Publications
OUR RESEARCH
Scientific Publications
Here you can find the comprehensive list of publications from the members of the Research Center on Computer Vision and eXtended Reality (xRAI).
Use the tag cloud to filter papers based on specific research topics, or use the menus to filter by year, type of publication, or authors.
For each paper, you have the option to view additional details such as the Abstract, Links, and BibTex record.
Research is formalized curiosity. It is poking and prying with a purpose
Zora Neale Hurston
2025
Pirani, Massimiliano; Cucchiarelli, Alessandro; Naeem, Tariq; Spalazzi, Luca
A Blockchain-Driven Cyber-Systemic Approach to Hybrid Reality Journal Article
In: Systems, vol. 13, no. 4, pp. 294, 2025, ISSN: 2079-8954.
Abstract | Links | BibTeX | Tags:
@article{pirani_blockchain-driven_2025,
title = {A Blockchain-Driven Cyber-Systemic Approach to Hybrid Reality},
author = {Massimiliano Pirani and Alessandro Cucchiarelli and Tariq Naeem and Luca Spalazzi},
url = {https://www.mdpi.com/2079-8954/13/4/294},
doi = {10.3390/systems13040294},
issn = {2079-8954},
year = {2025},
date = {2025-04-01},
urldate = {2025-09-30},
journal = {Systems},
volume = {13},
number = {4},
pages = {294},
abstract = {Hybrid Reality (HyR) is the place where human beings and artificial entities interact. HyR modelling relies simultaneously on the cognitive power of humans and artificial entities. In addition, HyR is an evolving paradigm where natural and artificial intelligence can intervene in processes that demand proper control. This work aims to lay the foundation for a systematic approach to understanding and modeling present and future human–machine symbiosis under a systems engineering perspective. It introduces a novel cyber-systemic methodology for managing the engineering of purposeful regulation for HyR phenomena by integrating the Blockchain technology framework and principled methods of cybernetics. This formalized interdisciplinary methodology integrates system dynamics, agent-based computation, artificial intelligence, and Blockchain-powered security and safety layers. The Blockchain framework, seen under a new cyber-systemic perspective, provides new opportunities and tools for the organization and control of HyR. A Cybersystemic Security Kit is here defined as a major component of the methodology, representing a candidate to offer viable breakthroughs in the field with respect to the best practices of Industry 5.0 when a systemically augmented perspective is adopted. Ongoing research and experimentation in the real field of sustainable supply chains is used as a motivating use case to support the proposed position. The industrial target is the primary one in its multi-dimensional and multi-faceted sustainability impacts, but this study will also reveal other potential societal areas of intervention.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Hybrid Reality (HyR) is the place where human beings and artificial entities interact. HyR modelling relies simultaneously on the cognitive power of humans and artificial entities. In addition, HyR is an evolving paradigm where natural and artificial intelligence can intervene in processes that demand proper control. This work aims to lay the foundation for a systematic approach to understanding and modeling present and future human–machine symbiosis under a systems engineering perspective. It introduces a novel cyber-systemic methodology for managing the engineering of purposeful regulation for HyR phenomena by integrating the Blockchain technology framework and principled methods of cybernetics. This formalized interdisciplinary methodology integrates system dynamics, agent-based computation, artificial intelligence, and Blockchain-powered security and safety layers. The Blockchain framework, seen under a new cyber-systemic perspective, provides new opportunities and tools for the organization and control of HyR. A Cybersystemic Security Kit is here defined as a major component of the methodology, representing a candidate to offer viable breakthroughs in the field with respect to the best practices of Industry 5.0 when a systemically augmented perspective is adopted. Ongoing research and experimentation in the real field of sustainable supply chains is used as a motivating use case to support the proposed position. The industrial target is the primary one in its multi-dimensional and multi-faceted sustainability impacts, but this study will also reveal other potential societal areas of intervention.
Moreschini, Sergio; Pour, Shahrzad; Lanese, Ivan; Balouek, Daniel; Bogner, Justus; Li, Xiaozhou; Pecorelli, Fabiano; Soldani, Jacopo; Truyen, Eddy; Taibi, Davide
AI Techniques in the Microservices Life-Cycle: a Systematic Mapping Study Journal Article
In: Computing, vol. 107, no. 4, pp. 100, 2025, ISSN: 0010-485X, 1436-5057.
Abstract | Links | BibTeX | Tags:
@article{moreschini_ai_2025,
title = {AI Techniques in the Microservices Life-Cycle: a Systematic Mapping Study},
author = {Sergio Moreschini and Shahrzad Pour and Ivan Lanese and Daniel Balouek and Justus Bogner and Xiaozhou Li and Fabiano Pecorelli and Jacopo Soldani and Eddy Truyen and Davide Taibi},
url = {https://link.springer.com/10.1007/s00607-025-01432-z},
doi = {10.1007/s00607-025-01432-z},
issn = {0010-485X, 1436-5057},
year = {2025},
date = {2025-04-01},
urldate = {2025-09-30},
journal = {Computing},
volume = {107},
number = {4},
pages = {100},
abstract = {Abstract
The use of AI in microservices (MSs) is an emerging field as indicated by a substantial number of surveys. However these surveys focus on a specific problem using specific AI techniques, therefore not fully capturing the growth of research and the rise and disappearance of trends. In our systematic mapping study, we take an exhaustive approach to reveal all possible connections between the use of AI techniques for improving any quality attribute (QA) of MSs during the DevOps phases. Our results include 16 research themes that connect to the intersection of particular QAs, AI domains and DevOps phases. Moreover by mapping identified future research challenges and relevant industry domains, we can show that many studies aim to deliver prototypes to be automated at a later stage, aiming at providing exploitable products in a number of key industry domains.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Abstract
The use of AI in microservices (MSs) is an emerging field as indicated by a substantial number of surveys. However these surveys focus on a specific problem using specific AI techniques, therefore not fully capturing the growth of research and the rise and disappearance of trends. In our systematic mapping study, we take an exhaustive approach to reveal all possible connections between the use of AI techniques for improving any quality attribute (QA) of MSs during the DevOps phases. Our results include 16 research themes that connect to the intersection of particular QAs, AI domains and DevOps phases. Moreover by mapping identified future research challenges and relevant industry domains, we can show that many studies aim to deliver prototypes to be automated at a later stage, aiming at providing exploitable products in a number of key industry domains.
The use of AI in microservices (MSs) is an emerging field as indicated by a substantial number of surveys. However these surveys focus on a specific problem using specific AI techniques, therefore not fully capturing the growth of research and the rise and disappearance of trends. In our systematic mapping study, we take an exhaustive approach to reveal all possible connections between the use of AI techniques for improving any quality attribute (QA) of MSs during the DevOps phases. Our results include 16 research themes that connect to the intersection of particular QAs, AI domains and DevOps phases. Moreover by mapping identified future research challenges and relevant industry domains, we can show that many studies aim to deliver prototypes to be automated at a later stage, aiming at providing exploitable products in a number of key industry domains.
Vergallo, Roberto; Cruz, Luís; Errico, Alessio; Mainetti, Luca
On the Effectiveness of the 'Follow-the-Sun' Strategy in Mitigating the Carbon Footprint of AI in Cloud Instances Miscellaneous
2025, (arXiv:2506.10990 [cs]).
Abstract | Links | BibTeX | Tags: and Cluster Computing, and Science, Computer Science - Artificial Intelligence, Computer Science - Computational Engineering, Computer Science - Distributed, Computer Science - Software Engineering, Finance, Parallel
@misc{vergallo_effectiveness_2025,
title = {On the Effectiveness of the 'Follow-the-Sun' Strategy in Mitigating the Carbon Footprint of AI in Cloud Instances},
author = {Roberto Vergallo and Luís Cruz and Alessio Errico and Luca Mainetti},
url = {http://arxiv.org/abs/2506.10990},
doi = {10.48550/arXiv.2506.10990},
year = {2025},
date = {2025-03-01},
urldate = {2025-09-30},
publisher = {arXiv},
abstract = {'Follow-the-Sun' (FtS) is a theoretical computational model aimed at minimizing the carbon footprint of computer workloads. It involves dynamically moving workloads to regions with cleaner energy sources as demand increases and energy production relies more on fossil fuels. With the significant power consumption of Artificial Intelligence (AI) being a subject of extensive debate, FtS is proposed as a strategy to mitigate the carbon footprint of training AI models. However, the literature lacks scientific evidence on the advantages of FtS to mitigate the carbon footprint of AI workloads. In this paper, we present the results of an experiment conducted in a partial synthetic scenario to address this research gap. We benchmarked four AI algorithms in the anomaly detection domain and measured the differences in carbon emissions in four cases: no strategy, FtS, and two strategies previously introduced in the state of the art, namely Flexible Start and Pause and Resume. To conduct our experiment, we utilized historical carbon intensity data from the year 2021 for seven European cities. Our results demonstrate that the FtS strategy not only achieves average reductions of up to 14.6% in carbon emissions (with peaks of 16.3%) but also helps in preserving the time needed for training.},
note = {arXiv:2506.10990 [cs]},
keywords = {and Cluster Computing, and Science, Computer Science - Artificial Intelligence, Computer Science - Computational Engineering, Computer Science - Distributed, Computer Science - Software Engineering, Finance, Parallel},
pubstate = {published},
tppubtype = {misc}
}
'Follow-the-Sun' (FtS) is a theoretical computational model aimed at minimizing the carbon footprint of computer workloads. It involves dynamically moving workloads to regions with cleaner energy sources as demand increases and energy production relies more on fossil fuels. With the significant power consumption of Artificial Intelligence (AI) being a subject of extensive debate, FtS is proposed as a strategy to mitigate the carbon footprint of training AI models. However, the literature lacks scientific evidence on the advantages of FtS to mitigate the carbon footprint of AI workloads. In this paper, we present the results of an experiment conducted in a partial synthetic scenario to address this research gap. We benchmarked four AI algorithms in the anomaly detection domain and measured the differences in carbon emissions in four cases: no strategy, FtS, and two strategies previously introduced in the state of the art, namely Flexible Start and Pause and Resume. To conduct our experiment, we utilized historical carbon intensity data from the year 2021 for seven European cities. Our results demonstrate that the FtS strategy not only achieves average reductions of up to 14.6% in carbon emissions (with peaks of 16.3%) but also helps in preserving the time needed for training.
Gatto, Carola; Barba, Maria Cristina; Chiarello, Sofia; Corchia, Laura; Faggiano, Federica; Nuzzo, Benito Luigi; Sumerano, Giada; Luca, Valerio De; Paolis, Lucio Tommaso De
Breaking the barriers: Extended reality and innovative technologies for enhanced accessibility of the Ceramics Museum of Cutrofiano Journal Article
In: Digital Applications in Archaeology and Cultural Heritage, vol. 36, pp. e00400, 2025, ISSN: 22120548.
Links | BibTeX | Tags: Cultural Heritage, Extended reality, Human Computer Interaction, User experience
@article{gatto_breaking_2025,
title = {Breaking the barriers: Extended reality and innovative technologies for enhanced accessibility of the Ceramics Museum of Cutrofiano},
author = {Carola Gatto and Maria Cristina Barba and Sofia Chiarello and Laura Corchia and Federica Faggiano and Benito Luigi Nuzzo and Giada Sumerano and Valerio De Luca and Lucio Tommaso De Paolis},
url = {https://linkinghub.elsevier.com/retrieve/pii/S2212054825000025},
doi = {10.1016/j.daach.2025.e00400},
issn = {22120548},
year = {2025},
date = {2025-03-01},
urldate = {2025-09-30},
journal = {Digital Applications in Archaeology and Cultural Heritage},
volume = {36},
pages = {e00400},
keywords = {Cultural Heritage, Extended reality, Human Computer Interaction, User experience},
pubstate = {published},
tppubtype = {article}
}
Intini, Paolo; Blasi, Gianni; Fracella, Francesco; Francone, Antonio; Vergallo, Roberto; Perrone, Daniele
Predicting traffic volumes on road infrastructures in the context of multi-risk assessment frameworks Journal Article
In: International Journal of Disaster Risk Reduction, vol. 117, pp. 105139, 2025, ISSN: 22124209.
Abstract | Links | BibTeX | Tags:
@article{intini_predicting_2025,
title = {Predicting traffic volumes on road infrastructures in the context of multi-risk assessment frameworks},
author = {Paolo Intini and Gianni Blasi and Francesco Fracella and Antonio Francone and Roberto Vergallo and Daniele Perrone},
url = {https://linkinghub.elsevier.com/retrieve/pii/S2212420924009014},
doi = {10.1016/j.ijdrr.2024.105139},
issn = {22124209},
year = {2025},
date = {2025-02-01},
urldate = {2025-01-20},
journal = {International Journal of Disaster Risk Reduction},
volume = {117},
pages = {105139},
abstract = {In multi-risk assessment frameworks involving road infrastructures, measures of exposure to natural hazards include traffic volumes. However, traffic counts are usually collected through traffic counter/radar stations which only cover a small part of the road network. In this study, country-wide Annual Average Daily Traffic (AADT) prediction models based on Italian data were developed to provide direct risk exposure measures both in terms of traffic volumes (continuous variable) and traffic volume discrete classes, using province-/municipality-related geographic, socio-economic and road-related variables as predictors. To ease transferability and applicability of the models, only publicly available predictors were selected. Traditional statistical techniques (generalized linear models for predicting traffic values and ordered logistic models for traffic classes) and Machine Learning (ML) approaches (XGBoost for both regression and classification problems) were used. Both the direct estimation of traffic volumes and the classification into traffic ranges provided satisfactory results in terms of goodness-of-fit and predictive accuracy metrics. Results show that population, occupation, tourism, density, number of lanes, urban environment, complex intersections and ring roads were generally related to an increase in traffic volumes. Distance from large cities and accessibility metrics are inversely related to traffic instead. The application of the XGBoost ML approach proved to be more accurate than traditional approaches only for heavy vehicles. It was discussed how the obtained models can be used as input modules for overall multi-risk assessment frameworks involving road infrastructures.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
In multi-risk assessment frameworks involving road infrastructures, measures of exposure to natural hazards include traffic volumes. However, traffic counts are usually collected through traffic counter/radar stations which only cover a small part of the road network. In this study, country-wide Annual Average Daily Traffic (AADT) prediction models based on Italian data were developed to provide direct risk exposure measures both in terms of traffic volumes (continuous variable) and traffic volume discrete classes, using province-/municipality-related geographic, socio-economic and road-related variables as predictors. To ease transferability and applicability of the models, only publicly available predictors were selected. Traditional statistical techniques (generalized linear models for predicting traffic values and ordered logistic models for traffic classes) and Machine Learning (ML) approaches (XGBoost for both regression and classification problems) were used. Both the direct estimation of traffic volumes and the classification into traffic ranges provided satisfactory results in terms of goodness-of-fit and predictive accuracy metrics. Results show that population, occupation, tourism, density, number of lanes, urban environment, complex intersections and ring roads were generally related to an increase in traffic volumes. Distance from large cities and accessibility metrics are inversely related to traffic instead. The application of the XGBoost ML approach proved to be more accurate than traditional approaches only for heavy vehicles. It was discussed how the obtained models can be used as input modules for overall multi-risk assessment frameworks involving road infrastructures.
Pirani, Massimiliano; Generosi, Andrea; Spalazzi, Luca
A perspective on generative artificial intelligence for the enhancement of methods in system dynamics Proceedings Article
In: 20th IRDO International Conference: Innovative, sustainable & socially responsible society 2025, I. Perko, M. Mulej, P. Glavic, A. Hrast, & T. Jere Jakulin (Eds.), 2025, ISBN: 978-961-7141-11-5.
Abstract | Links | BibTeX | Tags:
@inproceedings{pirani_perspective_2025,
title = {A perspective on generative artificial intelligence for the enhancement of methods in system dynamics},
author = {Massimiliano Pirani and Andrea Generosi and Luca Spalazzi},
url = {https://www.irdo.si/irdo2025/posters/45.pdf},
isbn = {978-961-7141-11-5},
year = {2025},
date = {2025-01-01},
urldate = {2025-11-05},
booktitle = {20th IRDO International Conference: Innovative, sustainable & socially responsible society 2025, I. Perko, M. Mulej, P. Glavic, A. Hrast, & T. Jere Jakulin (Eds.)},
abstract = {System Dynamics (SD) offers a set of indispensable tools for systems engineering, systems thinking, and cybernetics. To date, the development of such valuable models — crucial for ensuring causality and reproducibility in studies addressing socio-technical and social sustainability issues — has been predominantly manual. Furthermore, the criteria and processes involved in creating models such as Causal Loop Diagrams (CLD) and Stock and Flow Diagrams (SFD) are, in many cases, neither explainable nor reproducible. Recently, some studies have expressed some potential for generative artificial intelligence to be considered as a valuable tool for the automated construction of CLD models. Although promising, still many problems affect this technique but, most of all, there is a clear gap in the construction of SFD and in the transformation of a CLD into an equivalent SFD. SFDs are indeed essential when a dynamic has to be encoded into a mathematical system, in order to allow simulations, predictions and, most of all, synthesis of some control system. In this work, we would like to provide a perspective on the envisioned methodologies that have the potential to provide breakthroughs in this area, with a special focus on generative artificial intelligence, which is currently the best available technology for these kinds of transformations.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
System Dynamics (SD) offers a set of indispensable tools for systems engineering, systems thinking, and cybernetics. To date, the development of such valuable models — crucial for ensuring causality and reproducibility in studies addressing socio-technical and social sustainability issues — has been predominantly manual. Furthermore, the criteria and processes involved in creating models such as Causal Loop Diagrams (CLD) and Stock and Flow Diagrams (SFD) are, in many cases, neither explainable nor reproducible. Recently, some studies have expressed some potential for generative artificial intelligence to be considered as a valuable tool for the automated construction of CLD models. Although promising, still many problems affect this technique but, most of all, there is a clear gap in the construction of SFD and in the transformation of a CLD into an equivalent SFD. SFDs are indeed essential when a dynamic has to be encoded into a mathematical system, in order to allow simulations, predictions and, most of all, synthesis of some control system. In this work, we would like to provide a perspective on the envisioned methodologies that have the potential to provide breakthroughs in this area, with a special focus on generative artificial intelligence, which is currently the best available technology for these kinds of transformations.
Vergallo, Roberto; Aprile, Matteo; Cruz, Luís; Vadacca, Roberto; Mainetti, Luca
Large-scale Evaluation of Quantization for Reducing the Energy Footprint of Deep Learning Models Miscellaneous
2025.
@misc{vergallo_large-scale_2025,
title = {Large-scale Evaluation of Quantization for Reducing the Energy Footprint of Deep Learning Models},
author = {Roberto Vergallo and Matteo Aprile and Luís Cruz and Roberto Vadacca and Luca Mainetti},
url = {https://www.ssrn.com/abstract=5719661},
doi = {10.2139/ssrn.5719661},
year = {2025},
date = {2025-01-01},
urldate = {2025-11-10},
publisher = {SSRN},
keywords = {},
pubstate = {published},
tppubtype = {misc}
}
Vergallo, Roberto; Aprile, Matteo; Catalano, Christian; Vadacca, Roberto
Pruning LLMs for energy saving: is it worth the cost? Miscellaneous
2025.
@misc{vergallo_pruning_2025,
title = {Pruning LLMs for energy saving: is it worth the cost?},
author = {Roberto Vergallo and Matteo Aprile and Christian Catalano and Roberto Vadacca},
url = {https://www.ssrn.com/abstract=5551938},
doi = {10.2139/ssrn.5551938},
year = {2025},
date = {2025-01-01},
urldate = {2025-11-10},
publisher = {SSRN},
keywords = {},
pubstate = {published},
tppubtype = {misc}
}
Vergallo, Roberto; Mainetti, Luca
Incorporating Animals as Stakeholders in ICT: Towards Truly Inclusive Digital Sustainability: Proceedings Article
In: Proceedings of the 14th International Conference on Smart Cities and Green ICT Systems, pp. 147–154, SCITEPRESS - Science and Technology Publications, Porto, Portugal, 2025, ISBN: 978-989-758-751-1.
@inproceedings{vergallo_incorporating_2025,
title = {Incorporating Animals as Stakeholders in ICT: Towards Truly Inclusive Digital Sustainability:},
author = {Roberto Vergallo and Luca Mainetti},
url = {https://www.scitepress.org/DigitalLibrary/Link.aspx?doi=10.5220/0013431100003953},
doi = {10.5220/0013431100003953},
isbn = {978-989-758-751-1},
year = {2025},
date = {2025-01-01},
urldate = {2025-09-30},
booktitle = {Proceedings of the 14th International Conference on Smart Cities and Green ICT Systems},
pages = {147–154},
publisher = {SCITEPRESS - Science and Technology Publications},
address = {Porto, Portugal},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Vergallo, Roberto; Casciaro, Simone; Mainetti, Luca
From Fintech to Green Fintech: How to Decarbonize the AI in the Financial Domain Book Section
In: Chen, Haojun (Ed.): Advances in Finance, Accounting, and Economics, pp. 25–70, IGI Global, 2025, ISBN: 979-8-3693-8186-1 979-8-3693-8188-5.
Abstract | Links | BibTeX | Tags:
@incollection{chen_fintech_2025,
title = {From Fintech to Green Fintech: How to Decarbonize the AI in the Financial Domain},
author = {Roberto Vergallo and Simone Casciaro and Luca Mainetti},
editor = {Haojun Chen},
url = {https://services.igi-global.com/resolvedoi/resolve.aspx?doi=10.4018/979-8-3693-8186-1.ch002},
doi = {10.4018/979-8-3693-8186-1.ch002},
isbn = {979-8-3693-8186-1 979-8-3693-8188-5},
year = {2025},
date = {2025-01-01},
urldate = {2025-09-30},
booktitle = {Advances in Finance, Accounting, and Economics},
pages = {25–70},
publisher = {IGI Global},
abstract = {Machine Learning and Deep Learning are spreading at incredible pace in finance. However, the related energy consumption lacks in coherence with the sustainability requirements of banks and financial institutions. Particularly, as the energy required to run complex algorithms increases, more CO2 is released because of the power grid relying mostly on fossil fuels. While techniques are arising to mitigate the carbon footprint of AI, Fintech is facing the challenge of integrating such strategies within the peculiarities of the domain, which requires timeliness, data protection and compliance. In this work, we present the state-of-the-art Green AI techniques, reporting the available experimental results, and critically evaluating their applicability in the financial domain. This contribution is useful for financial institutions, Fintech startups and regulatory bodies who wants to measure and mitigate the impact of AI in the field, decarbonize old or new AI-based products and services, and assess the effectiveness of decarbonization efforts.},
keywords = {},
pubstate = {published},
tppubtype = {incollection}
}
Machine Learning and Deep Learning are spreading at incredible pace in finance. However, the related energy consumption lacks in coherence with the sustainability requirements of banks and financial institutions. Particularly, as the energy required to run complex algorithms increases, more CO2 is released because of the power grid relying mostly on fossil fuels. While techniques are arising to mitigate the carbon footprint of AI, Fintech is facing the challenge of integrating such strategies within the peculiarities of the domain, which requires timeliness, data protection and compliance. In this work, we present the state-of-the-art Green AI techniques, reporting the available experimental results, and critically evaluating their applicability in the financial domain. This contribution is useful for financial institutions, Fintech startups and regulatory bodies who wants to measure and mitigate the impact of AI in the field, decarbonize old or new AI-based products and services, and assess the effectiveness of decarbonization efforts.
Vergallo, Roberto; Mainetti, Luca; Caione, Adriana; Merico, Benedetta Maria; Matino, Sara
Digitalizing Bank Branches with a New Multichannel Standard to Address Office Closures in Europe Book Section
In: Paolis, Lucio Tommaso De; Arpaia, Pasquale; Sacco, Marco (Ed.): Extended Reality, vol. 15740, pp. 173–192, Springer Nature Switzerland, Cham, 2025, ISBN: 978-3-031-97771-8 978-3-031-97772-5, (Series Title: Lecture Notes in Computer Science).
@incollection{de_paolis_digitalizing_2025,
title = {Digitalizing Bank Branches with a New Multichannel Standard to Address Office Closures in Europe},
author = {Roberto Vergallo and Luca Mainetti and Adriana Caione and Benedetta Maria Merico and Sara Matino},
editor = {Lucio Tommaso De Paolis and Pasquale Arpaia and Marco Sacco},
url = {https://link.springer.com/10.1007/978-3-031-97772-5_12},
doi = {10.1007/978-3-031-97772-5_12},
isbn = {978-3-031-97771-8 978-3-031-97772-5},
year = {2025},
date = {2025-01-01},
urldate = {2025-09-30},
booktitle = {Extended Reality},
volume = {15740},
pages = {173–192},
publisher = {Springer Nature Switzerland},
address = {Cham},
note = {Series Title: Lecture Notes in Computer Science},
keywords = {},
pubstate = {published},
tppubtype = {incollection}
}
Gallo, Alba Maria; Ippolito, Adelaide; Margherita, Smarra; Sorrentino, Marco
Sustainability reporting: Weighing benefits and costs in the wake of the EU’s Omnibus Package Proceedings Article
In: pp. 55–60, Virtus Interpress, Ucraina, 2025, ISBN: 978-617-7309-35-1.
@inproceedings{gallo_sustainability_2025,
title = {Sustainability reporting: Weighing benefits and costs in the wake of the EU’s Omnibus Package},
author = {Alba Maria Gallo and Adelaide Ippolito and Smarra Margherita and Marco Sorrentino},
doi = {https://doi.org/10.22495/cgiop11},
isbn = {978-617-7309-35-1},
year = {2025},
date = {2025-01-01},
pages = {55–60},
publisher = {Virtus Interpress},
address = {Ucraina},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Angelis, Grazia De; Saviano, Emilio; Scuotto, Chiara; Iannello, Nicoló M.; Generosi, Andrea; Luca, Valerio De; Rega, Angelo; Triberti, Stefano; Affuso, Gaetana; Gallo, Luigi; Limone, Pierpaolo
2025, (Accepted: 2025-10-16T18:01:10Z).
Abstract | Links | BibTeX | Tags: Psychology, User study, Virtual Reality
@unpublished{de_angelis_utilitarian_2025,
title = {Utilitarian choice and self-serving cognitive distortions: using VR to investigate adolescents’ responses and enhance interventions},
author = {Grazia De Angelis and Emilio Saviano and Chiara Scuotto and Nicoló M. Iannello and Andrea Generosi and Valerio De Luca and Angelo Rega and Stefano Triberti and Gaetana Affuso and Luigi Gallo and Pierpaolo Limone},
url = {https://www.unipegaso.iris.cineca.it/handle/20.500.12607/62203},
year = {2025},
date = {2025-01-01},
urldate = {2026-02-06},
abstract = {11},
note = {Accepted: 2025-10-16T18:01:10Z},
keywords = {Psychology, User study, Virtual Reality},
pubstate = {published},
tppubtype = {unpublished}
}
11
Neroni, Pietro; Caggianese, Giuseppe; Esposito, Ciro; Gallo, Luigi
Real-Time 3D Posture Tracking for Surgeons in Pediatric Minimally Invasive Surgery: Proceedings Article
In: Proceedings of the 17th International Conference on Computer Supported Education, pp. 921–928, SCITEPRESS - Science and Technology Publications, Porto, Portugal, 2025, ISBN: 978-989-758-746-7.
Abstract | Links | BibTeX | Tags: Healthcare, Human Computer Interaction, Surgery, Tracking
@inproceedings{neroni_real-time_2025,
title = {Real-Time 3D Posture Tracking for Surgeons in Pediatric Minimally Invasive Surgery:},
author = {Pietro Neroni and Giuseppe Caggianese and Ciro Esposito and Luigi Gallo},
url = {https://www.scitepress.org/DigitalLibrary/Link.aspx?doi=10.5220/0013504200003932},
doi = {10.5220/0013504200003932},
isbn = {978-989-758-746-7},
year = {2025},
date = {2025-01-01},
urldate = {2025-05-06},
booktitle = {Proceedings of the 17th International Conference on Computer Supported Education},
pages = {921–928},
publisher = {SCITEPRESS - Science and Technology Publications},
address = {Porto, Portugal},
abstract = {Minimally invasive pediatric surgery presents ergonomic challenges that significantly increase the risk of musculoskeletal disorders (MSDs) among surgeons due to prolonged periods of suboptimal posture. This study introduces a real-time posture monitoring and correction system designed to address this issue. The system utilizes depth camera technology, interactive feedback mechanisms, advanced skeletal tracking, and ergonomic assessment algorithms to continuously monitor and evaluate surgeons’ posture. Through rapid data processing, the system provides real-time feedback, enabling immediate posture adjustments during surgical procedures. It delivers non-intrusive alerts to inform medical staff when incorrect postures are detected, thereby promoting ergonomic well-being and reducing the incidence of MSDs. Designed for seamless integration into the perioperative environment, the system meets strict requirements for privacy, sterility, and operational efficiency. Beyond its application in surgical practice, the system can also enhance surgical education and training by providing real-time feedback, enabling personalized learning pathways, and gamified simulation exercises. It provides detailed analyses of trainee performance, enabling instructors to deliver targeted feedback and develop adaptive training strategies based on detected posture deviations.},
keywords = {Healthcare, Human Computer Interaction, Surgery, Tracking},
pubstate = {published},
tppubtype = {inproceedings}
}
Minimally invasive pediatric surgery presents ergonomic challenges that significantly increase the risk of musculoskeletal disorders (MSDs) among surgeons due to prolonged periods of suboptimal posture. This study introduces a real-time posture monitoring and correction system designed to address this issue. The system utilizes depth camera technology, interactive feedback mechanisms, advanced skeletal tracking, and ergonomic assessment algorithms to continuously monitor and evaluate surgeons’ posture. Through rapid data processing, the system provides real-time feedback, enabling immediate posture adjustments during surgical procedures. It delivers non-intrusive alerts to inform medical staff when incorrect postures are detected, thereby promoting ergonomic well-being and reducing the incidence of MSDs. Designed for seamless integration into the perioperative environment, the system meets strict requirements for privacy, sterility, and operational efficiency. Beyond its application in surgical practice, the system can also enhance surgical education and training by providing real-time feedback, enabling personalized learning pathways, and gamified simulation exercises. It provides detailed analyses of trainee performance, enabling instructors to deliver targeted feedback and develop adaptive training strategies based on detected posture deviations.
Niccoli, Niccolò; Galteri, Leonardo; Seidenari, Lorenzo
Diffusion Autoencoders are Foundation Video Compressors Proceedings Article
In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 3964–3972, 2025.
BibTeX | Tags:
@inproceedings{niccoli_diffusion_2025,
title = {Diffusion Autoencoders are Foundation Video Compressors},
author = {Niccolò Niccoli and Leonardo Galteri and Lorenzo Seidenari},
year = {2025},
date = {2025-01-01},
booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision},
pages = {3964–3972},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Ferrara, Rita; Galteri, Leonardo
Prompt-Engineered Detection of AI-Generated Images Proceedings Article
In: 2025.
Abstract | Links | BibTeX | Tags:
@inproceedings{ferrara_prompt-engineered_2025,
title = {Prompt-Engineered Detection of AI-Generated Images},
author = {Rita Ferrara and Leonardo Galteri},
url = {https://ebooks.iospress.nl/volumearticle/74930},
year = {2025},
date = {2025-01-01},
abstract = {The proliferation of highly realistic AI-generated images presents significant challenges related to authenticity and misinformation. Although multimodal large language models (LLMs) possess advanced visual understanding capabilities, their effectiveness in distinguishing synthetic images from real ones requires systematic evaluation. This paper investigates the ability of four prominent LLMs—GPT-4V, Gemini, LLaVA, and Claude to detect AI-generated images. Using a case study methodology, twelve diverse synthetic images were presented to the models using ten distinct prompts, ranging from generic classification requests to detailed, forensically-guided queries. The study analyzes the accuracy and response patterns of each LLM, with a particular focus on the impact of prompt specificity and iterative refinement (prompt engineering) on detection performance. The results indicate that while GPT-4V demonstrated superior consistency, the performance of all tested LLMs, particularly Gemini, Claude, and LLaVA, was significantly influenced by prompt quality. Specific, detailed prompts markedly improved detection accuracy compared to generic ones. The findings underscore that effective prompt engineering is crucial for leveraging LLMs as tools for synthetic image detection and highlight the need for skilled human interaction to guide these systems to achieve reliable results. This research contributes to understanding the potential and current limitations of LLMs in addressing the challenges posed by synthetic media.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The proliferation of highly realistic AI-generated images presents significant challenges related to authenticity and misinformation. Although multimodal large language models (LLMs) possess advanced visual understanding capabilities, their effectiveness in distinguishing synthetic images from real ones requires systematic evaluation. This paper investigates the ability of four prominent LLMs—GPT-4V, Gemini, LLaVA, and Claude to detect AI-generated images. Using a case study methodology, twelve diverse synthetic images were presented to the models using ten distinct prompts, ranging from generic classification requests to detailed, forensically-guided queries. The study analyzes the accuracy and response patterns of each LLM, with a particular focus on the impact of prompt specificity and iterative refinement (prompt engineering) on detection performance. The results indicate that while GPT-4V demonstrated superior consistency, the performance of all tested LLMs, particularly Gemini, Claude, and LLaVA, was significantly influenced by prompt quality. Specific, detailed prompts markedly improved detection accuracy compared to generic ones. The findings underscore that effective prompt engineering is crucial for leveraging LLMs as tools for synthetic image detection and highlight the need for skilled human interaction to guide these systems to achieve reliable results. This research contributes to understanding the potential and current limitations of LLMs in addressing the challenges posed by synthetic media.
Rienzo, Marco Di; Bruni, Matteo; Galteri, Leonardo; Becattini, Federico; Bertini, Marco
Multi-Frame Alignment for Video Super-Resolution using Attention Proceedings Article
In: 2025.
@inproceedings{di_rienzo_multi-frame_2025,
title = {Multi-Frame Alignment for Video Super-Resolution using Attention},
author = {Marco Di Rienzo and Matteo Bruni and Leonardo Galteri and Federico Becattini and Marco Bertini},
year = {2025},
date = {2025-01-01},
abstract = {This paper presents a novel attention-based video super-resolution (VSR) method that avoids costly optical flow estimation while effectively exploiting temporal correlations between frames. We propose an aligner module that utilizes cross-attention to blend relevant patches from adjacent frames, gathering information from multiple frames simultaneously. This method improves upon traditional flow-based approaches by working at a block level and enabling the blending of several pixels, yielding better alignment for larger motions. The proposed VSR technique can upscale videos up to 4x while simultaneously removing compression artifacts, enhancing both resolution and quality. Experimental results demonstrate the effectiveness of this approach compared to classic flow-based methods, particularly in handling compressed videos where compression artifacts can severely impact optical flow estimation.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
This paper presents a novel attention-based video super-resolution (VSR) method that avoids costly optical flow estimation while effectively exploiting temporal correlations between frames. We propose an aligner module that utilizes cross-attention to blend relevant patches from adjacent frames, gathering information from multiple frames simultaneously. This method improves upon traditional flow-based approaches by working at a block level and enabling the blending of several pixels, yielding better alignment for larger motions. The proposed VSR technique can upscale videos up to 4x while simultaneously removing compression artifacts, enhancing both resolution and quality. Experimental results demonstrate the effectiveness of this approach compared to classic flow-based methods, particularly in handling compressed videos where compression artifacts can severely impact optical flow estimation.
Aversano, Lerina; Iammarino, Martina; Madau, Antonella; Montano, Debora; Verdone, Chiara
An Explainable Model for Waste Cost Prediction: A Study on Linked Open Data in Italy: Proceedings Article
In: Proceedings of the 20th International Conference on Software Technologies, pp. 446–453, SCITEPRESS - Science and Technology Publications, Bilbao, Spain, 2025, ISBN: 978-989-758-757-3.
@inproceedings{aversano_explainable_2025,
title = {An Explainable Model for Waste Cost Prediction: A Study on Linked Open Data in Italy:},
author = {Lerina Aversano and Martina Iammarino and Antonella Madau and Debora Montano and Chiara Verdone},
url = {https://www.scitepress.org/DigitalLibrary/Link.aspx?doi=10.5220/0013650800003964},
doi = {10.5220/0013650800003964},
isbn = {978-989-758-757-3},
year = {2025},
date = {2025-01-01},
urldate = {2025-11-04},
booktitle = {Proceedings of the 20th International Conference on Software Technologies},
pages = {446–453},
publisher = {SCITEPRESS - Science and Technology Publications},
address = {Bilbao, Spain},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
Aversano, Lerina; Iammarino, Martina; Madau, Antonella; Pirlo, Giuseppe; Semeraro, Gianfranco
Process mining applications in healthcare: a systematic literature review Journal Article
In: PeerJ Computer Science, vol. 11, pp. e2613, 2025, ISSN: 2376-5992.
Abstract | Links | BibTeX | Tags:
@article{aversano_process_2025,
title = {Process mining applications in healthcare: a systematic literature review},
author = {Lerina Aversano and Martina Iammarino and Antonella Madau and Giuseppe Pirlo and Gianfranco Semeraro},
url = {https://peerj.com/articles/cs-2613},
doi = {10.7717/peerj-cs.2613},
issn = {2376-5992},
year = {2025},
date = {2025-01-01},
urldate = {2025-11-04},
journal = {PeerJ Computer Science},
volume = {11},
pages = {e2613},
abstract = {Process mining applications in healthcare is a field widely investigated in the last years. Its diffusion is driven by increasing digitalization and the availability of large quantities of clinical data, enabling hospitals, clinics, and other healthcare organizations to optimize workflows, reduce operational costs, and improve asset management. The importance of process mining lies in its potential to identify inefficiencies in processes, standardize clinical practices, support evidence-based decisions and, in general, improve the quality of care provided. The article aims to systematically review the research landscape in the field of process mining in healthcare, providing an in-depth understanding of how process mining is applied in healthcare. It contributes to the existing literature by highlighting the following aspects: the specific research topics covered (i), the extent of use of various process mining algorithms in different healthcare applications, showing their adaptability and effectiveness in specific contexts (ii), and, finally, the types and characteristics of data employed in these studies, highlighting the needs and challenges related to data in healthcare process mining (iii). Through this systematic literature review, the article can support researchers in identifying the most valuable research topic to be explored by the scientific community working on process mining in healthcare. To achieve this goal, several articles focusing on the algorithms and data employed were selected and analyzed. The final discussion highlights current research gaps, suggesting future areas of investigation, and identifies critical issues and vulnerabilities of existing process mining applications in healthcare.},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Process mining applications in healthcare is a field widely investigated in the last years. Its diffusion is driven by increasing digitalization and the availability of large quantities of clinical data, enabling hospitals, clinics, and other healthcare organizations to optimize workflows, reduce operational costs, and improve asset management. The importance of process mining lies in its potential to identify inefficiencies in processes, standardize clinical practices, support evidence-based decisions and, in general, improve the quality of care provided. The article aims to systematically review the research landscape in the field of process mining in healthcare, providing an in-depth understanding of how process mining is applied in healthcare. It contributes to the existing literature by highlighting the following aspects: the specific research topics covered (i), the extent of use of various process mining algorithms in different healthcare applications, showing their adaptability and effectiveness in specific contexts (ii), and, finally, the types and characteristics of data employed in these studies, highlighting the needs and challenges related to data in healthcare process mining (iii). Through this systematic literature review, the article can support researchers in identifying the most valuable research topic to be explored by the scientific community working on process mining in healthcare. To achieve this goal, several articles focusing on the algorithms and data employed were selected and analyzed. The final discussion highlights current research gaps, suggesting future areas of investigation, and identifies critical issues and vulnerabilities of existing process mining applications in healthcare.
Aversano, Lerina; Iammarino, Martina; Madau, Antonella; Montano, Debora; Verdone, Chiara
Repairing Missing Activity Labels in Healthcare Process Logs: a Machine Learning Approach Book Section
In: Chen, Yen-Wei; Tanaka, Satoshi; Howlett, Robert J.; Jain, Lakhmi C. (Ed.): Innovation in Medicine and Healthcare, vol. 412, pp. 91–101, Springer Nature Singapore, Singapore, 2025, ISBN: 978-981-97-7497-5 978-981-97-7498-2, (Series Title: Smart Innovation, Systems and Technologies).
@incollection{chen_repairing_2025,
title = {Repairing Missing Activity Labels in Healthcare Process Logs: a Machine Learning Approach},
author = {Lerina Aversano and Martina Iammarino and Antonella Madau and Debora Montano and Chiara Verdone},
editor = {Yen-Wei Chen and Satoshi Tanaka and Robert J. Howlett and Lakhmi C. Jain},
url = {https://link.springer.com/10.1007/978-981-97-7498-2_9},
doi = {10.1007/978-981-97-7498-2_9},
isbn = {978-981-97-7497-5 978-981-97-7498-2},
year = {2025},
date = {2025-01-01},
urldate = {2025-10-22},
booktitle = {Innovation in Medicine and Healthcare},
volume = {412},
pages = {91–101},
publisher = {Springer Nature Singapore},
address = {Singapore},
note = {Series Title: Smart Innovation, Systems and Technologies},
keywords = {},
pubstate = {published},
tppubtype = {incollection}
}
Ambretti, Antinea; Monacis, Domenico; Forte, Pasqualina; Morsanuto, Stefania; Savoia, Teresa; D’Anna, Cristiana; Iammarino, Martina
Future perspectives for physical education in primary school to promote innovative co-planning practices Journal Article
In: Journal of Inclusive Methodology and Technology in Learning and Teaching, vol. 5, no. 1, 2025, ISSN: 2785-5104.
@article{ambretti_future_2025,
title = {Future perspectives for physical education in primary school to promote innovative co-planning practices},
author = {Antinea Ambretti and Domenico Monacis and Pasqualina Forte and Stefania Morsanuto and Teresa Savoia and Cristiana D’Anna and Martina Iammarino},
url = {https://www.inclusiveteaching.it/index.php/inclusiveteaching/article/view/300},
issn = {2785-5104},
year = {2025},
date = {2025-01-01},
journal = {Journal of Inclusive Methodology and Technology in Learning and Teaching},
volume = {5},
number = {1},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Martens, Julian; Kumara, Indika; Nucci, Dario Di; Pecorelli, Fabiano; Monsieur, Geert; Tamburri, Damian Andrew; Heuvel, Willem-Jan Van Den
Acceptance and Development of Quantum Computing in the Netherlands and Germany: Barriers and Remedies From a Multistakeholder Perspective Journal Article
In: IEEE Transactions on Engineering Management, vol. 72, pp. 62–77, 2025, ISSN: 0018-9391, 1558-0040.
@article{martens_acceptance_2025,
title = {Acceptance and Development of Quantum Computing in the Netherlands and Germany: Barriers and Remedies From a Multistakeholder Perspective},
author = {Julian Martens and Indika Kumara and Dario Di Nucci and Fabiano Pecorelli and Geert Monsieur and Damian Andrew Tamburri and Willem-Jan Van Den Heuvel},
url = {https://ieeexplore.ieee.org/document/10746616/},
doi = {10.1109/TEM.2024.3493600},
issn = {0018-9391, 1558-0040},
year = {2025},
date = {2025-01-01},
urldate = {2025-09-30},
journal = {IEEE Transactions on Engineering Management},
volume = {72},
pages = {62–77},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Barbareschi, Mario; Barone, Salvatore; Mazzocca, Nicola; Moriconi, Alberto
Designing Energy-Efficient Approximate Circuits for the FPGA Technology Proceedings Article
In: Proceedings of the 28th Euromicro Conference on Digital System Design (DSD), Salerno, 2025.
@inproceedings{barbareschi_designing_2025,
title = {Designing Energy-Efficient Approximate Circuits for the FPGA Technology},
author = {Mario Barbareschi and Salvatore Barone and Nicola Mazzocca and Alberto Moriconi},
year = {2025},
date = {2025-01-01},
booktitle = {Proceedings of the 28th Euromicro Conference on Digital System Design (DSD)},
address = {Salerno},
abstract = {A significant number of contributions focusing on approximation techniques for Application Specific Integrated Circuits (ASIC) has become part of the scientific literature. Conversely, Field Programmable Gate Arrays (FPGAs) are often overlooked, despite their increasing spread. ASIC-based techniques are, however, often unsuitable when it comes to FPGA, providing little or no advantages at all, due to the inherent differences in the two target technologies. Most of the FPGA-based approximation techniques being recently proposed either rely on manual approximation, or are too tightly coupled with a particular FPGA fabric, making them ineffective or even inapplicable to other devices. Furthermore, they rely on machine-learning based predictors to drive the Design Space Exploration (DSE), that, given the high fidelity required, are usually burdensome to achieve, or even unfeasible when little or no training data is available. In this paper, we discuss a fabricand workload-independent approach to design power-optimized approximate circuits for FPGA. We exploit the existing bond between Look-Up Table (LUT)-mapping in FPGA synthesis and cut-enumeration in And-Inverter graph representation of digital circuits, and we resort to an analytical model to estimate the power consumption during the DSE, avoiding costly synthesis as well as machine-learning based predictors for hardware resources during the DSE. Several benchmark circuits are considered for evaluation purposes, and the significant savings achieved allow us to claim our approach is suitable for addressing approximate circuit design while targeting the FPGA.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
A significant number of contributions focusing on approximation techniques for Application Specific Integrated Circuits (ASIC) has become part of the scientific literature. Conversely, Field Programmable Gate Arrays (FPGAs) are often overlooked, despite their increasing spread. ASIC-based techniques are, however, often unsuitable when it comes to FPGA, providing little or no advantages at all, due to the inherent differences in the two target technologies. Most of the FPGA-based approximation techniques being recently proposed either rely on manual approximation, or are too tightly coupled with a particular FPGA fabric, making them ineffective or even inapplicable to other devices. Furthermore, they rely on machine-learning based predictors to drive the Design Space Exploration (DSE), that, given the high fidelity required, are usually burdensome to achieve, or even unfeasible when little or no training data is available. In this paper, we discuss a fabricand workload-independent approach to design power-optimized approximate circuits for FPGA. We exploit the existing bond between Look-Up Table (LUT)-mapping in FPGA synthesis and cut-enumeration in And-Inverter graph representation of digital circuits, and we resort to an analytical model to estimate the power consumption during the DSE, avoiding costly synthesis as well as machine-learning based predictors for hardware resources during the DSE. Several benchmark circuits are considered for evaluation purposes, and the significant savings achieved allow us to claim our approach is suitable for addressing approximate circuit design while targeting the FPGA.
Barbareschi, Mario; Barone, Salvatore; Casola, Valentina; Torca, Salvatore Della
Error Resiliency and Adversarial Robustness in Convolutional Neural Networks: An Empirical Analysis Proceedings Article
In: Rey, Gaëtan; Tigli, Jean-Yves; Franquet, Erwin (Ed.): Internet of Things, pp. 149–160, Springer Nature Switzerland, Cham, 2025, ISBN: 978-3-031-81900-1.
Abstract | Links | BibTeX | Tags:
@inproceedings{barbareschi_error_2025,
title = {Error Resiliency and Adversarial Robustness in Convolutional Neural Networks: An Empirical Analysis},
author = {Mario Barbareschi and Salvatore Barone and Valentina Casola and Salvatore Della Torca},
editor = {Gaëtan Rey and Jean-Yves Tigli and Erwin Franquet},
url = {https://link.springer.com/chapter/10.1007/978-3-031-81900-1_9},
doi = {10.1007/978-3-031-81900-1_9},
isbn = {978-3-031-81900-1},
year = {2025},
date = {2025-01-01},
booktitle = {Internet of Things},
pages = {149–160},
publisher = {Springer Nature Switzerland},
address = {Cham},
abstract = {The increasing pervasiveness of Artificial Intelligence (AI), and Convolutional Neural Networks (CNNs) in edge-computing and Internet of Things applications pose several challenges, including the hunger for computational and power resources of predictive models, and their robustness w.r.t. security threats, e.g., adversarial attacks. As for the former, the approximate computing emerged as one of the most promising solutions to lower the computational effort of AI, since the output of approximate application is usually barely distinguishable from the exact one. Nevertheless, alterations to predictive models through approximation may actually jeopardize inner characteristics of CNNs, such as their adversarial robustness, that is their ability to discern legitimate inputs from systematically crafted malicious ones.},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
The increasing pervasiveness of Artificial Intelligence (AI), and Convolutional Neural Networks (CNNs) in edge-computing and Internet of Things applications pose several challenges, including the hunger for computational and power resources of predictive models, and their robustness w.r.t. security threats, e.g., adversarial attacks. As for the former, the approximate computing emerged as one of the most promising solutions to lower the computational effort of AI, since the output of approximate application is usually barely distinguishable from the exact one. Nevertheless, alterations to predictive models through approximation may actually jeopardize inner characteristics of CNNs, such as their adversarial robustness, that is their ability to discern legitimate inputs from systematically crafted malicious ones.
Barbareschi, Mario; Barone, Salvatore; Bosio, Alberto; Deveautour, Bastien; Piri, Ali; Traiola, Marcello
Automatic generation of input-aware approximate arithmetic circuits Proceedings Article
In: 2025 IEEE 28th International Symposium on Design and Diagnostics of Electronic Circuits and Systems (DDECS), pp. 139–144, IEEE, 2025.
@inproceedings{barbareschi_automatic_2025,
title = {Automatic generation of input-aware approximate arithmetic circuits},
author = {Mario Barbareschi and Salvatore Barone and Alberto Bosio and Bastien Deveautour and Ali Piri and Marcello Traiola},
url = {https://ieeexplore.ieee.org/abstract/document/11006680/},
year = {2025},
date = {2025-01-01},
urldate = {2025-09-16},
booktitle = {2025 IEEE 28th International Symposium on Design and Diagnostics of Electronic Circuits and Systems (DDECS)},
pages = {139–144},
publisher = {IEEE},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}