US 12,073,301 B2
Learning device, and learning method
Takuya Tanaka, Tokyo (JP); and Ryosuke Kasahara, Kanagawa (JP)
Assigned to RICOH COMPANY, LTD., Tokyo (JP)
Appl. No. 15/734,619
Filed by RICOH COMPANY, LTD., Tokyo (JP)
PCT Filed Jun. 7, 2019, PCT No. PCT/JP2019/022834
§ 371(c)(1), (2) Date Dec. 3, 2020,
PCT Pub. No. WO2019/235636, PCT Pub. Date Dec. 12, 2019.
Claims priority of application No. 2018-109735 (JP), filed on Jun. 7, 2018.
Prior Publication US 2021/0158224 A1, May 27, 2021
Int. Cl. G06N 20/20 (2019.01); G06N 3/08 (2023.01); G06N 5/01 (2023.01); G06N 20/00 (2019.01); G06V 10/776 (2022.01); G06V 10/82 (2022.01)
CPC G06N 20/20 (2019.01) [G06N 5/01 (2023.01); G06N 3/08 (2013.01); G06N 20/00 (2019.01); G06V 10/776 (2022.01); G06V 10/82 (2022.01)] 10 Claims
OG exemplary drawing
 
1. A learning device configured to perform learning by gradient boosting, comprising:
a plurality of data memories configured to store learning data including at least one type of feature amount and gradient information corresponding to the learning data;
a plurality of gradient outputters provided corresponding to the plurality of data memories, and each gradient outputter is configured to receive an input of the feature amount and the gradient information corresponding to the feature amount from a corresponding one of the plurality of data memories, and output the gradient information corresponding to the feature amount through an output port corresponding to each value of the input feature amount;
an adder configured to add up one or more pieces of the gradient information corresponding to a same value of the feature amount among pieces of the gradient information output from the plurality of gradient outputters, and output an added value of the gradient information corresponding to each value of the feature amount; and
a histogram memory configured to store a histogram obtained by integrating added values of the gradient information corresponding to each value of the feature amount output from the adder for each bin in a case in which the each value is set as a bin.