Bandwidth controlled multi-party joint data processing methods and apparatuses

Patent metadata

Field	Value
Publication number	US-12574356-B2
Application number	US-202318482706-A
Country	US
Kind code	B2
Filing date	Oct 6, 2023
Priority date	Apr 8, 2021
Publication date	Mar 10, 2026
Grant date	Mar 10, 2026

Abstract

Bandwidth-controlled, private, multi-party joint data processing is described. When performing an n th iteration of joint data processing on service data, multiple participants separately determine a same n th identifier at the current iteration and includes sub-identifiers separately corresponding to service data of the multiple participants. The multiple participants are enabled to separately determine their respective random arrays in a predetermined trusted interaction manner and based on the n th identifier and initial seeds of the multiple participants. When a first sub-identifier exists in a correspondence between sub-identifiers and confidential data, the multiple participants separately obtain first confidential data corresponding to the first sub-identifier and perform the current iteration of joint data processing with another participant using secure multi-party computation (MPC) based on exchange of other confidential data and respective first confidential data. The first confidential data is obtained by fusing first service data and a corresponding first random number.

First claim

What is claimed is:

1 . A computer-implemented method for bandwidth-controlled, private, multi-party joint data processing, comprising:
when an n th iteration of joint data processing on service data of multiple participants is performed, determining a same n th identifier of the multiple participants at a current iteration of joint data processing, wherein the service data of multiple participants comprises feature data of multiple objects, and wherein the same n th identifier comprises sub-identifiers respectively corresponding to the service data of multiple participants; in a predetermined trusted interaction manner and based on the same n th identifier and initial seeds of the multiple participants, the multiple participants separately determine respective random arrays, wherein the respective random arrays comprise multiple random numbers respectively corresponding to multiple sub-identifiers; when a first sub-identifier in the same n th identifier exists in a correspondence between sub-identifiers and confidential data, obtaining first confidential data corresponding to the first sub-identifier, wherein the first confidential data is obtained by fusing first service data and a corresponding first random number in advance, and wherein, before obtaining first confidential data corresponding to the first sub-identifier: separately matching the sub-identifiers in the same n th identifier against corresponding sub-identifiers in the correspondence; and
determining a successfully matched sub-identifier as the first sub-identifier; and
performing the current iteration of joint data processing with another participant using secure multi-party computation (MPC) based on exchange of other confidential data and respective first confidential data, wherein other confidential data is obtained by fusing other service data and a corresponding random number.

2 . The computer-implemented method of claim 1 , wherein:
the feature data of multiple objects comprises at least one of: basic attribute information, association relationship information, interaction information, and historical behavior information; and an object comprises one of: a user, a product, and an event.

3 . The computer-implemented method of claim 1 , wherein the first service data is service data of a participant or service data of the other participant.

4 . The computer-implemented method of claim 1 , wherein the correspondence is obtained based on exchange of confidential data between the multiple participants in an m th iteration of joint data processing process that has been performed.

5 . The computer-implemented method of claim 1 , wherein, when a second sub-identifier in the same n th identifier does not exist in the correspondence and the second sub-identifier corresponds to second service data:
obtaining second confidential data, wherein the second confidential data is obtained by fusing the second service data and a second random number corresponding to the second sub-identifier; and adding the second sub-identifier and corresponding second confidential data to the correspondence.

6 . The computer-implemented method of claim 1 , wherein when the sub-identifiers in the same n th identifier do not exist in the correspondence:
determining confidential data by fusing service data of a participant and a corresponding random number in a random array of the participant; and performing the current iteration of joint data processing by using the MPC based on exchange between the confidential data and confidential data of the other participant.

7 . The computer-implemented method of claim 1 , wherein the first confidential data is obtained based on a difference between the first service data and the corresponding first random number.

8 . The computer-implemented method of claim 1 , wherein the step of determining a same n th identifier of the multiple participants at a current iteration of joint data processing, comprises:
when service data of a participant in the current iteration of joint data processing are the same as service data of the participant in joint data processing that has been performed, determining a sub-identifier corresponding to the service data in the joint data processing that has been performed as a sub-identifier of the service data at the current iteration of joint data processing; and when the service data of the participant in the current iteration of joint data processing are not the same as the service data of the participant in the joint data processing that has been performed, determining the sub-identifier of the service data at the current iteration of joint data processing based on a value different from the sub-identifier corresponding to the service data in the joint data processing that has been performed.

9 . The computer-implemented method of claim 8 , wherein the step of determining the sub-identifier of the service data at the current iteration of joint data processing based on a value different from the sub-identifier corresponding to the service data in the joint data processing that has been performed, comprises: determining the sub-identifier of the service data at the current iteration of joint data processing based on an updated value of a counter corresponding to the service data, wherein the multiple participants have a same initial value of the counter corresponding to the service data, and update modes are the same.

10 . The computer-implemented method of claim 1 , wherein the predetermined trusted interaction manner, comprises:
an interaction manner between the multiple participants using MPC; or an interaction manner between the multiple participants using a trusted third party.

11 . The computer-implemented method of claim 10 , wherein, in an interaction manner using a trusted third party, enabling the multiple participants to separately determine respective random arrays, comprises:
generating an obtaining request carrying the same n th identifier, and sending the obtaining request to the trusted third party; and receiving a random array returned by the trusted third party for a participant, wherein the trusted third party separately determines corresponding random arrays for the multiple participants based on the same n th identifier, the respective initial seeds of the multiple participants, and the predetermined trusted interaction manner.

12 . The computer-implemented method of claim 10 , wherein, in an interaction manner using a trusted third party, the trusted third party comprises several computing units; and the computing units comprise a trusted computing unit that has a trusted execution environment (TEE).

13 . The computer-implemented method of claim 10 , wherein, in an interaction manner using a trusted third party, enabling the multiple participants to separately determine respective random arrays, comprises:
when a participant is a selected participant, generating a random array by using the same n th identifier and an initial seed of the participant, wherein the random array comprises a to-be-corrected random number; and sending an obtaining request carrying the same n th identifier to the trusted third party, receiving a correction value that is of the to-be-corrected random number and returned by the trusted third party, and updating a locally generated random array by using the correction value, wherein the correction value is determined based on the same n th identifier and respective initial seeds of the multiple participants; and when the participant is not a selected participant, generating the random array by using the same n th identifier and the initial seed of

Assignees

• Alipay Hangzhou Inf Tech Co Ltd

Inventors

• Yu Chaofan
• Wang Lei

Classifications

• G06F7/588

  Random number generators, i.e. based on natural stochastic processes · CPC title

• H04L63/0421

  Anonymous communication, i.e. the party's identifiers are hidden from the other party or parties, e.g. using an anonymizer · CPC title

• H04L2209/46

  Secure multiparty computation, e.g. millionaire problem · CPC title

• G06F21/6245

  Protecting personal data, e.g. for financial or medical purposes · CPC title

• H04L63/0407 Primary

  wherein the identity of one or more communicating identities is hidden (cryptographic mechanisms or cryptographic arrangements for anonymous credentials or for identity based cryptographic systems H04L9/00) · CPC title