Föderiertes Lernen mit Privatsphärenerhaltung von Kundendaten und Vorbedingungenprüfung

Details

Veröffentlichungs-Nr.

EP4787245

Anmeldetag

30. Januar 2025

Veröffentlichung

5. August 2026

Rechtsraum

EP

Abstract

There is disclosed a method for training and using a ML model on customer engineering data with increased customer privacy preservation in an industrial context. The method comprises, in a training phase of the ML model: obtaining the customer engineering data from customer project data associated with a customer; training the ML model on the customer engineering data by using federated ML; and applying customer privacy preservation mechanisms on information inferred by the ML model. The method further comprises, when using the ML model for inference: verifying, by using an Entailment Checking method, that the information inferred by the ML model and determined by the ML model to be associated with the customer were obtained from the customer project data.

Anmelder

Firma

ABB SCHWEIZ AG

Land

🇨🇭 Schweiz