| CPC H04L 9/008 (2013.01) [G06F 16/2468 (2019.01); H04L 9/0618 (2013.01)] | 16 Claims |
|
1. A computer-implemented cryptographic method for facilitating a data similarity query across a network of data memories, the method comprising:
receiving, from a first memory in the network, a first ciphertext for a first aggregation of first encodings of data items of a first group, the first group of data items being of a first party;
receiving, from a second memory in the network, a second ciphertext for a second aggregation of at least one second encoding of at least one data item of a second group, the second group of data items being of a second party; and
processing the first and second ciphertexts into a third ciphertext, the third ciphertext decryptable into plaintext output, wherein the plaintext output is associable with similarity measures between data items from the two groups, wherein the similarity measures establish a similarity between the data items of the first group and the at least one data item of the second group as held in the data memories,
wherein the first encodings and the at least one second encoding are, by representation, in a respective hash-based probabilistic data structure, PDS,
the first and second aggregations previously obtained by forming respective matrices from the said first encodings and the at least one second encoding,
the first ciphertext previously obtained through homomorphic encryption of the first aggregation of the first encodings, the second ciphertext previously obtained through homomorphic encryption of the second aggregation of the second encoding,
and further wherein the processing of the first and the second ciphertexts includes a matrix multiplication of the at least two matrices.
|