Patents by Inventor Kunal Talwar
Kunal Talwar 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: 11955732Abstract: Millimeter wave (mmWave) technology, apparatuses, and methods that relate to transceivers, receivers, and antenna structures for wireless communications are described. The various aspects include co-located millimeter wave (mmWave) and near-field communication (NFC) antennas, scalable phased array radio transceiver architecture (SPARTA), phased array distributed communication system with MIMO support and phase noise synchronization over a single coax cable, communicating RF signals over cable (RFoC) in a distributed phased array communication system, clock noise leakage reduction, IF-to-RF companion chip for backwards and forwards compatibility and modularity, on-package matching networks, 5G scalable receiver (Rx) architecture, among others.Type: GrantFiled: December 27, 2022Date of Patent: April 9, 2024Assignee: Intel CorporationInventors: Erkan Alpman, Arnaud Lucres Amadjikpe, Omer Asaf, Kameran Azadet, Rotem Banin, Miroslav Baryakh, Anat Bazov, Stefano Brenna, Bryan K. Casper, Anandaroop Chakrabarti, Gregory Chance, Debabani Choudhury, Emanuel Cohen, Claudio Da Silva, Sidharth Dalmia, Saeid Daneshgar Asl, Kaushik Dasgupta, Kunal Datta, Brandon Davis, Ofir Degani, Amr M. Fahim, Amit Freiman, Michael Genossar, Eran Gerson, Eyal Goldberger, Eshel Gordon, Meir Gordon, Josef Hagn, Shinwon Kang, Te Yu Kao, Noam Kogan, Mikko S. Komulainen, Igal Yehuda Kushnir, Saku Lahti, Mikko M. Lampinen, Naftali Landsberg, Wook Bong Lee, Run Levinger, Albert Molina, Resti Montoya Moreno, Tawfiq Musah, Nathan G. Narevsky, Hosein Nikopour, Oner Orhan, Georgios Palaskas, Stefano Pellerano, Ron Pongratz, Ashoke Ravi, Shmuel Ravid, Peter Andrew Sagazio, Eren Sasoglu, Lior Shakedd, Gadi Shor, Baljit Singh, Menashe Soffer, Ra'anan Sover, Shilpa Talwar, Nebil Tanzi, Moshe Teplitsky, Chintan S. Thakkar, Jayprakash Thakur, Avi Tsarfati, Yossi Tsfati, Marian Verhelst, Nir Weisman, Shuhei Yamada, Ana M. Yepes, Duncan Kitchin
-
Patent number: 11726769Abstract: Systems and methods for learning differentially private machine-learned models are provided. A computing system can include one or more server computing devices comprising one or more processors and one or more non-transitory computer-readable media that collectively store instructions that, when executed by the one or more processors cause the one or more server computing devices to perform operations. The operations can include selecting a subset of client computing devices from a pool of available client computing devices; providing a machine-learned model to the selected client computing devices; receiving, from each selected client computing device, a local update for the machine-learned model; determining a differentially private aggregate of the local updates; and determining an updated machine-learned model based at least in part on the data-weighted average of the local updates.Type: GrantFiled: October 12, 2022Date of Patent: August 15, 2023Assignee: GOOGLE LLCInventors: Hugh Brendan McMahan, Kunal Talwar, Li Zhang, Daniel Ramage
-
Publication number: 20230066545Abstract: Systems and methods for learning differentially private machine-learned models are provided. A computing system can include one or more server computing devices comprising one or more processors and one or more non-transitory computer-readable media that collectively store instructions that, when executed by the one or more processors cause the one or more server computing devices to perform operations. The operations can include selecting a subset of client computing devices from a pool of available client computing devices; providing a machine-learned model to the selected client computing devices; receiving, from each selected client computing device, a local update for the machine-learned model; determining a differentially private aggregate of the local updates; and determining an updated machine-learned model based at least in part on the data-weighted average of the local updates.Type: ApplicationFiled: October 12, 2022Publication date: March 2, 2023Inventors: Hugh Brendan McMahan, Kunal Talwar, Li Zhang, Daniel Ramage
-
Patent number: 11475350Abstract: Systems and methods for learning differentially private machine-learned models are provided. A computing system can include one or more server computing devices comprising one or more processors and one or more non-transitory computer-readable media that collectively store instructions that, when executed by the one or more processors cause the one or more server computing devices to perform operations. The operations can include selecting a subset of client computing devices from a pool of available client computing devices; providing a machine-learned model to the selected client computing devices; receiving, from each selected client computing device, a local update for the machine-learned model; determining a differentially private aggregate of the local updates; and determining an updated machine-learned model based at least in part on the data-weighted average of the local updates.Type: GrantFiled: January 22, 2018Date of Patent: October 18, 2022Assignee: GOOGLE LLCInventors: Hugh Brendan McMahan, Kunal Talwar, Li Zhang, Daniel Ramage
-
Publication number: 20210158211Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media for training a machine learning model. The method includes obtaining a training data set comprising a plurality of training examples; determining i) a stochastic gradient descent step size schedule, ii) a stochastic gradient descent noise schedule, and iii) a stochastic gradient descent batch size schedule, wherein the stochastic gradient descent batch size schedule comprises a sequence of varying batch sizes; and training a machine learning model on the training data set, comprising performing stochastic gradient descent according to the i) stochastic gradient descent step size schedule, ii) stochastic gradient descent noise schedule, and iii) stochastic gradient descent batch size schedule to adjust a machine learning model loss function.Type: ApplicationFiled: November 20, 2020Publication date: May 27, 2021Inventors: Kunal Talwar, Vitaly Feldman, Tomer Koren
-
Patent number: 10540519Abstract: The privacy of linear queries on histograms is protected. A database containing private data is queried. Base decomposition is performed to recursively compute an orthonormal basis for the database space. Using correlated (or Gaussian) noise and/or least squares estimation, an answer having differential privacy is generated and provided in response to the query. In some implementations, the differential privacy is ?-differential privacy (pure differential privacy) or is (?,?)-differential privacy (i.e., approximate differential privacy). In some implementations, the data in the database may be dense. Such implementations may use correlated noise without using least squares estimation. In other implementations, the data in the database may be sparse. Such implementations may use least squares estimation with or without using correlated noise.Type: GrantFiled: October 24, 2018Date of Patent: January 21, 2020Assignee: Microsoft Technology Licensing, LLCInventors: Li Zhang, Kunal Talwar, Aleksandar Nikolov
-
Publication number: 20190227980Abstract: Systems and methods for learning differentially private machine-learned models are provided. A computing system can include one or more server computing devices comprising one or more processors and one or more non-transitory computer-readable media that collectively store instructions that, when executed by the one or more processors cause the one or more server computing devices to perform operations. The operations can include selecting a subset of client computing devices from a pool of available client computing devices; providing a machine-learned model to the selected client computing devices; receiving, from each selected client computing device, a local update for the machine-learned model; determining a differentially private aggregate of the local updates; and determining an updated machine-learned model based at least in part on the data-weighted average of the local updates.Type: ApplicationFiled: January 22, 2018Publication date: July 25, 2019Inventors: Hugh Brendan McMahan, Kunal Talwar, Li Zhang, Daniel Ramage
-
Publication number: 20190065225Abstract: A method for packing virtual machines onto host devices may calculate scarcity values for several different parameters. A host's scarcity for a parameter may be determined by multiplying the host's capacity for a parameter with the overall scarcity of that parameter. The sum of a host's scarcity for all the parameters determines the host's overall scarcity. Hosts having the highest scarcity are attempted to be populated with a group of virtual machines selected for compatibility with the host. In many cases, several different scenarios may be evaluated and an optimal scenario implemented. The method gives a high priority to those virtual machines that consume scarce resources, with the scarcity being a function of the available hardware and the virtual machines that may be placed on them.Type: ApplicationFiled: February 11, 2016Publication date: February 28, 2019Inventors: Lincoln K. Uyeda, Rina Panigrahy, Ehud Wieder, Kunal Talwar
-
Publication number: 20190057224Abstract: The privacy of linear queries on histograms is protected. A database containing private data is queried. Base decomposition is performed to recursively compute an orthonormal basis for the database space. Using correlated (or Gaussian) noise and/or least squares estimation, an answer having differential privacy is generated and provided in response to the query. In some implementations, the differential privacy is ?-differential privacy (pure differential privacy) or is (?,?)-differential privacy (i.e., approximate differential privacy). In some implementations, the data in the database may be dense. Such implementations may use correlated noise without using least squares estimation. In other implementations, the data in the database may be sparse. Such implementations may use least squares estimation with or without using correlated noise.Type: ApplicationFiled: October 24, 2018Publication date: February 21, 2019Inventors: Li Zhang, Kunal Talwar, Aleksandar Nikolov
-
Patent number: 10121024Abstract: The privacy of linear queries on histograms is protected. A database containing private data is queried. Base decomposition is performed to recursively compute an orthonormal basis for the database space. Using correlated (or Gaussian) noise and/or least squares estimation, an answer having differential privacy is generated and provided in response to the query. In some implementations, the differential privacy is ?-differential privacy (pure differential privacy) or is (?,?)-differential privacy (i.e., approximate differential privacy). In some implementations, the data in the database may be dense. Such implementations may use correlated noise without using least squares estimation. In other implementations, the data in the database may be sparse. Such implementations may use least squares estimation with or without using correlated noise.Type: GrantFiled: May 4, 2017Date of Patent: November 6, 2018Assignee: Microsoft Technology Licensing, LLCInventors: Li Zhang, Kunal Talwar, Aleksandar Nikolov
-
Patent number: 9934311Abstract: Weighted features associated with a document are scaled using scales to generate a set of unweighted elements for each scale. A sketch is generated for each scale by sampling the unweighted elements generated for the scale. The scales are chosen based on a selected cutoff factor so that documents that have a similarity that is less than the cutoff factor might have no scales in common, while documents that have a similarity that is greater than the cutoff factor will have at sufficiently many but at least one scale in common. The similarity of these documents can be estimated using the sketches associated with each of the documents for the common scales.Type: GrantFiled: April 24, 2014Date of Patent: April 3, 2018Assignee: Microsoft Technology Licensing, LLCInventors: Bernhard Haeupler, Kunal Talwar, Mark S. Manasse
-
Publication number: 20170235974Abstract: The privacy of linear queries on histograms is protected. A database containing private data is queried. Base decomposition is performed to recursively compute an orthonormal basis for the database space. Using correlated (or Gaussian) noise and/or least squares estimation, an answer having differential privacy is generated and provided in response to the query. In some implementations, the differential privacy is ?-differential privacy (pure differential privacy) or is (?,?)-differential privacy (i.e., approximate differential privacy). In some implementations, the data in the database may be dense. Such implementations may use correlated noise without using least squares estimation. In other implementations, the data in the database may be sparse. Such implementations may use least squares estimation with or without using correlated noise.Type: ApplicationFiled: May 4, 2017Publication date: August 17, 2017Inventors: Li ZHANG, Kunal TALWAR, Aleksandar NIKOLOV
-
Patent number: 9672364Abstract: The privacy of linear queries on histograms is protected. A database containing private data is queried. Base decomposition is performed to recursively compute an orthonormal basis for the database space. Using correlated (or Gaussian) noise and/or least squares estimation, an answer having differential privacy is generated and provided in response to the query. In some implementations, the differential privacy is ?-differential privacy (pure differential privacy) or is (?,?)-differential privacy (i.e., approximate differential privacy). In some implementations, the data in the database may be dense. Such implementations may use correlated noise without using least squares estimation. In other implementations, the data in the database may be sparse. Such implementations may use least squares estimation with or without using correlated noise.Type: GrantFiled: March 15, 2013Date of Patent: June 6, 2017Assignee: MICROSOFT TECHNOLOGY LICENSING, LLCInventors: Li Zhang, Kunal Talwar, Aleksandar Nikolov
-
Publication number: 20160162309Abstract: A method for packing virtual machines onto host devices may calculate scarcity values for several different parameters. A host's scarcity for a parameter may be determined by multiplying the host's capacity for a parameter with the overall scarcity of that parameter. The sum of a host's scarcity for all the parameters determines the host's overall scarcity. Hosts having the highest scarcity are attempted to be populated with a group of virtual machines selected for compatibility with the host. In many cases, several different scenarios may be evaluated and an optimal scenario implemented. The method gives a high priority to those virtual machines that consume scarce resources, with the scarcity being a function of the available hardware and the virtual machines that may be placed on them.Type: ApplicationFiled: February 11, 2016Publication date: June 9, 2016Inventors: Lincoln K. Uyeda, Rina Panigrahy, Ehud Wieder, Kunal Talwar
-
Patent number: 9292320Abstract: A method for packing virtual machines onto host devices may calculate scarcity values for several different parameters. A host's scarcity for a parameter may be determined by multiplying the host's capacity for a parameter with the overall scarcity of that parameter. The sum of a host's scarcity for all the parameters determines the host's overall scarcity. Hosts having the highest scarcity are attempted to be populated with a group of virtual machines selected for compatibility with the host. In many cases, several different scenarios may be evaluated and an optimal scenario implemented. The method gives a high priority to those virtual machines that consume scarce resources, with the scarcity being a function of the available hardware and the virtual machines that may be placed on them.Type: GrantFiled: June 10, 2013Date of Patent: March 22, 2016Assignee: Microsoft Technology Licensing, LLCInventors: Lincoln K. Uyeda, Rina Panigrahy, Ehud Wieder, Kunal Talwar
-
Publication number: 20150310102Abstract: Weighted features associated with a document are scaled using scales to generate a set of unweighted elements for each scale. A sketch is generated for each scale by sampling the unweighted elements generated for the scale. The scales are chosen based on a selected cutoff factor so that documents that have a similarity that is less than the cutoff factor might have no scales in common, while documents that have a similarity that is greater than the cutoff factor will have at sufficiently many but at least one scale in common. The similarity of these documents can be estimated using the sketches associated with each of the documents for the common scales.Type: ApplicationFiled: April 24, 2014Publication date: October 29, 2015Applicant: Microsoft CorporationInventors: Bernhard Haeupler, Kunal Talwar, Mark S. Manasse
-
Publication number: 20140283091Abstract: The privacy of linear queries on histograms is protected. A database containing private data is queried. Base decomposition is performed to recursively compute an orthonormal basis for the database space. Using correlated (or Gaussian) noise and/or least squares estimation, an answer having differential privacy is generated and provided in response to the query. In some implementations, the differential privacy is ?-differential privacy (pure differential privacy) or is (?,?)-differential privacy (i.e., approximate differential privacy). In some implementations, the data in the database may be dense. Such implementations may use correlated noise without using least squares estimation. In other implementations, the data in the database may be sparse. Such implementations may use least squares estimation with or without using correlated noise.Type: ApplicationFiled: March 15, 2013Publication date: September 18, 2014Inventors: Li Zhang, Kunal Talwar, Aleksandar Nikolov
-
Patent number: 8661047Abstract: A system for answering sets of queries on a set of private data while providing differential privacy protection is provided. The set of queries is received and applied to the set of private data to generate a set of results or answers. A geometric representation of the set of queries is generated. Example geometric representations include polytopes. Error values are generated for the set of queries using a K-norm mechanism based on values sampled from the geometric representation. The sampled values are added to the set of results to provide the differential privacy protection. By generating the error values based on the set of queries rather than the set of results or the set of private data, the amount of error added to the generated results to achieve a level of differential privacy protection is reduced.Type: GrantFiled: May 17, 2010Date of Patent: February 25, 2014Assignee: Microsoft CorporationInventors: Kunal Talwar, Moritz A. W. Hardt
-
Publication number: 20130275977Abstract: A method for packing virtual machines onto host devices may calculate scarcity values for several different parameters. A host's scarcity for a parameter may be determined by multiplying the host's capacity for a parameter with the overall scarcity of that parameter. The sum of a host's scarcity for all the parameters determines the host's overall scarcity. Hosts having the highest scarcity are attempted to be populated with a group of virtual machines selected for compatibility with the host. In many cases, several different scenarios may be evaluated and an optimal scenario implemented. The method gives a high priority to those virtual machines that consume scarce resources, with the scarcity being a function of the available hardware and the virtual machines that may be placed on them.Type: ApplicationFiled: June 10, 2013Publication date: October 17, 2013Inventors: Lincoln K. Uyeda, Rina Panigrafy, Ehud Wieder, Kunal Talwar
-
Patent number: 8464267Abstract: A method for packing virtual machines onto host devices may calculate scarcity values for several different parameters. A host's scarcity for a parameter may be determined by multiplying the host's capacity for a parameter with the overall scarcity of that parameter. The sum of a host's scarcity for all the parameters determines the host's overall scarcity. Hosts having the highest scarcity are attempted to be populated with a group of virtual machines selected for compatibility with the host. In many cases, several different scenarios may be evaluated and an optimal scenario implemented. The method gives a high priority to those virtual machines that consume scarce resources, with the scarcity being a function of the available hardware and the virtual machines that may be placed on them.Type: GrantFiled: April 10, 2009Date of Patent: June 11, 2013Assignee: Microsoft CorporationInventors: Lincoln K. Uyeda, Rina Panigrahy, Ehud Wieder, Kunal Talwar
