US 12,149,607 B2
Graph based embedding under fully homomorphic encryption
Allon Adir, Kiryat Tivon (IL); Ramy Masalha, Kafr Qari (IL); Eyal Kushnir, Kfar Vradim (IL); Omri Soceanu, Haifa (IL); Ehud Aharoni, Kfar Saba (IL); Nir Drucker, Zichron Yaakov (IL); and Guy Moshkowich, Nes Ziyona (IL)
Assigned to International Business Machines Corporation, Armonk, NY (US)
Filed by International Business Machines Corporation, Armonk, NY (US)
Filed on Oct. 10, 2022, as Appl. No. 17/962,843.
Prior Publication US 2024/0121074 A1, Apr. 11, 2024
Int. Cl. H04L 9/00 (2022.01); G06F 9/38 (2018.01); H04L 9/40 (2022.01)
CPC H04L 9/008 (2013.01) [G06F 9/3887 (2013.01); H04L 63/0428 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method, in a data processing system, for fully homomorphic encryption enabled graph embedding, the method comprising:
receiving an encrypted graph data structure comprising a plurality of encrypted entities and a plurality of encrypted predicates;
generating, for each encrypted entity in the plurality of encrypted entities, a corresponding set of entity ciphertexts based on an initial embedding of entity features of the encrypted entity;
generating, for each encrypted predicate in the plurality of encrypted predicates, a corresponding predicate ciphertext based on an initial embedding of predicate features of the encrypted predicate;
iteratively executing a machine learning process, on the sets of entity ciphertexts and the predicate ciphertexts to update embeddings of the entity features of the encrypted entities and update embeddings of predicate features of the encrypted predicates, to generate a computer model for embedding entities and predicates; and
outputting a final embedding based on the updated embeddings of the entity features and predicate features of the computer model.