US 12,079,814 B2
Privacy-preserving graph compression with automated fuzzy variable detection
Theodore Harris, San Francisco, CA (US); Tatiana Korolevskaya, Mountain View, CA (US); and Yue Li, San Mateo, CA (US)
Assigned to Visa International Service Association, San Francisco, CA (US)
Appl. No. 17/256,945
Filed by Visa International Service Association, San Francisco, CA (US)
PCT Filed Jul. 11, 2019, PCT No. PCT/US2019/041318
§ 371(c)(1), (2) Date Dec. 29, 2020,
PCT Pub. No. WO2020/014425, PCT Pub. Date Jan. 16, 2020.
Claims priority of provisional application 62/696,480, filed on Jul. 11, 2018.
Prior Publication US 2021/0272121 A1, Sep. 2, 2021
Int. Cl. G06Q 20/40 (2012.01); G06F 21/62 (2013.01); G06N 3/126 (2023.01); G06N 5/048 (2023.01); G06N 7/02 (2006.01); G06N 20/10 (2019.01); G06Q 20/38 (2012.01)
CPC G06Q 20/4016 (2013.01) [G06F 21/6245 (2013.01); G06N 3/126 (2013.01); G06N 5/048 (2013.01); G06N 7/023 (2013.01); G06N 20/10 (2019.01); G06Q 20/38 (2013.01)] 18 Claims
OG exemplary drawing
 
1. A method comprising:
a) receiving, by a first server computer, network data comprising a plurality of transaction data for a plurality of transactions, wherein each transaction data comprises a plurality of data elements with data values, wherein at least one of the plurality of data elements comprises a user identifier for a user;
b) generating, by the first server computer, one or more graphs comprising a plurality of communities based on the network data;
c) anonymizing the network data by determining, by the first server computer, fuzzy values for at least some of the data values for each transaction of the plurality of transactions, wherein determining fuzzy values for at least some of the data values comprises:
determining, by the first server computer, a set of membership functions for each data element corresponding to the at least some of the data values, wherein at least one membership function of the set of membership functions overlaps with a second membership function of the set of membership functions and a percentage of overlap between the at least one membership function and the second membership function is greater than a predetermined overlap threshold; and
determining, by the first server computer, the fuzzy values for the at least some of the data values using the set of membership functions;
d) for each user, determining, by the first server computer, fuzzy values for communities within the plurality of communities;
e) transmitting, by the first server computer to a second server computer, the fuzzy values obtained in steps c) and d), wherein the second server computer is configured to return a determination that the fuzzy values obtained in steps c) and d) are privacy-preserving of the plurality of transaction data;
f) receiving, from the second server computer, a responsive message indicating at least one of the fuzzy values obtained in steps c) and d) is not privacy-preserving;
g) determining, by the first server computer, an updated set of membership functions and using the updated set of membership functions to generate updated fuzzy values for at least some of the data values;
h) generating, by the first server computer, a model using the fuzzy values obtained in step c) and the updated fuzzy values obtained in step g), and at least some of the data values; and
i) responsive to a request from an evaluation computer, providing a prediction of a fraudulent transaction to the evaluation computer based on the model generated using the anonymized network data.