Patents by Inventor Han Young YIM
Han Young YIM 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: 11823028Abstract: An artificial neural network (ANN) quantization method for generating an output ANN by quantizing an input ANN includes: obtaining second parameters by quantizing first parameters of the input ANN; obtaining a sample distribution from an intermediate ANN in which the obtained second parameters have been applied to the input ANN; and obtaining a fractional length for the sample distribution by quantizing the obtained sample distribution.Type: GrantFiled: July 24, 2018Date of Patent: November 21, 2023Assignee: SAMSUNG ELECTRONICS CO., LTD.Inventors: Do-yun Kim, Han-young Yim, Byeoung-su Kim, Nak-woo Sung, Jong-han Lim, Sang-hyuck Ha
-
Patent number: 11694073Abstract: A method and apparatus for generating a fixed point neural network are provided. The method includes selecting at least one layer of a neural network as an object layer, wherein the neural network includes a plurality of layers, each of the plurality of layers corresponding to a respective one of plurality of quantization parameters; forming a candidate parameter set including candidate parameter values with respect to a quantization parameter of the plurality of quantization parameters corresponding to the object layer; determining an update parameter value from among the candidate parameter values based on levels of network performance of the neural network, wherein each of the levels of network performance correspond to a respective one of the candidate parameter values; and updating the quantization parameter with respect to the object layer based on the update parameter value.Type: GrantFiled: November 20, 2018Date of Patent: July 4, 2023Assignee: SAMSUNG ELECTRONICS CO., LTD.Inventors: Han-young Yim, Do-yun Kim, Byeoung-su Kim, Nak-Woo Sung, Jong-han Lim, Sang-hyuck Ha
-
Publication number: 20220207373Abstract: An operation method of a computing device includes selecting first data on which a first pruning is to be performed, down-scaling a first plurality of weights included in a first output channel associated with the first data, up-scaling a second plurality of weights used to generate second data to be multiplied by a weight having a major value from among the first plurality of weights included in the first output channel, calculating the second data based on the up-scaled second plurality of weights, and performing the first pruning.Type: ApplicationFiled: June 16, 2021Publication date: June 30, 2022Inventors: JUHEE PARK, HAN YOUNG YIM, SANGHYUCK HA
-
Patent number: 11373087Abstract: A method of generating a fixed-point type neural network by quantizing a floating-point type neural network, includes obtaining, by a device, a plurality of post-activation values by applying an activation function to a plurality of activation values that are received from a layer included in the floating-point type neural network, and deriving, by the device, a plurality of statistical characteristics for at least some of the plurality of post-activation values. The method further includes determining, by the device, a step size for the quantizing of the floating-point type neural network, based on the plurality of statistical characteristics, and determining, by the device, a final fraction length for the fixed-point type neural network, based on the step size.Type: GrantFiled: July 12, 2018Date of Patent: June 28, 2022Assignee: SAMSUNG ELECTRONICS CO., LTD.Inventors: Han-young Yim, Do-yun Kim, Byeoung-su Kim, Nak-woo Sung, Jong-han Lim, Sang-hyuck Ha
-
Patent number: 11275986Abstract: A method of quantizing an artificial neural network includes dividing an input distribution of the artificial neural network into a plurality of segments, generating an approximated density function by approximating each of the plurality of segments, calculating at least one quantization error corresponding to at least one step size for quantizing the artificial neural network, based on the approximated density function, and determining a final step size for quantizing the artificial neural network based on the at least one quantization error.Type: GrantFiled: June 14, 2018Date of Patent: March 15, 2022Assignee: SAMSUNG ELECTRONICS CO., LTD.Inventors: Do-yun Kim, Han-young Yim, In-yup Kang, Byeoung-su Kim, Nak-woo Sung, Jong-Han Lim, Sang-hyuck Ha
-
Patent number: 10854215Abstract: A residual echo estimator estimates a residual echo of a microphone signal input to a microphone as a reference signal is output through a speaker. The residual echo estimator estimates a magnitude of a residual echo of the reference signal at a current frame by using a magnitude of the linear echo of the reference signal at the current frame and a magnitude of the linear echo of the reference signal at a past frame, and updates weights to be applied to the magnitudes of the linear echo.Type: GrantFiled: August 21, 2019Date of Patent: December 1, 2020Assignees: Samsung Electronics Co., Ltd., Gwangju Institute of Science and TechnologyInventors: Seongwoo Kim, Han Young Yim, Dongil Hyun, Jong Won Shin, Hyungchan Song
-
Publication number: 20200243104Abstract: A residual echo estimator estimates a residual echo of a microphone signal input to a microphone as a reference signal is output through a speaker. The residual echo estimator estimates a magnitude of a residual echo of the reference signal at a current frame by using a magnitude of the linear echo of the reference signal at the current frame and a magnitude of the linear echo of the reference signal at a past frame, and updates weights to be applied to the magnitudes of the linear echo.Type: ApplicationFiled: August 21, 2019Publication date: July 30, 2020Applicants: Samsung Electronics Co., Ltd., GWANGJU INSTITUTE OF SCIENCE AND TECHNOLOGYInventors: Seongwoo KIM, Han Young YIM, Dongil HYUN, Jong Won SHIN, Hyungchan SONG
-
Publication number: 20190180177Abstract: A method and apparatus for generating a fixed point neural network are provided. The method includes selecting at least one layer of a neural network as an object layer, wherein the neural network includes a plurality of layers, each of the plurality of layers corresponding to a respective one of plurality of quantization parameters; forming a candidate parameter set including candidate parameter values with respect to a quantization parameter of the plurality of quantization parameters corresponding to the object layer; determining an update parameter value from among the candidate parameter values based on levels of network performance of the neural network, wherein each of the levels of network performance correspond to a respective one of the candidate parameter values; and updating the quantization parameter with respect to the object layer based on the update parameter value.Type: ApplicationFiled: November 20, 2018Publication date: June 13, 2019Applicant: SAMSUNG ELECTRONICS CO., LTD.Inventors: Han-young Yim, Do-yun Kim, Byeoung-su Kim, Nak-woo Sung, Jong-han Lim, Sang-hyuck Ha
-
Publication number: 20190147322Abstract: An artificial neural network (ANN) quantization method for generating an output ANN by quantizing an input ANN includes: obtaining second parameters by quantizing first parameters of the input ANN; obtaining a sample distribution from an intermediate ANN in which the obtained second parameters have been applied to the input ANN; and obtaining a fractional length for the sample distribution by quantizing the obtained sample distribution.Type: ApplicationFiled: July 24, 2018Publication date: May 16, 2019Applicant: SAMSUNG ELECTRONICS CO., LTD.Inventors: Do-yun KIM, Han-young YIM, Byeoung-su KIM, Nak-woo SUNG, Jong-han LIM, Sang-hyuck HA
-
Publication number: 20190130255Abstract: A method of generating a fixed-point type neural network by quantizing a floating-point type neural network, includes obtaining, by a device, a plurality of post-activation values by applying an activation function to a plurality of activation values that are received from a layer included in the floating-point type neural network, and deriving, by the device, a plurality of statistical characteristics for at least some of the plurality of post-activation values. The method further includes determining, by the device, a step size for the quantizing of the floating-point type neural network, based on the plurality of statistical characteristics, and determining, by the device, a final fraction length for the fixed-point type neural network, based on the step size.Type: ApplicationFiled: July 12, 2018Publication date: May 2, 2019Applicant: SAMSUNG ELECTRONICS CO., LTD.Inventors: Han-young Yim, Do-yun Kim, Byeoung-su Kim, Nak-woo Sung, Jong-han Lim, Sang-hyuck Ha
-
Publication number: 20190095777Abstract: A method of quantizing an artificial neural network includes dividing an input distribution of the artificial neural network into a plurality of segments, generating an approximated density function by approximating each of the plurality of segments, calculating at least one quantization error corresponding to at least one step size for quantizing the artificial neural network, based on the approximated density function, and determining a final step size for quantizing the artificial neural network based on the at least one quantization error.Type: ApplicationFiled: June 14, 2018Publication date: March 28, 2019Applicant: SAMSUNG ELECTRONICS CO., LTD.Inventors: Do-yun KIM, Han-young YIM, In-yup KANG, Byeoung-su KIM, Nak-woo SUNG, Jong-Han LIM, Sang-hyuck HA
-
Patent number: 9894658Abstract: Provided is a joint pattern beam sectorization method and apparatuses for performing the same, wherein the joint pattern beam sectorization method including generating, in a service target region, a pattern sector corresponding to an antenna array including antennas having an identical radiation pattern by using the antenna array and generating pattern beam sectors in the pattern sector through a beamforming using the antennas.Type: GrantFiled: October 16, 2015Date of Patent: February 13, 2018Assignee: Korea Advanced Institute of Science and TechnologyInventors: Gye Tae Gil, Ju Yong Lee, Dong Ho Cho, Dae Hee Park, Chang Bae Yoon, Sang Min Oh, Han Young Yim, Yong Hoon Lee
-
Publication number: 20160219567Abstract: Provided is a joint pattern beam sectorization method and apparatuses for performing the same, wherein the joint pattern beam sectorization method including generating, in a service target region, a pattern sector corresponding to an antenna array including antennas having an identical radiation pattern by using the antenna array and generating pattern beam sectors in the pattern sector through a beamforming using the antennas.Type: ApplicationFiled: October 16, 2015Publication date: July 28, 2016Applicant: Korea Advanced Institute of Science and TechnologyInventors: Gye Tae Gil, Ju Yong Lee, Dong Ho Cho, Dae Hee Park, Chang Bae Yoon, Sang Min Oh, Han Young Yim, Yong Hoon Lee
-
Patent number: 9363730Abstract: A method and apparatus for improving a mobile problem caused by a narrow handover region in a boundary region between virtual cells constructed by a plurality of small base stations are provided. A plurality of Virtual Cells (VCs) include a plurality of Distributed Base Stations (DBSs) whose VCs cooperatively communicate with each other. An Intermediate Distributed Base Station (I-DBS) is located in a region where adjacent at least two VCs among the plurality of VCs are superimposed, and belongs to a different VC according to a time division scheme.Type: GrantFiled: October 15, 2013Date of Patent: June 7, 2016Assignees: Samsung Electronics Co., Ltd., Korea Advanced Institute of Science and TechnologyInventors: Hyun-Jeong Kang, Tae-Soo Kwon, Ju Yong Lee, Han Young Yim, Tae-Young Kim
-
Publication number: 20140105050Abstract: A method and apparatus for improving a mobile problem caused by a narrow handover region in a boundary region between virtual cells constructed by a plurality of small base stations are provided. A plurality of Virtual Cells (VCs) include a plurality of Distributed Base Stations (DBSs) whose VCs cooperatively communicate with each other. An Intermediate Distributed Base Station (I-DBS) is located in a region where adjacent at least two VCs among the plurality of VCs are superimposed, and belongs to a different VC according to a time division scheme.Type: ApplicationFiled: October 15, 2013Publication date: April 17, 2014Applicants: Korea Advanced Institute of Science and Technology, Samsung Electronics Co., Ltd.Inventors: Hyun-Jeong KANG, Tae-Soo KWON, Ju Yong LEE, Han Young YIM, Tae-Young KIM