El pròxim dijous 7 de maig, el Dipòsit Digital no estarà operatiu de 8:00 a 12:00 h per tasques d'actualització. Disculpeu les molèsties.
El próximo jueves 7 de mayo, el Dipòsit Digital no estará operativo de 8:00 a 12:00 h debido a tareas de actualización. Disculpen las molestias.
Our digital repository will be temporarily unavailable on Thursday, May 7th, from 8:00 a.m. to 12:00 p.m. due to a system update.
 
Carregant...
Miniatura

Tipus de document

Treball de fi de màster

Data de publicació

Llicència de publicació

cc-by-nc-nd (c) Rubén Jiménez Lumbreras, 2026
Si us plau utilitzeu sempre aquest identificador per citar o enllaçar aquest document: https://hdl.handle.net/2445/228768

Distance-based copying of machine learning classifiers

Títol de la revista

Director/Tutor

ISSN de la revista

Títol del volum

Recurs relacionat

Resum

Copying machine learning black box classifiers is a key framework that allows practitioners to upgrade their old models, enriching them with new properties, changing their architectures or adapting them to comply with the current AI legislations. Thanks to the copying techniques and assumptions, these improvements can be done even in settings where retraining the original system from scratch is not possible, due to resource, protocol or availability constraints. In this work, we propose the use of signed distances to the decision boundary as a replacement of the black box hard labels used to build the copies, and introduce two different algorithms to compute these distances. In addition, we observe that distance-based copying could behave as a model-agnostic regularization technique and develop a flexible framework to reduce the generalization error of the copies. Then, we validate these proposals through a series of experiments on synthetic datasets and real problems. Results show that distance-based copying is successful across multiple relevant settings and evaluation metrics. Furthermore, results also validate the quality of the predicted distances and their potential as uncertainty measures.

Descripció

Treballs finals del Màster de Fonaments de Ciència de Dades, Facultat de matemàtiques, Universitat de Barcelona. Any: 2026. Tutor: Oriol Pujol Vila

Citació

Citació

JIMÉNEZ LUMBRERAS, Rubén. Distance-based copying of machine learning classifiers. [consulta: 5 de maig de 2026]. [Disponible a: https://hdl.handle.net/2445/228768]

Exportar metadades

JSON - METS

Compartir registre