US Patent Issued to International Business Machines on Aug. 25 for "Training neural networks with non-polynomial elements for homomorphic encryption computations using sub-networks and multi-loss" (Israeli Inventors)

ALEXANDRIA, Va., Aug. 25 -- United States Patent no. 12,719,678, issued on Aug. 25, was assigned to International Business Machines Corp. (Armonk, N.Y.).

"Training neural networks with non-polynomial elements for homomorphic encryption computations using sub-networks and multi-loss" was invented by Itamar Zimerman (Tel-Aviv, Israel) and Nir Drucker (Zichron Yaakov, Israel).

According to the abstract* released by the U.S. Patent & Trademark Office: "An example system includes a processor to receive a non-homomorphic encryption (HE)-friendly analytics model including a non-polynomial element. The processor is to train a substitution model in which the non-polynomial element of the non-homomorphic encryption (HE)-friendly analytics model is ...