementation, the first set of one or more values may be associated with the first lock.
    Type: Grant
    Filed: December 10, 2013
    Date of Patent: February 14, 2017
    Assignee: Google Inc.
    Inventors: Tushar Deepak Chandra, Tal Shaked, Yoram Singer, Tze Way Eugene Ie, Joshua Redstone
  • Patent number: 9418343
    Abstract: Implementations of the disclosed subject matter provide methods and systems for using a multistage learner for efficiently boosting large datasets in a machine learning system. A method may include obtaining a first plurality of examples for a machine learning system and selecting a first point in time. Next, a second point in time occurring subsequent to the first point in time may be selected. The machine learning system may be trained using m of the first plurality of examples. Each of the m examples may include a feature initially occurring after the second point in time. In addition, the machine learning system may be trained using n of the first plurality of examples, and each of the n examples may include a feature initially occurring after the first point in time.
    Type: Grant
    Filed: December 30, 2013
    Date of Patent: August 16, 2016
    Assignee: Google Inc.
    Inventors: Tushar Deepak Chandra, Tal Shaked, Yoram Singer, Tze Way Eugene Ie, Joshua Redstone
  • Patent number: 9390382
    Abstract: Systems and techniques are disclosed for training a machine learning model based on one or more regularization penalties associated with one or more features. A template having a lower regularization penalty may be given preference over a template having a higher regularization penalty. A regularization penalty may be determined based on domain knowledge. A restrictive regularization penalty may be assigned to a template based on determining that a template occurrence is below a stability threshold and may be modified if the template occurrence meets or exceeds the stability threshold.
    Type: Grant
    Filed: December 30, 2013
    Date of Patent: July 12, 2016
    Assignee: Google Inc.
    Inventors: Yoram Singer, Tal Shaked, Tushar Deepak Chandra, Tze Way Eugene Ie, James Vincent McFadden, Jeremiah Harmsen, Kristen Riedt LeFevre
  • Patent number: 9269057
    Abstract: Systems and techniques are disclosed for generating weighted machine learned models using multi-shard combiners. A learner in a machine learning system may receive labeled positive and negative examples and workers within the learner may be configured to receive either positive or negative examples. A positive and negative statistic may be calculated for a given feature and may either be applied separately in a model or may be combined to generate an overall statistic.
    Type: Grant
    Filed: December 11, 2013
    Date of Patent: February 23, 2016
    Assignee: Google, Inc.
    Inventors: Tushar Deepak Chandra, Tal Shaked, Tze Way Eugene Ie, Yoram Singer, Joshua Redstone
  • Publication number: 20150186795
    Abstract: Implementations of the disclosed subject matter provide methods and systems for using a multistage learner for efficiently boosting large datasets in a machine learning system. A method may include obtaining a first plurality of examples for a machine learning system and selecting a first point in time. Next, a second point in time occurring subsequent to the first point in time may be selected. The machine learning system may be trained using m of the first plurality of examples. Each of the m examples may include a feature initially occurring after the second point in time. In addition, the machine learning system may be trained using n of the first plurality of examples, and each of the n examples may include a feature initially occurring after the first point in time.
    Type: Application
    Filed: December 30, 2013
    Publication date: July 2, 2015
    Applicant: Google Inc.
    Inventors: Tushar Deepak Chandra, Tal Shaked, Yoram Singer, Tze Way Eugene Ie, Joshua Redstone
  • Publication number: 20150186794
    Abstract: Systems and techniques are disclosed for training a machine learning model based on one or more regularization penalties associated with one or more features. A template having a lower regularization penalty may be given preference over a template having a higher regularization penalty. A regularization penalty may be determined based on domain knowledge. A restrictive regularization penalty may be assigned to a template based on determining that a template occurrence is below a stability threshold and may be modified if the template occurrence meets or exceeds the stability threshold.
    Type: Application
    Filed: December 30, 2013
    Publication date: July 2, 2015
    Applicant: Google Inc.
    Inventors: Yoram Singer, Tal Shaked, Tushar Deepak Chandra, Tze Way Eugene Ie, James Vincent McFadden, Jeremiah Harmsen, Kristen Riedt LeFevre
  • Publication number: 20150178383
    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for classifying data objects. One of the methods includes obtaining data that associates each term in a vocabulary of terms with a respective high-dimensional representation of the term; obtaining classification data for a data object, wherein the classification data includes a respective score for each of a plurality of categories, and wherein each of the categories is associated with a respective category label; computing an aggregate high-dimensional representation for the data object from high-dimensional representations for the category labels associated with the categories and the respective scores; identifying a first term in the vocabulary of terms having a high-dimensional representation that is closest to the aggregate high-dimensional representation; and selecting the first term as a category label for the data object.
    Type: Application
    Filed: December 19, 2014
    Publication date: June 25, 2015
    Inventors: Gregory Sean Corrado, Tomas Mikolov, Samy Bengio, Yoram Singer, Jonathon Shlens, Andrea L. Frome, Jeffrey Adgate Dean, Mohammad Norouzi
  • Publication number: 20150178596
    Abstract: Systems and techniques are disclosed for labeling objects within an image. The objects may be labeled by selecting an option from a plurality of options such that each option is a potential label for the object. An option may have an option score associated with. Additionally, a relation score may be calculated for a first option and a second option corresponding to a second object in an image. The relation score may be based on a frequency, probability, or observance corresponding to the co-occurrence of text associated with the first option and the second option in a text corpus such as the World Wide Web. An option may be selected as a label for an object based on a global score calculated based at least on an option score and relation score associated with the option.
    Type: Application
    Filed: December 20, 2013
    Publication date: June 25, 2015
    Applicant: Google Inc.
    Inventors: Samy Bengio, Jeffrey Adgate Dean, Quoc Le, Jonathon Shlens, Yoram Singer
  • Patent number: 9053115
    Abstract: Methods, systems and apparatus for identifying result images for a query image. One or more labels that are associated with the query image are obtained. Candidate images matching the query labels are identified. Visual similarity scores are generated for the candidate images. Each visual similarity score represents the visual similarity of a respective candidate image to the query image. Relevance scores are generated for each of the candidate images based on the visual similarity scores. Each relevance score represents a measure of relevance of the respective candidate images to the query image. The candidate images are ranked based on the relevance scores, a highest ranking subset of the candidate images being identified as result images and referenced by image search results. The result images can be candidate images that satisfy a similarity condition relative to the query image and other result images.
    Type: Grant
    Filed: August 23, 2012
    Date of Patent: June 9, 2015
    Assignee: Google Inc.
    Inventors: Charles J. Rosenberg, Jingbin Wang, Sarah Moussa, Erik Murphy-Chutorian, Andrea Frome, Yoram Singer, Radhika Malpani