Web-App for knowledge transfer between complex systems via Explainable AI and Network Science

# Record card
326
Thematic areas
Aerospace and Earth Science
Aerospace and Earth Science / Space sciences
Materials
Materials / Properties of materials, corrosion, degradation
Materials / Processes of production & treatment of materials
ICT & Electronics
ICT & Electronics / Big Data
ICT & Electronics / Artificial Intelligence
Health & Biotech
Health & Biotech / Bio-medicals
Health & Biotech / Bio-informatics
Energy and environmental sustainability
Energy and environmental sustainability / Simulation
Tourism, social sciences and cultural heritage / Tourism
Tourism, social sciences and cultural heritage / Socio-economic models
Description

The technology is an innovative Web-App integrating Network Science and Explainable AI (XAI) to model, analyze, and transfer knowledge across heterogeneous scientific domains (e.g., neuroscience, materials, social systems, astrophysics). The system converts complex data into a common abstract framework: Network Science maps dynamic relationships, while XAI identifies relevant features, ensuring process transparency. A key strength is transfer learning, enabling the application of models learned in data-rich sectors to data-scarce contexts, thereby reducing time and resources. For instance, an algorithm trained on vast human neural networks can be readapted to study galaxy distribution or the resilience of new materials, where data is limited. The solution supports interdisciplinary research, the discovery of universal patterns, and the development of advanced applications in different domains (e.g., biomedicine, smart materials, economy, astrophysics, social science, etc.).

Type of innovation
Product / process innovation in integration with an already existing technology
Description of innovative features / Competitive advantages

The innovation lies in the unique integration of Network Science and XAI into a framework that makes different systems comparable (from the brain to galaxies). Compared to current vertical, domain-specific solutions, this platform is transversal: it allows identifying common structural principles and transferring predictive models from "data-rich" to "data-poor" sectors. This transfer learning approach drastically reduces costs and training times. Furthermore, the use of XAI ensures the transparency ("white-box") of results, overcoming the opacity of traditional "black-box" models and facilitating scientific and industrial validation. The web-app offers a scalable tool for research institutions and the high-tech industry.

Reference market
Creation of new markets
Development stage
Feasibility
TRL
2
3
Advantages
New product/process/service/technology
Patentable technology
No
Patented technology
No
Publication of technology
Published
Technology validation/demonstration
Internal validation
Market positioning
Italian
European
International
Partner required
Enteprise
Public research center/university
Private research center
Cooperation in national /european / international project

Information
For more information and/or to be put in contact with the Research Team, please contact the Project Manager:

Barbara Angelini - Project Manager
CNR - Unità Valorizzazione della Ricerca
Phone number 06.49932415
E-mail barbara.angelini@cnr.it