Data processing method and device, electronic equipment and storage medium

By combining the Cuckoo Hash and OPRF algorithms with secret sharing technology, data is partitioned and hashed, solving the problems of low efficiency and insufficient security in privacy set union operations. This enables efficient and secure parallel processing of datasets, making it suitable for federated learning environments.

CN115757624BActive Publication Date: 2026-07-07BAIDU INT TECH (SHENZHEN) CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
BAIDU INT TECH (SHENZHEN) CO LTD
Filing Date
2022-11-17
Publication Date
2026-07-07

AI Technical Summary

Technical Problem

In existing technologies, the development of privacy set union operations in federated learning has been relatively slow. It has high computational complexity and poses a risk of data leakage, especially when the dataset is small or the participants have insufficient data, making public key operations a bottleneck.

Method used

The data is segmented and hashed using the Cuckoo Hash algorithm and the Unintentional Pseudo-Random Function (OPRF) combined with the Secret Sharing algorithm. A distributed computing platform is used to achieve parallel transmission and processing of the data. The union is determined by transmitting selected bits of data, thus avoiding the direct exposure of the original data.

Benefits of technology

It improves the efficiency of union operations on privacy-preserving data, ensures data security, reduces computational complexity, and enables efficient parallel processing of datasets in untrusted environments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115757624B_ABST
    Figure CN115757624B_ABST
Patent Text Reader

Abstract

The present disclosure provides a data processing method, relates to the technical field of artificial intelligence, and particularly relates to the technical field of computer security, deep learning, federated learning and the like. The specific implementation scheme is: first source data of a data receiver is split to obtain at least one first sub-data; first hash data is generated according to the first sub-data; the first hash data is split to obtain a split result, wherein the split result comprises first sub-hash data and second sub-hash data; a comparison result is determined according to a second preset function output set corresponding to the first sub-hash data and a first preset function output set corresponding to the second sub-hash data; and a union set data corresponding to the first source data is determined according to the first sub-data, second target sub-data and the comparison result, wherein the second preset function output set is from a data sender. The present disclosure also provides a data processing device, an electronic device and a storage medium.
Need to check novelty before this filing date? Find Prior Art