Antonio Carta
Post-doc at Pisa University
Learning continually from non-stationary data streams is a challenging research topic of growing popularity in the last few years. Being able to learn, adapt and generalize continually, in an efficient, effective and scalable way appears to be fundamental for a more sustainable development of Artificial Intelligent systems. However, access to raw data can be unavailable due to privacy constraints. More in general, we may want to exploit the wide availability of compressed information in the form of trained models.
In this project, we aim to study a recently proposed paradigm named ”Ex-Model Continual Learning”, where an agent learns from a sequence of trained models instead of raw data. Continual learning from models is a challenging and unexplored scenario that allows independent models to share their knowledge efficiently while ensuring the privacy of the data.