logo SBA

ETD

Archivio digitale delle tesi discusse presso l’Università di Pisa

Tesi etd-02062023-172422


Tipo di tesi
Tesi di laurea magistrale
Autore
SALVATI, DARIO
URN
etd-02062023-172422
Titolo
Prompt Continual Learning for Concepts-based Scenarios
Dipartimento
INFORMATICA
Corso di studi
INFORMATICA
Relatori
relatore Prof. Lomonaco, Vincenzo
Parole chiave
  • computer vision
  • continual learning
  • deep learning
  • machine learning
  • prompt continual learning
Data inizio appello
24/02/2023
Consultabilità
Completa
Riassunto
In this thesis work, we propose a new family of Continual Learning scenarios, called
Concepts-based scenarios, for realizing Continual Learning benchmarks that are compliant
with real-world applications. Then, we illustrate how such scenarios can be faced using
Prompt-based methodologies for Continual Learning. Lastly, we discuss how the afore-
mentioned methods compare with the state-of-the-art Continual Learning techniques, and
we explore interesting future research directions.
File