Published in Soft Computing, to appear, 2024, 2024
We showcase our tool NeVer2 along with a complete experimental campaign.
Recommended citation: S. Demarchi, D. Guidotti, L. Pulina and A. Tacchella (2024), "NeVer2: Learning and Verification of Neural Networks." in Soft Computing, to appear, 2024.
Published in International Conference on Modelling and Simulation, ECMS 2024, Cracow, Poland, 2024
We experiment with different techniques to improve our verification algorithm.
Recommended citation: S. Demarchi, A. Gimelli and Armando Tacchella (2024), "Improving Abstract Propagation for Verification of Neural Networks." in International Conference on Modelling and Simulation, ECMS 2024, Cracow, Poland, June 4-7th, 2024, Proceedings, 2024.
Published in 6th Workshop on Formal Methods for ML-Enabled Autonomous Systems, FoMLAS 2023, Paris, France, 2023
We propose a standard for neural networks verification and a tool for generating benchmarks.
Recommended citation: S. Demarchi, D. Guidotti, L. Pulina and A. Tacchella (2023), "Supporting Standardization of Neural Networks Verification with VNN-LIB and CoCoNet." in 6th Workshop on Formal Methods for ML-Enabled Autonomous Systems, FoMLAS 2023, Paris, France, July 17-18, 2023, Proceedings, 2023.
Published in Università degli Studi di Genova, 2023
PhD Thesis on the work about Design and Verification.
Recommended citation: S. Demarchi (2023), "Experimenting with Constraint Programming Techniques in Artificial Intelligence: Automated System Design and Verification of Neural Networks." PhD Thesis, 2023.
Published in Cyber-Physical Systems Summer School Workshop, CPSWS 2022, Pula, Italy, 2022
We experiment with abstraction refinement in order to achieve fast and precise results.
Recommended citation: S. Demarchi and D. Guidotti (2022), "Counter-Example Guided Abstract Refinement for Verification of Neural Networks." in Cyber-Physical Systems Summer School Workshop, CPSWS 2022, Pula, Italy, September 19, 2022, Proceedings, 2022.
Published in Intelligenza Artificiale 16 (1), 131-150, 2022, 2022
We present several declarative encodings for the automated design of elevator systems.
Recommended citation: G. Cicala, S. Demarchi, M. Menapace, L. Annunziata and A. Tacchella (2022), "A Comparison of Declarative AI Techniques for Computer Automated Design of Elevator Systems." in Intelligenza Artificiale 16 (1), 131-150, 2022.
Published in International Conference on Modelling and Simulation, ECMS 2022, Aalesund, Norway, 2022
We provide a real case study from automotive to try and verify a NN controller.
Recommended citation: S. Demarchi, D. Guidotti, A. Pitto and Armando Tacchella (2022), "Formal Verification of Neural Networks: a Case Study about Adaptive Cruise Control." in International Conference on Modelling and Simulation, ECMS 2022, Aalesund, Norway, May 30th-June 3rd, 2022, Proceedings, 2022.
Published in International Conference of the Italian Association for Artificial Intelligence, AIxIA 2021, Milano, Italy December 1st-3rd, 2021, 2019
In the context of elevator systems design, we challenge SMT and CP solvers to enable domain-agnostic design.
Recommended citation: S. Demarchi, M. Menapace and A. Tacchella (2021), "Automated Design of Elevator Systems: Experimenting with Constraint-Based Approaches." in International Conference of the Italian Association for Artificial Intelligence, AIxIA 2021, Milano, Italy December 1st-3rd, 2021, Proceedings, 2021.
Published in IEEE International Conference on Tools with Artificial Intelligence, ICTAI 2019, Portland, Oregon, 2019
In the context of elevator systems design, we use SMT as a mean to enable domain-agnostic configuration.
Recommended citation: S. Demarchi, M. Menapace and Armando Tacchella (2019), "Automating elevator design with Satisfiability Modulo Theories." in IEEE International Conference on Tools with Artificial Intelligence, ICTAI 2019, Portland, Oregon, November 4-6, 2019, Proceedings, 2019.
Published in Cyber-Physical Systems Summer School Workshop, CPSWS 2019, Alghero, Italy, 2019
Starting from a success story in elevator systems design, we try and use SMT as a mean to enable domain-agnostic system configuration.
Recommended citation: S. Demarchi (2019), "Automated Design of Complex Systems with Constraint Programming Techniques." in Cyber-Physical Systems Summer School Workshop, CPSWS 2019, Alghero, Italy, September 23, 2019, Proceedings, 2019.