Computer-implemented training method, classification method and system and computer-readable recording medium

US12073322B2 · kind B2 · utility

Assignees

• TOYOTA JIDOSHA KABUSHIKI KAISHA · JP
• Katholieke Universiteit Leuven · BE

Inventors

• Wim Abbeloos · Hove, BE
• Gabriel Othmezouri · Brussels, BE
• Wouter Van Gansbeke · Dilbeek, BE
• Simon Vandenhende · Leuven, BE
• Marc Proesmans · Leuven, BE
• Stamatios Georgoulis · Zürich, CH
• Luc Van Gool · Leuven, BE

Key dates

• Filing date: May 21, 2021
• Grant date: Aug 27, 2024
• Expiry date: Jan 8, 2043

Classification

• Technology area (CPC G): Physics
• CPC primary: G06N3/045
• WIPO field: Computer technology
• WIPO sector: Electrical engineering

Abstract

A computer-implemented method for training a classifier (Φη), including: training a pretext model (ΦΘ) to learn a pretext task, so as to minimize a distance between an output of a source sample via the pretext model (ΦΘ) and an output of a corresponding transformed sample via the pretext model (ΦΘ), 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 (Φη) 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.