Verfahren, Vorrichtung und Computerlesbares Medium zur Datensatzerweiterung und -Verifizierung unter Verwendung von Merkmalseinbettungstechniken

Details

Veröffentlichungs-Nr.

EP4760595

Anmeldetag

11. Dezember 2024

Veröffentlichung

17. Juni 2026

Rechtsraum

EP

Abstract

Embodiments of the present disclosure relate to a method for generating training data for an application-specific task. A training method and a dataset used for training a machine learning model are also disclosed. A machine learning model, a computer program, a computer-readable data carrier and an apparatus are also presented. The method for generating the training data for the application-specific task comprises: fine-tuning a pre-trained machine learning model using a first dataset; obtaining, using said model, original embeddings for input data obtained from a second application-specific dataset; applying, to each of the obtained original embeddings, a corresponding augmentation function, thereby obtaining corresponding augmented embeddings; and adding the augmented embeddings and the original embeddings to an embeddings dataset.

Anmelder

Firma

AUMOVIO Germany GmbH

Land

🇩🇪 Deutschland