Lightweight Edge AI & GenAI
Develop and optimise AI and generative models for resource-limited devices. The programme includes model compression, knowledge distillation, federated training and lightweight explainability.
MUR · PN RIC 2021–2027 · ICSC
ECHO-TWIN
Research, infrastructure and skills for intelligent Digital Twin ecosystems
Edge-Cloud-HPC Optimized Twins based on Workflow-enhanced Inference Networks
ONE INITIATIVEThree connected components
RISEResearch & innovationModels, algorithms and applications
↗NETInfrastructure & transferShared resources and innovation hubs
↗UPSkills & trainingFrom business needs to competencies
↗EDGE ↔ CLOUD ↔ HPC
- Programme
- PN RIC 2021–2027
- Coordinator
- ICSC Foundation
- Operational launch
- 5–7 May 2026 · Naples
- Project components
- RISE · NET · UP
Digital Twin-enabled Smart Ecosystems
A continuum of computing, not an isolated digital replica
ECHO-TWIN develops Digital Twin-enabled Smart Ecosystems (DTSE): connected systems in which multiple digital twins exchange information and support monitoring, prediction and operational decisions. Research is coupled with infrastructure development and skills transfer.
The initiative connects local processing at the edge with cloud services and high-performance computing. Its three components have different roles: RISE develops the scientific methods; NET strengthens shared infrastructure and technology transfer; UP develops the skills needed to use them.
01 · EDGE
↔Observe & infer
Sensors, local data and resource-constrained AI close to the physical system
02 · CLOUD
↔Connect & coordinate
Shared services, workflows and distributed Digital Twin applications
03 · HPC
Train & simulate
Compute-intensive models, large-scale analysis and simulation
Digital Twin-enabled Smart Ecosystems · Performance · Interoperability · Security · Sustainability
RISEAction 1.1.2
Research and Innovation for Scalable Edge-computing
The research component
Industrial research and experimental development of methods, models and technologies for Digital Twin-enabled Smart Ecosystems.
M.O.D.A.L. · UNINA-DMA
Federated and edge intelligence for Digital Twin ecosystems
M.O.D.A.L. is involved in RISE through UNINA’s research contribution. The programme connects AI model development with distributed execution and intelligent services. The UNINA research agenda connects the following complementary strands. The programme-level tasks are presented separately from unit-specific responsibilities.
Distributed learning & inference
Study distributed training, fine-tuning and inference, including federated mechanisms. Investigate intelligent-agent coordination for tasks executed across the computing continuum.
Intelligent services for DTSE
Study data-driven behavioural models and federated aggregation for anomaly detection, alongside prediction and what-if simulation as cross-application services.
Application link · WP5 / T5.1
Smart mobility
The UNINA contribution includes traffic prediction, alternative routing and mobility services built from distributed observations.
Application link · WP8 / T8.1
Environmental monitoring
The programme assigns UNINA the investigation of anomalous air pollution through low-cost sensor fleets and edge–cloud training and inference.
RISE research scope: Technical Programme v2.0, WP3, WP5 and WP8. A public UNINA-DMA research notice also documents work on scalable Edge Computing for ECHO-TWIN-RISE.
NETAction 1.1.3b
Networking for Excellence and Technology Transfer
From research methods to shared experimental infrastructure
NET strengthens the innovation cluster connected to ICSC through infrastructure, laboratories and technology-transfer services.
The infrastructure programme connects heterogeneous resources, including low-power devices, and supports benchmarking, interoperability and operational resilience. It gives research groups and enterprises shared environments in which to test Digital Twin and AI solutions.
UNINA-DMA contributes to validating distributed architectures in federated-training scenarios and edge–cloud simulations. This activity is explicitly identified in the NET unit programme.
UNINA-DMA contribution
Validate, compare, transfer
Use federated learning workloads to examine the behaviour of distributed architectures and connect methodological research to test environments.
UPAction 1.4.3
Upskilling for Progress
Turn technological capability into usable skills
Advanced training, specialist support and networking for enterprises, with particular attention to Southern Italy.
WP2 · Led by UNINA-DMA
Needs analysis and training design
UNINA-DMA coordinates the methodology, scientific supervision and integration of WP2. The partner units contribute evidence and content in their respective areas. The objective is to connect documented business needs with coherent training pathways, not simply to compile a list of courses.
01
Identify needs
Map enterprise requirements and emerging skill gaps.
02
Define competencies
Relate professional profiles to ECHO-TWIN technologies.
03
Design pathways
Specify modules, learning outcomes and assessment.
04
Prepare delivery
Transfer coherent pathways to the training-delivery activities.
A national collaboration
Research organisations, universities and enterprises
ICSC coordinates a public–private partnership. Participation and responsibilities differ between RISE, NET and UP.
Research and academic partners
ICSC · CNR · INFN · INAF · FBK · University of Bologna · University of Turin · University of Naples Federico II · University of Salento · University of Bari Aldo Moro · Polytechnic University of Bari · University of Catania · University of Calabria
Enterprise partners
Istituto Oncologico del Mediterraneo · Net Service · Nurjana Technologies · Planetek Italia · Parsec 3.26
5–7 May 2026 · Naples
The operational launch
The opening event took place at Federico II’s engineering campus in Piazzale Tecchio.
Read the institutional announcementConnected scientific directions
Distributed AI and scientific modelling
Explore the group’s research that connects to this programme