From artificial intelligence, tools for reactive Digital Twins.
Creating a Digital Twin requires the continuous exchange of a vast amount of data with the real system, often heterogeneous and difficult to interpret. In this context, artificial intelligence, and particularly machine learning, plays a key role in analyzing, filtering, and making sense of this data in real time, enhancing the interaction between the physical system and its digital counterpart. Research Topic 3 (RT3) focuses on integrating machine learning into numerical computation algorithms to make Digital Twins even smarter and more efficient. In this scenario, artificial neural networks play a crucial role thanks to their ability to recognize hidden patterns and generate reliable predictions, even under changing conditions. In particular, through the computational power of deep learning, Digital Twins can respond in near real-time, maintaining a constant connection with their physical counterparts.
The branch of artificial intelligence that focuses on developing algorithms capable of learning or improving their performance based on the data provided.
Artificial neural network
A computational model composed of many interconnected units (neurons), capable of making decisions through processes inspired by the functioning of the human brain.
Deep learning
A subset of machine learning algorithms composed of multiple layers of interconnected networks that progressively extract higher-level features from raw data.
Tasks and outputs
Research Topic 3 is organized into three tasks, each corresponding to a different step in the research workflow, from data and model integration to modularization and adaptation of digital twins.
Each scientific output of Research Topic 3 is associated with a particular task and with a milestone that situates it within the iNEST project, either along the project timeline (2022, 2023, 2024, 2025) or within specific project activities (e.g., those involving Young Researcher Grants).
1 – Model integration
Task RT3.1 is focused on automatic learning techniques for continuous data and model integration.
Stochastic Quasi-Newton Trust-Region Training for Deep Networks
Yousefi, M., & Martínez, Á. (2023). Deep Neural Networks Training by Stochastic Quasi-Newton Trust-Region Methods. Algorithms, 16(10), 490. https://doi.org/10.3390/a16100490
University of Trieste
2023
Journal paper
Generative Abstraction of Population Markov Models
Cairoli, F., Anselmi, F., d’Onofrio, A., & Bortolussi, L. (2023). Generative abstraction of Markov population processes. Theoretical Computer Science, 977, 114169.
University of Trieste
2023
Journal paper
Neural Network Compression via Knowledge Distillation and Tensors
Meneghetti, L., Bianchi, E., Demo, N., & Rozza, G. (2025). KD-AHOSVD: Neural network compression via knowledge distillation and tensor decomposition. In Lecture notes in computer science (pp. 81–92). https://doi.org/10.1007/978-3-031-87897-8_7
SISSA
2024
Conference paper
Non-Monotone Trust-Region Methods with Noisy Oracles
Krejić, N. et al. (2024). A non-monotone trust-region method with noisy oracles and additional sampling. Comput Optim Appl 89, 247–278. https://doi.org/10.1007/s10589-024-00580-w
University of Trieste
2024
Journal paper
Probabilistic Programming for Collective Adaptive Systems
Randone, F., Doz, R., Cairoli, F., & Bortolussi, L. (2024, October). Towards a probabilistic programming approach to analyse collective adaptive systems. In International Symposium on Leveraging Applications of Formal Methods (pp. 168-185). Cham: Springer Nature Switzerland.
University of Trieste
2024
Conference paper
First- and Second-Order Directions for Deep Network Training
Martínez, Á., Viola, M., Yousefi, M. (2025). Combined First- and Second-Order Directions for Deep Neural Networks Training. In: Sergeyev, Y.D., Kvasov, D.E., Astorino, A. (eds) Numerical Computations: Theory and Algorithms. NUMTA 2023. Lecture Notes in Computer Science, vol 14476. Springer, Cham. https://doi.org/10.1007/978-3-031-81241-5_9
University of Trieste
2025
Book contribution
2 – Modular AI
Task RT3.2 investigates machine learning and AI techniques for developing modular digital twins.
Graph-Based Inference of Chemical Reaction Networks
Bortolussi, L., Cairoli, F., Klein, J., & Petrov, T. (2023, September). Data-driven inference of chemical reaction networks via graph-based variational autoencoders. In International Conference on Quantitative Evaluation of Systems (pp. 143-147). Cham: Springer Nature Switzerland.
University of Trieste
2023
Conference paper
Conformal Predictive Monitoring for Multi-Modal Systems
Cairoli, F., Bortolussi, L., Deshmukh, J. V., Lindemann, L., & Paoletti, N. (2025, September). Conformal Predictive Monitoring for Multi-modal Scenarios. In International Conference on Runtime Verification (pp. 336-356). Cham: Springer Nature Switzerland.
University of Trieste
2025
Conference paper
Neuro-Symbolic Discovery of Population Markov Processes
Bortolussi, L., Cairoli, F., Klein, J., & Petrov, T. (2025). Neuro-Symbolic Discovery of Markov Population Processes.
University of Trieste
2025
Conference paper
Evolutionary Synthesis of Probabilistic Programs
Doz, R., Randone, F., Medvet, E., & Bortolussi, L. (2025, July). Evolutionary Synthesis of Probabilistic Programs. In Proceedings of the Genetic and Evolutionary Computation Conference (pp. 999-1007).
University of Trieste
2025
Conference paper
Trace-Elites with Multi-Point Descriptors
Ludwig, H. M., Espeseth, A., & Medvet, E. (2025, April). Trace-Elites: Better Quality-Diversity with Multi-point Descriptors. In International Conference on the Applications of Evolutionary Computation (Part of EvoStar) (pp. 338-353). Cham: Springer Nature Switzerland.
University of Trieste
2025
Conference paper
Stepping Stones in MAP-Elites Search Dynamics
Nadizar, G., Rusin, F., Medvet, E., & Ochoa, G. (2025, April). The Role of Stepping Stones in MAP-Elites: Insights from Search Trajectory Networks. In European Conference on Genetic Programming (Part of EvoStar) (pp. 224-239). Cham: Springer Nature Switzerland.
University of Trieste
2025
Conference paper
3 – Adaptive digital twin
Task RT3.3 has the goal of integrating AI and automatic learning techniques for real-time simulation, optimization, and adaptation of digital twins.
Model-Free Control of Cable Robots
Blanchini, F., Della Schiava, L., Fenu, G., Giordano, G., Pellegrino, F. A., & Salvato, E. (2023). Model-free cable robot control. IFAC-PapersOnLine, 56(2), 550-555.
University of Trieste
2023
Conference paper
Data-Driven Relatively Optimal Control
Pellegrino, F. A., Blanchini, F., Fenu, G., & Salvato, E. (2023). Data-driven dynamic relatively optimal control. European Journal of Control, 74, 100839.
University of Trieste
2023
Journal paper
Variational Model Checking for Stochastic Verification
Bortolussi, L., Cairoli, F., Carbone, G., & Pulcini, P. (2023, October). Scalable stochastic parametric verification with stochastic variational smoothed model checking. In International Conference on Runtime Verification (pp. 45-65). Cham: Springer Nature Switzerland
University of Trieste
2023
Conference paper
Learning-Based Predictive Monitoring with Statistical Guarantees
Cairoli, F., Bortolussi, L., & Paoletti, N. (2023, October). Learning-based approaches to predictive monitoring with conformal statistical guarantees. In International Conference on Runtime Verification (pp. 461-487). Cham: Springer Nature Switzerland.
University of Trieste
2023
Conference paper
Model Abstraction with Score-Based Diffusion Models
Bortolussi, L., Cairoli, F., Giacomarra, F., & Scassola, D. (2023, September). Model Abstraction and Conditional Sampling with Score-Based Diffusion Models. In International Conference on Quantitative Evaluation of Systems (pp. 307-310). Cham: Springer Nature Switzerland.
University of Trieste
2023
Conference paper
Conformal Monitoring of STL Requirements
Cairoli, F., Paoletti, N., & Bortolussi, L. (2023). Conformal Quantitative Predictive Monitoring of STL Requirements for Stochastic Processes.
University of Trieste
2023
Conference paper
Shared Control for Traffic Wave Attenuation
Salvato, E., Elia, L., Fenu, G., & Parisini, T. (2024, December). Stop-and-Go Traffic Wave Attenuation: A Shared Control Approach. In 2024 IEEE 63rd Conference on Decision and Control (CDC) (pp. 4929-4934). IEEE.
University of Trieste
2024
Conference paper
Active Disturbance Rejection MPC for Over-Actuated Systems
Salvato, E., Fenu, G., Pellegrino, F. A., & Parisini, T. (2024, June). An Active Disturbance Rejection Model Predictive Controller for Constrained Over-Actuated Systems. In 2024 European Control Conference (ECC) (pp. 2547-2552). IEEE.
University of Trieste
2024
Conference paper
Resilient Coverage Control for UAV Teams
Rezaee, H., Salvato, E., Fenu, G., & Parisini, T. (2024). Resilient coverage by teams of quadrotor UAVs: Theory and experiments. IEEE Transactions on Control Systems Technology, 32(6), 2009-2022.
University of Trieste
2024
Journal paper
Calibration-Free Visual Servo Control
Salvato, E., Blanchini, F., Fenu, G., Giordano, G., & Pellegrino, F. A. (2025). Position-based visual servo control without hand-eye calibration. Robotics and Autonomous Systems, 105045.
University of Trieste
2025
Journal paper
Model-Free Kinematic Control for Robots
Salvato, E., Blanchini, F., Fenu, G., Giordano, G., & Pellegrino, F. A. (2025). Model-free kinematic control for robotic systems. Automatica, 173, 112030.
University of Trieste
2025
Journal paper
Conformal Predictive Monitoring with Conditional Guarantees
Cairoli, F., Kuipers, T., Bortolussi, L., & Paoletti, N. (2025). Conformal quantitative predictive monitoring of stochastic systems with conditional validity. Nonlinear Analysis: Hybrid Systems, 57, 101606.
University of Trieste
2025
Journal paper
Scalable Stochastic Parametric Verification
Cairoli, F., & Bortolussi, L. (2025). Scalable and reliable stochastic parametric verification with stochastic variational smoothed model checking. International Journal of Systems Science, 1-29.
University of Trieste
2025
Journal paper
Certified Planning with Deep Generative Models
Giacomarra, F., Hosseini, M., Paoletti, N., & Cairoli, F. (2025, May). Certified Guidance for Planning with Deep Generative Models. In Proceedings of the 24th International Conference on Autonomous Agents and Multiagent Systems (pp. 877-885).
University of Trieste
2025
Conference paper
Coordination
Research Topic 3 is led by the University of Trieste. The International School for Advanced Studies (SISSA) is also involved in RT3.
