面向多源异构数据融合的隐私集合求交研究

Multi-Source Heterogeneous Data Collaboration via Private Set Intersection

多源异构数据融合的痛点在于数据的低价值密度性和分散性。数据的多源异构性增加了数据聚合的难度,导致数据价值极度零散,使得数据融合方法面对多源异构大数据无的放矢,无法有效关联零散价值的数据。隐私集合求交(Private Set Intersection,PSI)不但可以使数据方放心提供数据,还可以将多源异构数据价值有效融合,是挖掘有效数据开展数据融合工作的新工具。为此,文章针对异构数据的整合、数据的多源以及大规模数据的并行处理3类问题,给出多源异构数据融合的3个新思路。

The main point of multi-source heterogeneous data fusion is the low value density and dispersion of data.The multi-source heterogeneity of data increases the difficulty of data aggregation,leading to extreme fragmentation of data value,making data fusion methods to face multi-source heterogeneous big data with no target,and unable to effectively correlate data with fragmented value.Private Set Intersection(PSI)not only enables data providers to provide data with peace of mind,but also effectively integrates the value of heterogeneous data from multiple sources,and mines effective data to carry out data fusion work as a new tool.To this end,the article gave three new ideas for the fusion of heterogeneous data from multiple sources with respect to three types of problems:integration of heterogeneous data,multiple sources of data,and parallel processing of large-scale data.