Computerimplementiertes Trainingsverfahren, Klassifizierungsverfahren und -System, Computerprogramm und Computerlesbares Aufzeichnungsmedium

Details

Veröffentlichungs-Nr.

EP3913544

Aktenzeichen

EP20176143

Anmeldetag

22. Mai 2020

Veröffentlichung

12. März 2025

Erteilung

12. März 2025

Rechtsraum

EP

IPC

G06N3/08// G06N3/04

Abstract

A computer-implemented method for training a classifier (Φ<η>), comprising: S10) training a pretext model (Φ<Θ>) to learn a pretext task, so as to minimize a distance between an output of a source sample via the pretext model (Φ<Θ>) and an output of a corresponding transformed sample via the pretext model (Φ<Θ>), the transformed sample being a sample obtained by applying a transformation (T) to the source sample ; S20) determining a neighborhood (NXi) of samples (Xi) of a dataset (SD) in the embedding space; S30) training the classifier (Φ<η>) to predict respective estimated probabilities Φηj(Xi), j=1..C, for a sample (Xi) to belong to respective clusters (Cj), by using a second training criterion which tends to: - maximize a likelihood for a sample and its neighbors (Xj) of its neighborhood (Nxi) to belong to the same cluster; and - force the samples to be distributed over several clusters.

Anmelder

Firmen

TOYOTA JIDOSHA KABUSHIKI KAISHA
Katholieke Universiteit Leuven

Land

🇯🇵 Japan

Fachgebiete