US 12,254,121 B2
Data processing method and apparatus, device, and computer-readable storage medium
Fangcheng Fu, Shenzhen (CN); Jie Jiang, Shenzhen (CN); Junwei Pan, Shenzhen (CN); Chen Hou, Shenzhen (CN); Huanran Xue, Shenzhen (CN); Yong Cheng, Shenzhen (CN); Yuhong Liu, Shenzhen (CN); Peng Chen, Shenzhen (CN); and Yangyu Tao, Shenzhen (CN)
Assigned to TENCENT TECHNOLOGY (SHENZHEN) COMPANY LTD, Shenzhen (CN)
Filed by Tencent Technology (Shenzhen) Company Limited, Guangdong (CN)
Filed on Nov. 30, 2022, as Appl. No. 18/072,313.
Application 18/072,313 is a continuation of application No. PCT/CN2022/089798, filed on Apr. 28, 2022.
Claims priority of application No. 202110576191.6 (CN), filed on May 26, 2021.
Prior Publication US 2023/0108682 A1, Apr. 6, 2023
Int. Cl. G06F 21/64 (2013.01); G06F 21/60 (2013.01)
CPC G06F 21/645 (2013.01) [G06F 21/602 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A data processing method, performed by a first node comprising a computer device, the data processing method comprising:
acquiring a first intersection set comprising N pieces of first intersection data, the first intersection data comprising a first plurality of user identifiers from raw data from a storage of the computer device, and N being a positive integer;
acquiring a second intersection set comprising M pieces of second intersection data, the second intersection data comprising a second plurality of user identifiers received from a second node, and M being a positive integer;
calculating an intersection between the first intersection set and the second intersection set to obtain an intersection result set that includes an intersecting portion of the first intersection data and the second intersection data;
obfuscating the intersection result set to obtain an obfuscation set that comprises:
obfuscated data based on data in the second intersection set, and
an intersection data set based on the intersection result set,
wherein the obfuscated data is not correlated with data in the intersection data set;
generating a first training data set based on the obfuscation set, wherein the first training data set comprises a plurality of obfuscated user identifiers of the obfuscation set aligned with first user data;
calculating a first gradient of the first training data set;
transmitting the obfuscation set to the second node, such that the raw data is not exposed to the second node, to cause the second node to calculate and transmit a second gradient to a third node;
transmitting the first gradient to the third node; and
receiving from the third node a total gradient based on the first gradient and the second gradient.