Green Deep Learning Approaches to Clinical AI - NAEL
Future Artificial Intelligence Research - FAIR (Spoke 9 – Green-aware AI)
Project Coordinator
Prof. Francesco Carlo Morabito, University of Reggio Calabria
Project
Green AI for clinical applications focuses on the development of environmentally sustainable and energy-efficient AI solutions to address the unique challenges and requirements of healthcare facilities. This approach aims to reduce the environmental impact and carbon footprint of AI technologies while maintaining or improving their efficiency and effectiveness in supporting clinical processes. In this context, the project will propose a range of innovative strategies that encompass most of the relevant aspects of machine and deep learning (ML/DL) and AI artifacts:
| meta-learning approaches to fine-tune hyperparameters and reduce the need for large data sets (few-shot learning); | |
| learning and deployment at the edge to minimize energy consumption and address privacy concerns; | |
| intelligent data pre-processing to avoid learning what is already known, using advanced signal processing strategies; | |
| explainability and interpretability issues to improve accuracy and robustness while providing understandable reasons for AI decisions; | |
| low-tech solutions to run complex spatiotemporal models to generate decision trees that can be interpreted with simple graphical models. |
Intervention
National Recovery and Resilience Plan (NRRP) (thematic area 1 – Artificial Intelligence: Foundational Aspects, provided for in the MUR Notice No. 341 of 03/15/2022 )
SPOKE
Labs activities
To Be Advised
Team Lab
To Be Advised
GitHub Project
Data and source code are hosted on GitHub
Disseminations
Intelligenza Artificiale: una rivoluzione nella Sanità Tecniche innovative di IA per la medicina e la clinica
RealAIze - Esperienze Real Life e AI nella SM (Siracusa, 2025, Italy)