Sicheres Verfahren zur Validierung von Maschinenlernmodellen mittels Homomorpher Verschlüsselungstechniken

Details

Veröffentlichungs-Nr.

EP4095769

Aktenzeichen

EP21175794

Anmeldetag

25. Mai 2021

Veröffentlichung

18. Oktober 2023

Erteilung

18. Oktober 2023

Rechtsraum

EP

IPC

G06N20/00H04L9/00

Abstract

The invention relates to a method (100) for secure validation of machine learning models and parallel validation data using homomorphic encryption, comprising the steps: providing, by a provider, a machine learning model (101) and providing, by a user, validation data (102); encrypting, by the provider, the machine learning model (103); sending, by the provider, a public encryption parameter to the user (104); selecting, by the user and provider, a unifying encoding method (105); encrypting, by the user, the validation data (106); sending, by the user, the encrypted validation data (107); processing, the encrypted validation data with the encrypted machine learning model (108); providing encrypted results of said processing to the provider and the user (109); decrypting the results and evaluating whether the performance of the machine learning model is satisfactory with the given valuation data of the user (110).

Anmelder

Firma

Unify Patente GmbH & Co. KG

Land

🇩🇪 Deutschland

Fachgebiete