Patents by Inventor George Ke
George Ke 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: 20220121993Abstract: A disclosed example includes implementing a first worker instance and a second worker instance to operate in parallel running a first tuning operation via the first worker instance to tune first hyperparameters; running a second tuning operation via the second worker instance using a Bayesian-based optimization to determine a hyperparameter configuration to evaluate next; evaluating the hyperparameter configuration for an external model using a surrogate model; and selecting the hyperparameter configuration for the external model.Type: ApplicationFiled: December 23, 2021Publication date: April 21, 2022Inventors: Patrick Hayes, Michael McCourt, Alexandra Johnson, George Ke, Scott Clark
-
Patent number: 11301781Abstract: A system and method includes receiving a tuning work request for tuning an external machine learning model; implementing a plurality of distinct queue worker machines that perform various tuning operations based on the tuning work data of the tuning work request; implementing a plurality of distinct tuning sources that generate values for each of the one or more hyperparameters of the tuning work request; selecting, by one or more queue worker machines of the plurality of distinct queue worker machines, one or more tuning sources of the plurality of distinct tuning sources for tuning the one or more hyperparameters; and using the selected one or more tuning sources to generate one or more suggestions for the one or more hyperparameters, the one or more suggestions comprising values for the one or more hyperparameters of the tuning work request.Type: GrantFiled: February 20, 2020Date of Patent: April 12, 2022Assignee: Intel CorporationInventors: Patrick Hayes, Michael McCourt, Alexandra Johnson, George Ke, Scott Clark
-
Publication number: 20200202254Abstract: A system and method includes receiving a tuning work request for tuning an external machine learning model; implementing a plurality of distinct queue worker machines that perform various tuning operations based on the tuning work data of the tuning work request; implementing a plurality of distinct tuning sources that generate values for each of the one or more hyperparameters of the tuning work request; selecting, by one or more queue worker machines of the plurality of distinct queue worker machines, one or more tuning sources of the plurality of distinct tuning sources for tuning the one or more hyperparameters; and using the selected one or more tuning sources to generate one or more suggestions for the one or more hyperparameters, the one or more suggestions comprising values for the one or more hyperparameters of the tuning work request.Type: ApplicationFiled: February 20, 2020Publication date: June 25, 2020Inventors: Patrick Hayes, Michael McCourt, Alexandra Johnson, George Ke, Scott Clark
-
Patent number: 10607159Abstract: A system and method includes receiving a tuning work request for tuning an external machine learning model; implementing a plurality of distinct queue worker machines that perform various tuning operations based on the tuning work data of the tuning work request; implementing a plurality of distinct tuning sources that generate values for each of the one or more hyperparameters of the tuning work request; selecting, by one or more queue worker machines of the plurality of distinct queue worker machines, one or more tuning sources of the plurality of distinct tuning sources for tuning the one or more hyperparameters; and using the selected one or more tuning sources to generate one or more suggestions for the one or more hyperparameters, the one or more suggestions comprising values for the one or more hyperparameters of the tuning work request.Type: GrantFiled: January 9, 2019Date of Patent: March 31, 2020Assignee: SigOpt, Inc.Inventors: Patrick Hayes, Michael McCourt, Alexandra Johnson, George Ke, Scott Clark
-
Publication number: 20190147362Abstract: A system and method includes receiving a tuning work request for tuning an external machine learning model; implementing a plurality of distinct queue worker machines that perform various tuning operations based on the tuning work data of the tuning work request; implementing a plurality of distinct tuning sources that generate values for each of the one or more hyperparameters of the tuning work request; selecting, by one or more queue worker machines of the plurality of distinct queue worker machines, one or more tuning sources of the plurality of distinct tuning sources for tuning the one or more hyperparameters; and using the selected one or more tuning sources to generate one or more suggestions for the one or more hyperparameters, the one or more suggestions comprising values for the one or more hyperparameters of the tuning work request.Type: ApplicationFiled: January 9, 2019Publication date: May 16, 2019Inventors: Patrick Hayes, Michael McCourt, Alexandra Johnson, George Ke, Scott Clark
-
Patent number: 10217061Abstract: A system and method includes receiving a tuning work request for tuning an external machine learning model; implementing a plurality of distinct queue worker machines that perform various tuning operations based on the tuning work data of the tuning work request; implementing a plurality of distinct tuning sources that generate values for each of the one or more hyperparameters of the tuning work request; selecting, by one or more queue worker machines of the plurality of distinct queue worker machines, one or more tuning sources of the plurality of distinct tuning sources for tuning the one or more hyperparameters; and using the selected one or more tuning sources to generate one or more suggestions for the one or more hyperparameters, the one or more suggestions comprising values for the one or more hyperparameters of the tuning work request.Type: GrantFiled: May 11, 2018Date of Patent: February 26, 2019Assignee: SigOpt, Inc.Inventors: Patrick Hayes, Michael McCourt, Alexandra Johnson, George Ke, Scott Clark
-
Publication number: 20180336493Abstract: A system and method includes receiving a tuning work request for tuning an external machine learning model; implementing a plurality of distinct queue worker machines that perform various tuning operations based on the tuning work data of the tuning work request; implementing a plurality of distinct tuning sources that generate values for each of the one or more hyperparameters of the tuning work request; selecting, by one or more queue worker machines of the plurality of distinct queue worker machines, one or more tuning sources of the plurality of distinct tuning sources for tuning the one or more hyperparameters; and using the selected one or more tuning sources to generate one or more suggestions for the one or more hyperparameters, the one or more suggestions comprising values for the one or more hyperparameters of the tuning work request.Type: ApplicationFiled: May 11, 2018Publication date: November 22, 2018Inventors: Patrick Hayes, Michael McCourt, Alexandra Johnson, George Ke, Scott Clark
-
Patent number: 8592935Abstract: A UV detector is designed to provide a photoresponse with a cutoff wavelength below a predetermined wavelength. The detector uses a sensor element having an active layer comprising a MgS component grown directly on a substrate. A thin layer metal layer is deposited over the active layer and forms a transparent Schottky metal layer.Type: GrantFiled: June 5, 2012Date of Patent: November 26, 2013Assignees: The Hong Kong University of Science and Technology, University of MacauInventors: Iam Keong Sou, Ying Hoi Lai, Shu Kin Lok, Wai Yip Cheung, George Ke Lun Wong, Kam Weng Tam, Sut Kam Ho
-
Publication number: 20120306042Abstract: A UV detector is designed to provide a photoresponse with a cutoff wavelength below a predetermined wavelength. The detector uses a sensor element having an active layer comprising a MgS component grown directly on a substrate. A thin layer metal layer is deposited over the active layer and forms a transparent Schottky metal layer.Type: ApplicationFiled: June 5, 2012Publication date: December 6, 2012Applicant: THE HONG KONG UNIVERSITY OF SCIENCE AND TECHNOLOGYInventors: Iam Keong SOU, Ying Hoi LAI, Shu Kin LOK, Wai Yip CHEUNG, George Ke Lun WONG, Kam Weng TAM, Sut Kam HO
-
Publication number: 20030020021Abstract: The present invention relates to UV detectors comprising undoped Zn1-xMgxS as the UV responsive active material. Where x exceeds 0.3 the thickness of the active material must be below a critical value, for example if 0.30<x<1.00, and the active material is formed as a layer of a thickness t wherein 5000 Å≧t≧100 Å. A particularly preferred combination of x and thickness is x=0.57 and t≦1400 Å because at around these values the UV response of the active material is similar to the UV response of human skin.Type: ApplicationFiled: July 23, 2001Publication date: January 30, 2003Inventors: Iam Keong Sou, Marcus Chi Wai Wu, Kam Sing Wong, George Ke-Lun Wong
-
Patent number: 6362483Abstract: Visible-blind UV detectors are disclosed comprising an active layer of ZnSTe alloy. The Te composition can be varied to provide good lattice matching depending on the nature of the substrate (eg Si, GaP or GaAs) and a novel structure is provided to give high quantum efficiency. The invention also discloses UV detectors with an active layer of pure ZnS and with an active layer of ZnSSe.Type: GrantFiled: December 29, 1998Date of Patent: March 26, 2002Assignee: The Hong Kong University of Science & TechnologyInventors: Iam Keong Sou, Zhaohui Ma, Choi Lai Man, Zhi Yu Yang, Kam Sang Wong, George Ke-Lun Wong