Skip to main content
Vai all'homepage della Commissione europea (si apre in una nuova finestra)
italiano it
CORDIS - Risultati della ricerca dell’UE
CORDIS

Algorithms and Representations for Continual Learning

Obiettivo

Massive data poses a fundamental challenge to machine learning (ML) since algorithm running times scale
with data size. As data grows faster than computational power, future problems can take longer to solve
without economizing training via continual learning (CL). However, CL can result in performance drops in
prior tasks, known as catastrophic forgetting (CF), a persistent issue in ML, including fine-tuning foundation
models (FMs), blocking their full potential.
Contrary to the dogma that increasing more resources are critical for future FMs, our recent research
shows that computational investments can be recovered near-optimally without CF, hinting at a sustainable,
future-proof ML paradigm. Our observation calls for new formulations in robust representation learning
that prevent catastrophic overfitting (CO) and new game-theoretic algorithms that avoid limit cycles (LC).
However, understanding the trade-offs between data and computation in CL is at its infancy, especially for
non-linear models. Existing methods are case-specific and do not rigorously offer sustainability, robustness, or
trustworthiness.
The proposed ERC project ARC addresses these challenges by establishing mathematical foundations to
manage continuous data growth without CF or CO or LC and to optimize dynamic representations with
new robust and reinforcement learning formulations. We build new adaptive algorithms as well as decision
verification techniques that optimally handle increasing data sizes as well as dynamic models while matching
theoretical runtime bounds. We demonstrate our rigorous theory and robust learning frameworks in real-world
scenarios, such as neuro-rehabilitation and circuit design, to validate methods and explore trade-offs across
domains.
Our approach views learning problems as dynamic and interactive, potentially revolutionizing ML by creating
holistic, sustainable systems with optimal performance across diverse applications under limited resources.

Campo scientifico (EuroSciVoc)

Questo progetto non è ancora stato classificato con EuroSciVoc.
Suggerisci i campi scientifici che ritieni più rilevanti e aiutaci a migliorare il nostro servizio di classificazione.

Parole chiave

Programma(i)

Argomento(i)

Meccanismo di finanziamento

HORIZON-ERC - HORIZON ERC Grants

Vedi tutti i progetti finanziati nell’ambito di questo schema di finanziamento

Invito a presentare proposte

(si apre in una nuova finestra) ERC-2025-ADG

Vedi tutti i progetti finanziati nell’ambito del bando

Istituzione ospitante

ECOLE POLYTECHNIQUE FEDERALE DE LAUSANNE
Contributo netto dell'UE
€ 3 344 552,00
Indirizzo
BATIMENT CE 3316 STATION 1
1015 Lausanne
Svizzera
Regione
Schweiz/Suisse/Svizzera Région lémanique Vaud
Tipo di attività
Higher or Secondary Education Establishments
Collegamenti
Costo totale
Nessun dato

Beneficiari (1)