About me
Roberta De Fazio is a postdoctoral researcher in the Department of Mathematics and Physics at the University of Campania "Luigi Vanvitelli".
She is an active member of the Smartists Research Group (SMArt cRiTIcal SysTems and Software), where her research focuses on improving the reliability of critical systems by combining model-based and data-driven approaches in AI-intensive contexts, such as Clinical Decision Support Systems, Digital Twins and Predictive Maintenance.
Her interests lie in Predictive Maintenance, Process Mining, Causality Analysis and Artificial Intelligence.
Roberta earned her Bachelor's Degree and her Master's Degree in Mathematics (both with honors), and her Ph.D. in Mathematics, Physics and Engineering Applications from the University of Campania "Luigi Vanvitelli".
As part of her Ph.D. studies, Roberta spent three months as a visiting researcher at Hasselt University in Belgium (Erasmus+ programme) and six months at a corporate research laboratory with Gematica S.r.l., gaining valuable international and industry experience.
She also holds the National Habilitation for High School Teaching in Mathematics and Physics (A027) and is a member of the European Association for Theoretical Computer Science (EATCS).
0Publications
0Research projects
0Collaborations
0Talks
Research
Research interest
The research focuses on improving the reliability of critical systems, in particular, focusing on maintenance aspects of large networked transportation infrastructures:
- Critical Infrastructures design and modelling (e.g., rail, automotive, avionics, healthcare)
- Analysis techniques for critical systems performance and dependability evaluation
- Formal models for fault predictions
- Causality analysis and Process Mining
The research activities focus on the design and the implementation of analysis techniques that combine model-based and data-driven approaches in AI-intensive contexts (Clinical Decision Support Systems, Digital Twins, Prediction Maintenance) and on the evaluation and assurance of their trustworthiness.
In particular, the efforts of the research are devoted to extending existent formalism like Fault Trees, Petri Nets, and Bayesian Networks and to the integration with Process Mining (traditional and Object-Centric) and machine learning algorithms for Prognostic Health Monitoring, diagnosis support and clinical pathway predictions.
Research projects
- MISTICA - sMart Iot aneSThesIa Cart & cAbinet (CUP B69J23006090007), funded by Ministero delle Imprese e del Made in Italy (MIMIT) - "Accordi di innovazione DM 31/12/2021 secondo bando". Investigator, Sep 2025 - Nov 2025.
- SUD4VUP - Sistema di sUpporto Decisionale per la diagnosi precoce e il follow-up di pazienti affetti da Valvole Uretrali Posteriori (CUP B23C24000440001), funded by IIT/CNR within the "RAISE – Robotics and AI for Socio-Economic Empowerment" Ecosystem (NRRP, Next Generation EU). WP leader of WP4 (Testing & Validation), 2024/25.
- DARWNIST - aDversarial scenArios geneRation With dIgital twiNs In induSTry, funded by the young researchers programme, Università della Campania "Luigi Vanvitelli". Investigator, 2022/23.
Collaborations
- Hasselt University (Hasselt, BE) - integration of data-driven techniques, in particular causality analysis, into model-based frameworks (Erasmus+ visiting, Apr - Jul 2024, Business Informatics Research Group, Prof. Benoit Depaire).
- University of Chemistry and Technology (Prague, CZ) - voice signal analysis and Machine Learning support to the diagnosis of voice disorders.
- University of Cassino and Southern Lazio (IT) - Deep Learning for image diagnosis, Large Language Models for medical report analysis and clinical decision support.
- University of Bari "Aldo Moro" (IT) - Artificial Intelligence techniques for medical image processing.
- Gematica S.r.l. (Naples, IT) - analysis techniques for a decision support system applied to railway networks.
Software
- HECATE - HEpatic CAncer deTection via bayesian nEtwork
- CALYPSO - Causality AnaLYsis for Prediction of System Operation
- STELLA - Segmentation Tool for Enhanced Localisation and Labelling of Diagnostic Areas
All software is available on GitHub.
