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
2026
Mondal, Semanto; Ferraro, Antonino; Pecorelli, Fabiano; Iammarino, Martina; Pietro, Giuseppe De
Concept and rule guided neural network for early crop leaf nutrient deficiency diagnosis Journal Article
In: Computers and Electronics in Agriculture, vol. 248, pp. 111735, 2026, ISSN: 01681699.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Deep Learning, Early crop nutrition deficiency, Fuzzy logic, Neurosymbolic, RAG, ResNet
@article{mondal_concept_2026,
title = {Concept and rule guided neural network for early crop leaf nutrient deficiency diagnosis},
author = {Semanto Mondal and Antonino Ferraro and Fabiano Pecorelli and Martina Iammarino and Giuseppe De Pietro},
url = {https://linkinghub.elsevier.com/retrieve/pii/S0168169926003303},
doi = {10.1016/j.compag.2026.111735},
issn = {01681699},
year = {2026},
date = {2026-07-01},
urldate = {2026-06-16},
journal = {Computers and Electronics in Agriculture},
volume = {248},
pages = {111735},
abstract = {Crop nutrition deficiency poses a major challenge to achieving optimal yield, particularly in smallholder farming systems where timely expert diagnosis is limited. Early detection is crucial to minimize losses and reduce unnecessary fertilizer or pesticide usage. While deep learning offers potential for automated visual diagnosis, most existing approaches operate as black boxes and lack interpretability, explainability, or actionable recommendations. In this work, we present a neurosymbolic framework for early nutrient deficiency detection in ash gourd leaves using the EarlyNSD dataset. Our approach integrates a ResNet-50 backbone with a dual-head design: a classification head for deficiency prediction and a concept-prediction head that quantifies physiologically meaningful visual patterns such as yellowing, edge discoloration, spots, and vein greenness. These concept scores are combined with predefined domain rules to guide the learning of the neural component and to generate transparent, human-aligned explanations for each diagnosis. Building on the model outputs, we incorporate a Retrieval Augmented Generation (RAG)-based pipeline along with an agricultural knowledge base to generate targeted recommendations. This approach overcomes key shortcomings of pure neural models by incorporating domain knowledge in the form of differentiable fuzzy logic rules. The study demonstrates that the proposed framework improves both classification performance and interpretability compared to standard ResNet baselines. Grad-CAM analysis demonstrates that concept-guided attention aligns with symptom-specific regions, such as yellowed areas for Nitrogen deficiency or marginal discoloration for Potassium deficiency, providing visual validation of the reasoning process. Since EarlyNSD is limited in scale and visual diversity, the results are not directly comparable to large open-field datasets. Overall, our results establish a proof of concept for integrating neural detection with symbolic reasoning, enabling interpretable, actionable, and domain-informed nutrient management for practical applications.},
keywords = {Artificial Intelligence, Deep Learning, Early crop nutrition deficiency, Fuzzy logic, Neurosymbolic, RAG, ResNet},
pubstate = {published},
tppubtype = {article}
}
Crop nutrition deficiency poses a major challenge to achieving optimal yield, particularly in smallholder farming systems where timely expert diagnosis is limited. Early detection is crucial to minimize losses and reduce unnecessary fertilizer or pesticide usage. While deep learning offers potential for automated visual diagnosis, most existing approaches operate as black boxes and lack interpretability, explainability, or actionable recommendations. In this work, we present a neurosymbolic framework for early nutrient deficiency detection in ash gourd leaves using the EarlyNSD dataset. Our approach integrates a ResNet-50 backbone with a dual-head design: a classification head for deficiency prediction and a concept-prediction head that quantifies physiologically meaningful visual patterns such as yellowing, edge discoloration, spots, and vein greenness. These concept scores are combined with predefined domain rules to guide the learning of the neural component and to generate transparent, human-aligned explanations for each diagnosis. Building on the model outputs, we incorporate a Retrieval Augmented Generation (RAG)-based pipeline along with an agricultural knowledge base to generate targeted recommendations. This approach overcomes key shortcomings of pure neural models by incorporating domain knowledge in the form of differentiable fuzzy logic rules. The study demonstrates that the proposed framework improves both classification performance and interpretability compared to standard ResNet baselines. Grad-CAM analysis demonstrates that concept-guided attention aligns with symptom-specific regions, such as yellowed areas for Nitrogen deficiency or marginal discoloration for Potassium deficiency, providing visual validation of the reasoning process. Since EarlyNSD is limited in scale and visual diversity, the results are not directly comparable to large open-field datasets. Overall, our results establish a proof of concept for integrating neural detection with symbolic reasoning, enabling interpretable, actionable, and domain-informed nutrient management for practical applications.
Karthik, Gandrothu; Rupesh, Namburi; John, Joel; Raj, Rayappa David Amar; Tomazzoli, Claudio; Randieri, Cristian
In: Drones, vol. 10, no. 6, pp. 422, 2026, ISSN: 2504-446X.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Drones, Edge AI, Edge computing
@article{karthikLightweightEdgeAI2026,
title = {Lightweight Edge AI Hardware-Oriented Photovoltaic Fault Detection Using Generative Augmentation with Potential Drone-Based Inspection Applications},
author = {Gandrothu Karthik and Namburi Rupesh and Joel John and Rayappa David Amar Raj and Claudio Tomazzoli and Cristian Randieri},
url = {https://www.mdpi.com/2504-446X/10/6/422},
doi = {10.3390/drones10060422},
issn = {2504-446X},
year = {2026},
date = {2026-05-01},
urldate = {2026-06-18},
journal = {Drones},
volume = {10},
number = {6},
pages = {422},
abstract = {To ensure the reliability and sustained performance of industrial photovoltaic (PV) systems, fault detection frameworks must achieve both high detection accuracy and computational efficiency, particularly for deployment on resource-constrained edge platforms. This work proposes a lightweight and low-latency photovoltaic defect detection framework that integrates DCGAN-based generative augmentation with the proposed GhostViT-YOLOv10n architecture. The augmentation strategy helps address class imbalance, improve representation of rare defects, and enhance generalization capability in electroluminescence (EL) imagery through structured geometric and photometric transformations. The proposed framework integrates lightweight Ghost-based optimization, Cross-Stage Partial Fusion (C2f), Spatial Pyramid Pooling—Fast (SPPF), MobileViT contextual learning, and SimAM-based attention refinement to improve multi-scale feature extraction while maintaining low computational complexity. Experimental evaluation on the PVEL-AD and PV Multi Defect benchmark datasets demonstrates strong detection performance. On the PVEL-AD dataset, the BaseLine achieves a mAP@0.5 of 71.6% with only 2.7 M parameters and 8.4 GFLOPs, while our proposed GhostViT-YOLOv10n framework with DCGAN-enhanced version further improves detection performance to 93.6% mAP@0.5 with only 2.19 M parameters and 6.6 GFLOPs. On the PV Multi Defect dataset, the BaseLine achieves a mAP@0.5 of 74.0% with 2.71 M parameters and 8.4 GFLOPs, and the optimized framework with DCGAN-augmented configuration further improves performance to 95.4% mAP@0.5 with 2.58 M parameters and 7.7 GFLOPs. These results demonstrate the effectiveness of combining lightweight architectural optimization with generative augmentation for improving rare defect representation and multi-scale photovoltaic defect detection. To validate practical deployment feasibility, the optimized framework was deployed on a Raspberry Pi 5 using ONNX Runtime under CPU-only conditions. The deployed model achieved an average inference time of 43.05 ms and a real-time processing speed of 23.23 FPS while maintaining moderate CPU utilization and stable thermal behavior. These deployment results demonstrate the suitability of the proposed framework for lightweight edge-oriented photovoltaic inspection applications without requiring GPU acceleration. All evaluations were conducted exclusively on real test datasets, while synthetic samples were used only during training to improve data diversity and rare defect representation. Overall, the proposed framework provides a balanced solution that combines detection accuracy, computational efficiency, lightweight edge deployment capability, and generative augmentation for practical photovoltaic defect inspection applications with potential suitability for future drone-assisted inspection scenarios.},
keywords = {Artificial Intelligence, Drones, Edge AI, Edge computing},
pubstate = {published},
tppubtype = {article}
}
To ensure the reliability and sustained performance of industrial photovoltaic (PV) systems, fault detection frameworks must achieve both high detection accuracy and computational efficiency, particularly for deployment on resource-constrained edge platforms. This work proposes a lightweight and low-latency photovoltaic defect detection framework that integrates DCGAN-based generative augmentation with the proposed GhostViT-YOLOv10n architecture. The augmentation strategy helps address class imbalance, improve representation of rare defects, and enhance generalization capability in electroluminescence (EL) imagery through structured geometric and photometric transformations. The proposed framework integrates lightweight Ghost-based optimization, Cross-Stage Partial Fusion (C2f), Spatial Pyramid Pooling—Fast (SPPF), MobileViT contextual learning, and SimAM-based attention refinement to improve multi-scale feature extraction while maintaining low computational complexity. Experimental evaluation on the PVEL-AD and PV Multi Defect benchmark datasets demonstrates strong detection performance. On the PVEL-AD dataset, the BaseLine achieves a mAP@0.5 of 71.6% with only 2.7 M parameters and 8.4 GFLOPs, while our proposed GhostViT-YOLOv10n framework with DCGAN-enhanced version further improves detection performance to 93.6% mAP@0.5 with only 2.19 M parameters and 6.6 GFLOPs. On the PV Multi Defect dataset, the BaseLine achieves a mAP@0.5 of 74.0% with 2.71 M parameters and 8.4 GFLOPs, and the optimized framework with DCGAN-augmented configuration further improves performance to 95.4% mAP@0.5 with 2.58 M parameters and 7.7 GFLOPs. These results demonstrate the effectiveness of combining lightweight architectural optimization with generative augmentation for improving rare defect representation and multi-scale photovoltaic defect detection. To validate practical deployment feasibility, the optimized framework was deployed on a Raspberry Pi 5 using ONNX Runtime under CPU-only conditions. The deployed model achieved an average inference time of 43.05 ms and a real-time processing speed of 23.23 FPS while maintaining moderate CPU utilization and stable thermal behavior. These deployment results demonstrate the suitability of the proposed framework for lightweight edge-oriented photovoltaic inspection applications without requiring GPU acceleration. All evaluations were conducted exclusively on real test datasets, while synthetic samples were used only during training to improve data diversity and rare defect representation. Overall, the proposed framework provides a balanced solution that combines detection accuracy, computational efficiency, lightweight edge deployment capability, and generative augmentation for practical photovoltaic defect inspection applications with potential suitability for future drone-assisted inspection scenarios.
Nenna, Raffaella; Manti, Sara; Ferrante, Giuliana; Malizia, Velia; Alfano, Pietro; Parisi, Giuseppe Fabio; Regina, Domenico Paolo La; Gallo, Luigi; Pandolfo, Alessandra; Licari, Amelia; Grutta, Stefania La
Impact of digital technologies on pediatric asthma care Journal Article
In: Pediatric Respiratory Journal, vol. 04, no. 01, pp. 2–15, 2026, ISSN: 3035-2134.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Asthma, Healthcare, Mobile Application, Serious games
@article{nenna_impact_2026,
title = {Impact of digital technologies on pediatric asthma care},
author = {Raffaella Nenna and Sara Manti and Giuliana Ferrante and Velia Malizia and Pietro Alfano and Giuseppe Fabio Parisi and Domenico Paolo La Regina and Luigi Gallo and Alessandra Pandolfo and Amelia Licari and Stefania La Grutta},
url = {https://www.pediatric-respiratory-journal.com/impact-of-digital-technologies-on-pediatric-asthma-care/},
doi = {10.56164/PediatrRespirJ.2026.01},
issn = {3035-2134},
year = {2026},
date = {2026-04-01},
urldate = {2026-04-03},
journal = {Pediatric Respiratory Journal},
volume = {04},
number = {01},
pages = {2–15},
publisher = {Edra Media S.r.l.},
abstract = {Asthma is one of the most common chronic diseases in children, significantly impacting their health, quality of life, and healthcare systems globally. Pediatric asthma accounts for substantial morbidity, including frequent exacerbations, emergency department visits, and missed school days. Despite the availability of effective treatments and clear management guidelines, achieving optimal asthma control remains a challenge. In recent years, digital technologies have emerged as transformative tools in asthma care, offering new ways to monitor, educate, and treat pediatric patients. A systematic review was conducted to examine the impact of digital technologies on pediatric asthma care, synthesizing evidence on their effectiveness, challenges, and future directions. Covering studies from January 2020 to December 2024, the review analyzed 59 primary studies that involved mobile health (mHealth) applications, electronic medication monitoring systems, wearable devices, artificial intelligence (AI)-powered solutions, and school-based telemedicine programs. Findings reveal that mHealth applications and serious games promote self-management, improve medication adherence, and support patient education. Telemedicine, including school-based and remote patient monitoring, enhances care accessibility, reduces emergency visits, and promotes continuity of care, particularly in underserved populations. Wearable devices and electronic monitoring tools enhance symptom tracking and evaluation of inhaler technique. AI-driven interventions, such as digital twin systems, show promise in personalizing treatment and predicting exacerbations. Despite encouraging outcomes, challenges remain, including digital literacy gaps, limited access to devices and the internet, and difficulties integrating digital tools into clinical workflows. Usability and sustainability vary widely depending on design approaches, caregiver engagement, and infrastructure readiness.},
keywords = {Artificial Intelligence, Asthma, Healthcare, Mobile Application, Serious games},
pubstate = {published},
tppubtype = {article}
}
Asthma is one of the most common chronic diseases in children, significantly impacting their health, quality of life, and healthcare systems globally. Pediatric asthma accounts for substantial morbidity, including frequent exacerbations, emergency department visits, and missed school days. Despite the availability of effective treatments and clear management guidelines, achieving optimal asthma control remains a challenge. In recent years, digital technologies have emerged as transformative tools in asthma care, offering new ways to monitor, educate, and treat pediatric patients. A systematic review was conducted to examine the impact of digital technologies on pediatric asthma care, synthesizing evidence on their effectiveness, challenges, and future directions. Covering studies from January 2020 to December 2024, the review analyzed 59 primary studies that involved mobile health (mHealth) applications, electronic medication monitoring systems, wearable devices, artificial intelligence (AI)-powered solutions, and school-based telemedicine programs. Findings reveal that mHealth applications and serious games promote self-management, improve medication adherence, and support patient education. Telemedicine, including school-based and remote patient monitoring, enhances care accessibility, reduces emergency visits, and promotes continuity of care, particularly in underserved populations. Wearable devices and electronic monitoring tools enhance symptom tracking and evaluation of inhaler technique. AI-driven interventions, such as digital twin systems, show promise in personalizing treatment and predicting exacerbations. Despite encouraging outcomes, challenges remain, including digital literacy gaps, limited access to devices and the internet, and difficulties integrating digital tools into clinical workflows. Usability and sustainability vary widely depending on design approaches, caregiver engagement, and infrastructure readiness.
Gallo, Luigi; Carruba, Maria Concetta; Ferraro, Antonino; Lund, Henrik Hautop; Rega, Angelo; Triberti, Stefano
Editorial: AI innovations in education: adaptive learning and beyond Journal Article
In: Frontiers in Computer Science, vol. 8, pp. 1822456, 2026, ISSN: 2624-9898.
Abstract | Links | BibTeX | Tags: adaptive learning, AI literacy, Artificial Intelligence, artificial intelligence in education (AIeD), Education, ethical AI, Generative AI, human-AI synergy, multimodal learning analytics (MMLA), personalized learning
@article{gallo_editorial_2026,
title = {Editorial: AI innovations in education: adaptive learning and beyond},
author = {Luigi Gallo and Maria Concetta Carruba and Antonino Ferraro and Henrik Hautop Lund and Angelo Rega and Stefano Triberti},
url = {https://www.frontiersin.org/articles/10.3389/fcomp.2026.1822456/full},
doi = {10.3389/fcomp.2026.1822456},
issn = {2624-9898},
year = {2026},
date = {2026-03-01},
urldate = {2026-04-03},
journal = {Frontiers in Computer Science},
volume = {8},
pages = {1822456},
publisher = {Frontiers Media SA},
abstract = {Artificial Intelligence (AI) is gradually transforming educational practices. AI-powered teaching assistants, large language models, and multimodal analytics platforms are reshaping how learning experiences are designed and assessed. However, AI integration is not merely a technological matter: it is also heavily influenced by pedagogical, psychological, and sociocultural factors. Beyond technical implementation, AI systems can be framed within a human augmentation perspective, where technologies enhance sensory, motor, and cognitive processes in hybrid environments (Augello et al., 2022), including immersive contexts in which presence and cognition dynamically interact (Palombi et al., 2023). At the same time, advances in adaptive and data-driven AI, including explainable and diversity-aware approaches, highlight the role of algorithmic choices in shaping users' experiences and behaviors (Ferraro et al., 2025). Recent studies further show that, while educators are using AI tools for instructional purposes (e.g., creating teaching materials), concerns remain about the risk of unfair AI use and the difficulty of detecting it (Amato et al., 2023; Carruba et al., 2025). Accordingly, research on AI in education is becoming increasingly focused on factors that support implementation and adoption in real-life contexts, beyond mere improvement of algorithms from a computer science perspective (Triberti et al., 2024; Acosta-Enriquez et al., 2025; Galindo-Domĺnguez et al., 2024).
Against this backdrop, this Research Topic (RT), which spans four Frontiers journals, focuses on empirical and theoretical aspects of personalized and adaptive learning. More specifically, it examines the role of AI in fostering inclusive, data-driven, and emotionally responsive educational ecosystems, with particular attention to motivation, beliefs, creativity, self-regulation, and ethics.
For analytical clarity, the contributions are organized into six interrelated thematic areas: AI Adoption, Acceptance, and Self-Regulation; Teacher AI Literacy and Sustainable Integration; Adaptive, Immersive, and AI-Enhanced Learning Environments; Multimodal Analytics, Assessment, and Predictive AI; Learner Psychological, Cognitive, and Sociocultural Factors; Conceptual, Ethical, and Human-AI Synergy Perspectives. These areas reflect three broader dimensions of current research: the adoption of AI by learners and teachers, the design of intelligent learning environments, and the broader psychological and ethical implications of AI-supported education.},
keywords = {adaptive learning, AI literacy, Artificial Intelligence, artificial intelligence in education (AIeD), Education, ethical AI, Generative AI, human-AI synergy, multimodal learning analytics (MMLA), personalized learning},
pubstate = {published},
tppubtype = {article}
}
Artificial Intelligence (AI) is gradually transforming educational practices. AI-powered teaching assistants, large language models, and multimodal analytics platforms are reshaping how learning experiences are designed and assessed. However, AI integration is not merely a technological matter: it is also heavily influenced by pedagogical, psychological, and sociocultural factors. Beyond technical implementation, AI systems can be framed within a human augmentation perspective, where technologies enhance sensory, motor, and cognitive processes in hybrid environments (Augello et al., 2022), including immersive contexts in which presence and cognition dynamically interact (Palombi et al., 2023). At the same time, advances in adaptive and data-driven AI, including explainable and diversity-aware approaches, highlight the role of algorithmic choices in shaping users' experiences and behaviors (Ferraro et al., 2025). Recent studies further show that, while educators are using AI tools for instructional purposes (e.g., creating teaching materials), concerns remain about the risk of unfair AI use and the difficulty of detecting it (Amato et al., 2023; Carruba et al., 2025). Accordingly, research on AI in education is becoming increasingly focused on factors that support implementation and adoption in real-life contexts, beyond mere improvement of algorithms from a computer science perspective (Triberti et al., 2024; Acosta-Enriquez et al., 2025; Galindo-Domĺnguez et al., 2024).
Against this backdrop, this Research Topic (RT), which spans four Frontiers journals, focuses on empirical and theoretical aspects of personalized and adaptive learning. More specifically, it examines the role of AI in fostering inclusive, data-driven, and emotionally responsive educational ecosystems, with particular attention to motivation, beliefs, creativity, self-regulation, and ethics.
For analytical clarity, the contributions are organized into six interrelated thematic areas: AI Adoption, Acceptance, and Self-Regulation; Teacher AI Literacy and Sustainable Integration; Adaptive, Immersive, and AI-Enhanced Learning Environments; Multimodal Analytics, Assessment, and Predictive AI; Learner Psychological, Cognitive, and Sociocultural Factors; Conceptual, Ethical, and Human-AI Synergy Perspectives. These areas reflect three broader dimensions of current research: the adoption of AI by learners and teachers, the design of intelligent learning environments, and the broader psychological and ethical implications of AI-supported education.
Against this backdrop, this Research Topic (RT), which spans four Frontiers journals, focuses on empirical and theoretical aspects of personalized and adaptive learning. More specifically, it examines the role of AI in fostering inclusive, data-driven, and emotionally responsive educational ecosystems, with particular attention to motivation, beliefs, creativity, self-regulation, and ethics.
For analytical clarity, the contributions are organized into six interrelated thematic areas: AI Adoption, Acceptance, and Self-Regulation; Teacher AI Literacy and Sustainable Integration; Adaptive, Immersive, and AI-Enhanced Learning Environments; Multimodal Analytics, Assessment, and Predictive AI; Learner Psychological, Cognitive, and Sociocultural Factors; Conceptual, Ethical, and Human-AI Synergy Perspectives. These areas reflect three broader dimensions of current research: the adoption of AI by learners and teachers, the design of intelligent learning environments, and the broader psychological and ethical implications of AI-supported education.
2025
Aversano, Lerina; Iammarino, Martina; Madau, Antonella; Montano, Debora; Verdone, Chiara
A Hybrid Approach Integrating Clinical Data and Tomography to Improve Diagnosis of Parkinson’s Disease Journal Article
In: ACM Transactions on Computing for Healthcare, vol. 6, no. 4, pp. 1–22, 2025, ISSN: 2691-1957, 2637-8051.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Deep Learning, Deep Neural Network, Parkinson Disease
@article{aversano_hybrid_2025,
title = {A Hybrid Approach Integrating Clinical Data and Tomography to Improve Diagnosis of Parkinson’s Disease},
author = {Lerina Aversano and Martina Iammarino and Antonella Madau and Debora Montano and Chiara Verdone},
url = {https://dl.acm.org/doi/10.1145/3735660},
doi = {10.1145/3735660},
issn = {2691-1957, 2637-8051},
year = {2025},
date = {2025-10-01},
urldate = {2025-11-04},
journal = {ACM Transactions on Computing for Healthcare},
volume = {6},
number = {4},
pages = {1–22},
abstract = {Parkinson’s Disease (PD) is a neurodegenerative condition primarily affecting the elderly but also occurring in younger individuals. It is caused by a progressive loss of nerve cells in the brain’s substantia nigra that release dopamine, essential for controlling movements. Dopamine deficiency results in symptoms affecting both motor and non-motor functions, which vary among individuals. Diagnosis relies on clinical symptoms and medical history, often supported by brain scans, as there is no specific diagnostic test available. Diagnosis is challenging due to vague initial symptoms resembling other conditions. Current research indicates that AI can significantly enhance data and image analysis, aiding in the diagnosis and monitoring of PD progression. To this aim, this study proposes a hybrid model allowing the integrated use of clinical data and single photon emission computed tomography images of a patient to predict the presence of the disease. The approach consists of a combination of two types of neural networks, an LSTM for clinical data and a CNN for images. The validation is performed on a widely validated dataset belonging to the Parkinson’s Progression Markers Initiative, from which the data recording visits of 1,814 patients were extracted. The obtained results are interesting and useful to address further investigations.},
keywords = {Artificial Intelligence, Deep Learning, Deep Neural Network, Parkinson Disease},
pubstate = {published},
tppubtype = {article}
}
Parkinson’s Disease (PD) is a neurodegenerative condition primarily affecting the elderly but also occurring in younger individuals. It is caused by a progressive loss of nerve cells in the brain’s substantia nigra that release dopamine, essential for controlling movements. Dopamine deficiency results in symptoms affecting both motor and non-motor functions, which vary among individuals. Diagnosis relies on clinical symptoms and medical history, often supported by brain scans, as there is no specific diagnostic test available. Diagnosis is challenging due to vague initial symptoms resembling other conditions. Current research indicates that AI can significantly enhance data and image analysis, aiding in the diagnosis and monitoring of PD progression. To this aim, this study proposes a hybrid model allowing the integrated use of clinical data and single photon emission computed tomography images of a patient to predict the presence of the disease. The approach consists of a combination of two types of neural networks, an LSTM for clinical data and a CNN for images. The validation is performed on a widely validated dataset belonging to the Parkinson’s Progression Markers Initiative, from which the data recording visits of 1,814 patients were extracted. The obtained results are interesting and useful to address further investigations.
Krilavičius, Tomas; Paolis, Lucio Tommaso De; Luca, Valerio De; Spjut, Josef
eXtended Reality and Artificial Intelligence in Medicine and Rehabilitation Journal Article
In: Information Systems Frontiers, 2025, ISSN: 13873326.
Abstract | Links | BibTeX | Tags: 3D modeling, Artificial Intelligence, Augmented Reality, Extended reality, Minimally-invasive surgery, Personalized medicine, Pre-operative planning, Surgery
@article{krilavicius_extended_2025,
title = {eXtended Reality and Artificial Intelligence in Medicine and Rehabilitation},
author = {Tomas Krilavičius and Lucio Tommaso De Paolis and Valerio De Luca and Josef Spjut},
url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-85217159420&doi=10.1007%2fs10796-025-10580-8&partnerID=40&md5=3bc0bae0925d3f2f6d1a6e1e659b9aae},
doi = {10.1007/s10796-025-10580-8},
issn = {13873326},
year = {2025},
date = {2025-01-01},
journal = {Information Systems Frontiers},
abstract = {This special issue focuses on the application of eXtended Reality (XR) technologies—comprising Virtual Reality (VR), Augmented Reality (AR), and Mixed Reality (MR)—and Artificial Intelligence (AI) in the fields of medicine and rehabilitation. AR provides support in minimally invasive surgery, where it visualises internal anatomical structures on the patient’s body and provides real-time feedback to improve accuracy, keep the surgeon’s attention and reduce the risk of errors. Furthermore, XR technologies can be used to develop applications for pre-operative planning or for training surgeons through serious games. AI finds applications both in medical image processing, for the recognition of anatomical structures and the reconstruction of 3D models, and in the analysis of biological data for patient monitoring and disease diagnosis. In rehabilitation, XR and AI can enable personalised therapy plans, increase patient engagement through immersive environments and provide real-time feedback to improve recovery outcomes. The papers in this special issue deal with rehabilitation through serious games, AI-enhanced XR applications for healthcare, digital twins and the analysis of bio/neuro-adaptive signals. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2025.},
keywords = {3D modeling, Artificial Intelligence, Augmented Reality, Extended reality, Minimally-invasive surgery, Personalized medicine, Pre-operative planning, Surgery},
pubstate = {published},
tppubtype = {article}
}
This special issue focuses on the application of eXtended Reality (XR) technologies—comprising Virtual Reality (VR), Augmented Reality (AR), and Mixed Reality (MR)—and Artificial Intelligence (AI) in the fields of medicine and rehabilitation. AR provides support in minimally invasive surgery, where it visualises internal anatomical structures on the patient’s body and provides real-time feedback to improve accuracy, keep the surgeon’s attention and reduce the risk of errors. Furthermore, XR technologies can be used to develop applications for pre-operative planning or for training surgeons through serious games. AI finds applications both in medical image processing, for the recognition of anatomical structures and the reconstruction of 3D models, and in the analysis of biological data for patient monitoring and disease diagnosis. In rehabilitation, XR and AI can enable personalised therapy plans, increase patient engagement through immersive environments and provide real-time feedback to improve recovery outcomes. The papers in this special issue deal with rehabilitation through serious games, AI-enhanced XR applications for healthcare, digital twins and the analysis of bio/neuro-adaptive signals. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2025.
Barbareschi, Mario; Barone, Salvatore
Investigating the Resilience Source of Classification Systems for Approximate Computing Techniques Journal Article
In: IEEE Transactions on Emerging Topics in Computing, pp. 12, 2025, ISSN: 2168-6750.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Classification systems, Neural networks, Tree Ensemble
@article{barbareschi_investigating_2025,
title = {Investigating the Resilience Source of Classification Systems for Approximate Computing Techniques},
author = {Mario Barbareschi and Salvatore Barone},
url = {https://ieeexplore.ieee.org/document/10542568},
doi = {10.1109/TETC.2024.3403757},
issn = {2168-6750},
year = {2025},
date = {2025-01-01},
journal = {IEEE Transactions on Emerging Topics in Computing},
pages = {12},
abstract = {During the last decade, classification systems (CSs) received significant research attention, with new learning algorithms achieving high accuracy in various applications. However, their resource-intensive nature, in terms of hardware and computation time, poses new design challenges.
CSs exhibit inherent error resilience, due to redundancy of training sets, and self-healing properties, making them suitable for Approximate Computing (AxC).
AxC enables efficient computation by using reduced precision or approximate values, leading to energy, time, and silicon area savings.
Exploiting AxC involves estimating the introduced error for each approximate variant found during a Design-Space Exploration (DSE). This estimation has to be both rapid and meaningful, considering a substantial number of test samples, which are utterly conflicting demands.
In this paper, we investigate on sources of error resiliency of CSs, and we propose a technique to haste the DSE that reduces the computational time for error estimation by systematically reducing the test set. In particular, we cherry-pick samples that are likely to be more sensitive to approximation and perform accuracy-loss estimation just by exploiting such a sample subset.
In order to demonstrate its efficacy, we integrate our technique into two different approaches for generating approximate CSs, showing an average speed-up up to approx18.},
keywords = {Artificial Intelligence, Classification systems, Neural networks, Tree Ensemble},
pubstate = {published},
tppubtype = {article}
}
During the last decade, classification systems (CSs) received significant research attention, with new learning algorithms achieving high accuracy in various applications. However, their resource-intensive nature, in terms of hardware and computation time, poses new design challenges.
CSs exhibit inherent error resilience, due to redundancy of training sets, and self-healing properties, making them suitable for Approximate Computing (AxC).
AxC enables efficient computation by using reduced precision or approximate values, leading to energy, time, and silicon area savings.
Exploiting AxC involves estimating the introduced error for each approximate variant found during a Design-Space Exploration (DSE). This estimation has to be both rapid and meaningful, considering a substantial number of test samples, which are utterly conflicting demands.
In this paper, we investigate on sources of error resiliency of CSs, and we propose a technique to haste the DSE that reduces the computational time for error estimation by systematically reducing the test set. In particular, we cherry-pick samples that are likely to be more sensitive to approximation and perform accuracy-loss estimation just by exploiting such a sample subset.
In order to demonstrate its efficacy, we integrate our technique into two different approaches for generating approximate CSs, showing an average speed-up up to approx18.
CSs exhibit inherent error resilience, due to redundancy of training sets, and self-healing properties, making them suitable for Approximate Computing (AxC).
AxC enables efficient computation by using reduced precision or approximate values, leading to energy, time, and silicon area savings.
Exploiting AxC involves estimating the introduced error for each approximate variant found during a Design-Space Exploration (DSE). This estimation has to be both rapid and meaningful, considering a substantial number of test samples, which are utterly conflicting demands.
In this paper, we investigate on sources of error resiliency of CSs, and we propose a technique to haste the DSE that reduces the computational time for error estimation by systematically reducing the test set. In particular, we cherry-pick samples that are likely to be more sensitive to approximation and perform accuracy-loss estimation just by exploiting such a sample subset.
In order to demonstrate its efficacy, we integrate our technique into two different approaches for generating approximate CSs, showing an average speed-up up to approx18.
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).
Abstract | Links | BibTeX | Tags: Healthcare, Machine Learning, Process Mining
@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},
abstract = {Process mining and machine learning models are playing a vital role in the medical field today. These models are enabling the development of new technologies that improve the study of hospital processes. The focus of this work is on optimally reconstructing a process that has defects such as missing activity labels. In order to achieve this, a prediction is made on the activities present in hospital processes, exploring the previous and subsequent activities in an ordered trace. The experiments were conducted on real data that refer to hospital event logs.},
note = {Series Title: Smart Innovation, Systems and Technologies},
keywords = {Healthcare, Machine Learning, Process Mining},
pubstate = {published},
tppubtype = {incollection}
}
Process mining and machine learning models are playing a vital role in the medical field today. These models are enabling the development of new technologies that improve the study of hospital processes. The focus of this work is on optimally reconstructing a process that has defects such as missing activity labels. In order to achieve this, a prediction is made on the activities present in hospital processes, exploring the previous and subsequent activities in an ordered trace. The experiments were conducted on real data that refer to hospital event logs.
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.
Abstract | Links | BibTeX | Tags: Explainability, Machine Learning, Open Data
@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},
abstract = {Artificial intelligence and machine learning models are emerging as essential tools for optimizing municipal solid waste management and supporting policy decisions. However, transparency and interpretability of these models’ predictions continue to be major obstacles. Recent advances in Explainable Artificial Intelligence (XAI) techniques have made it possible to explain specific model decisions and guarantee that the outcomes are intelligible and useful. Using high-quality Italian open data in the form of Linked Open Data (LOD), this study investigates the benefits and viability of creating explainable models in italian municipalities. To achieve this, a method for using connected and open statistical data to create explainable models is provided. Addi- tionally, a case study is presented, covering four years, in which waste management expenses are predicted and interpreted using connected data about Italian municipalities, categorizing them into three cost bands. CatBoost was selected as the predictive model’s algorithm, and the SHAP framework was used to guarantee the predictions’ transparency. Through transparent and accountable data management, this effort seeks to il- lustrate how cutting-edge technologies can enhance the sustainability of public programs.},
keywords = {Explainability, Machine Learning, Open Data},
pubstate = {published},
tppubtype = {inproceedings}
}
Artificial intelligence and machine learning models are emerging as essential tools for optimizing municipal solid waste management and supporting policy decisions. However, transparency and interpretability of these models’ predictions continue to be major obstacles. Recent advances in Explainable Artificial Intelligence (XAI) techniques have made it possible to explain specific model decisions and guarantee that the outcomes are intelligible and useful. Using high-quality Italian open data in the form of Linked Open Data (LOD), this study investigates the benefits and viability of creating explainable models in italian municipalities. To achieve this, a method for using connected and open statistical data to create explainable models is provided. Addi- tionally, a case study is presented, covering four years, in which waste management expenses are predicted and interpreted using connected data about Italian municipalities, categorizing them into three cost bands. CatBoost was selected as the predictive model’s algorithm, and the SHAP framework was used to guarantee the predictions’ transparency. Through transparent and accountable data management, this effort seeks to il- lustrate how cutting-edge technologies can enhance the sustainability of public programs.
2024
Agostinelli, Thomas; Generosi, Andrea; Ceccacci, Silvia; Mengoni, Maura
Validation of computer vision-based ergonomic risk assessment tools for real manufacturing environments Journal Article
In: Scientific Reports, vol. 14, no. 1, pp. 27785, 2024, ISSN: 2045-2322.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Computer Vision and Pattern Recognition, Human-Centered Design, Industry 4.0
@article{agostinelli_validation_2024,
title = {Validation of computer vision-based ergonomic risk assessment tools for real manufacturing environments},
author = {Thomas Agostinelli and Andrea Generosi and Silvia Ceccacci and Maura Mengoni},
url = {https://www.nature.com/articles/s41598-024-79373-4},
doi = {10.1038/s41598-024-79373-4},
issn = {2045-2322},
year = {2024},
date = {2024-11-01},
urldate = {2024-12-28},
journal = {Scientific Reports},
volume = {14},
number = {1},
pages = {27785},
abstract = {This study contributes to understanding semi-automated ergonomic risk assessments in industrial manufacturing environments, proposing a practical tool for enhancing worker safety and operational efficiency. In the Industry 5.0 era, the human-centric approach in manufacturing is crucial, especially considering the aging workforce and the dynamic nature of the entire modern industrial sector, today integrating digital technology, automation, and sustainable practices to enhance productivity and environmental responsibility. This approach aims to adapt work conditions to individual capabilities, addressing the high incidence of work-related musculoskeletal disorders (MSDs). The traditional, subjective methods of ergonomic assessment are inadequate for dynamic settings, highlighting the need for affordable, automatic tools for continuous monitoring of workers’ postures to evaluate ergonomic risks effectively during tasks. To enable this perspective, 2D RGB Motion Capture (MoCap) systems based on computer vision currently seem the technologies of choice, given their low intrusiveness, cost, and implementation effort. However, the reliability and applicability of these systems in the dynamic and varied manufacturing environment remain uncertain. This research benchmarks various literature proposed MoCap tools and examines the viability of MoCap systems for ergonomic risk assessments in Industry 5.0 by exploiting one of the benchmarked semi-automated, low-cost and non-intrusive 2D RGB MoCap system, capable of continuously monitoring and analysing workers’ postures. By conducting experiments across varied manufacturing environments, this research evaluates the system’s effectiveness in assessing ergonomic risks and its adaptability to different production lines. Results reveal that the accuracy of risk assessments varies by specific environmental conditions and workstation setups. Although these systems are not yet optimized for expert-level risk certification, they offer significant potential for enhancing workplace safety and efficiency by providing continuous posture monitoring. Future improvements could explore advanced computational techniques like machine learning to refine ergonomic assessments further.},
keywords = {Artificial Intelligence, Computer Vision and Pattern Recognition, Human-Centered Design, Industry 4.0},
pubstate = {published},
tppubtype = {article}
}
This study contributes to understanding semi-automated ergonomic risk assessments in industrial manufacturing environments, proposing a practical tool for enhancing worker safety and operational efficiency. In the Industry 5.0 era, the human-centric approach in manufacturing is crucial, especially considering the aging workforce and the dynamic nature of the entire modern industrial sector, today integrating digital technology, automation, and sustainable practices to enhance productivity and environmental responsibility. This approach aims to adapt work conditions to individual capabilities, addressing the high incidence of work-related musculoskeletal disorders (MSDs). The traditional, subjective methods of ergonomic assessment are inadequate for dynamic settings, highlighting the need for affordable, automatic tools for continuous monitoring of workers’ postures to evaluate ergonomic risks effectively during tasks. To enable this perspective, 2D RGB Motion Capture (MoCap) systems based on computer vision currently seem the technologies of choice, given their low intrusiveness, cost, and implementation effort. However, the reliability and applicability of these systems in the dynamic and varied manufacturing environment remain uncertain. This research benchmarks various literature proposed MoCap tools and examines the viability of MoCap systems for ergonomic risk assessments in Industry 5.0 by exploiting one of the benchmarked semi-automated, low-cost and non-intrusive 2D RGB MoCap system, capable of continuously monitoring and analysing workers’ postures. By conducting experiments across varied manufacturing environments, this research evaluates the system’s effectiveness in assessing ergonomic risks and its adaptability to different production lines. Results reveal that the accuracy of risk assessments varies by specific environmental conditions and workstation setups. Although these systems are not yet optimized for expert-level risk certification, they offer significant potential for enhancing workplace safety and efficiency by providing continuous posture monitoring. Future improvements could explore advanced computational techniques like machine learning to refine ergonomic assessments further.
Recupito, Gilberto; Pecorelli, Fabiano; Catolino, Gemma; Lenarduzzi, Valentina; Taibi, Davide; Nucci, Dario Di; Palomba, Fabio
Technical debt in AI-enabled systems: On the prevalence, severity, impact, and management strategies for code and architecture Journal Article
In: Journal of Systems and Software, vol. 216, pp. 112151, 2024, ISSN: 0164-1212.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Software Engineering, Technical Debt Management
@article{recupitoTechnicalDebtAIenabled2024,
title = {Technical debt in AI-enabled systems: On the prevalence, severity, impact, and management strategies for code and architecture},
author = {Gilberto Recupito and Fabiano Pecorelli and Gemma Catolino and Valentina Lenarduzzi and Davide Taibi and Dario Di Nucci and Fabio Palomba},
url = {https://linkinghub.elsevier.com/retrieve/pii/S0164121224001961},
doi = {10.1016/j.jss.2024.112151},
issn = {0164-1212},
year = {2024},
date = {2024-10-01},
urldate = {2024-07-07},
journal = {Journal of Systems and Software},
volume = {216},
pages = {112151},
abstract = {Context: Artificial Intelligence (AI) is pervasive in several application domains and promises to be even more diffused in the next decades. Developing high-quality AI-enabled systems — software systems embedding one or multiple AI components, algorithms, and models — could introduce critical challenges for mitigating specific risks related to the systems' quality. Such development alone is insufficient to fully address socio-technical consequences and the need for rapid adaptation to evolutionary changes. Recent work proposed the concept of AI technical debt, a potential liability concerned with developing AI-enabled systems whose impact can affect the overall systems' quality. While the problem of AI technical debt is rapidly gaining the attention of the software engineering research community, scientific knowledge that contributes to understanding and managing the matter is still limited. Objective: In this paper, we leverage the expertise of practitioners to offer useful insights to the research community, aiming to enhance researchers' awareness about the detection and mitigation of AI technical debt. Our ultimate goal is to empower practitioners by providing them with tools and methods. Additionally, our study sheds light on novel aspects that practitioners might not be fully acquainted with, contributing to a deeper understanding of the subject. Method: We develop a survey study featuring 53 AI practitioners, in which we collect information on the practical prevalence, severity, and impact of AI technical debt issues affecting the code and the architecture other than the strategies applied by practitioners to identify and mitigate them. Results: The key findings of the study reveal the multiple impacts that AI technical debt issues may have on the quality of AI-enabled systems (e.g., the high negative impact that Undeclared consumers has on security, whereas Jumbled Model Architecture can induce the code to be hard to maintain) and the little support practitioners have to deal with them, limited to apply manual effort for identification and refactoring. Conclusion: We conclude the article by distilling lessons learned and actionable insights for researchers.},
keywords = {Artificial Intelligence, Software Engineering, Technical Debt Management},
pubstate = {published},
tppubtype = {article}
}
Context: Artificial Intelligence (AI) is pervasive in several application domains and promises to be even more diffused in the next decades. Developing high-quality AI-enabled systems — software systems embedding one or multiple AI components, algorithms, and models — could introduce critical challenges for mitigating specific risks related to the systems' quality. Such development alone is insufficient to fully address socio-technical consequences and the need for rapid adaptation to evolutionary changes. Recent work proposed the concept of AI technical debt, a potential liability concerned with developing AI-enabled systems whose impact can affect the overall systems' quality. While the problem of AI technical debt is rapidly gaining the attention of the software engineering research community, scientific knowledge that contributes to understanding and managing the matter is still limited. Objective: In this paper, we leverage the expertise of practitioners to offer useful insights to the research community, aiming to enhance researchers' awareness about the detection and mitigation of AI technical debt. Our ultimate goal is to empower practitioners by providing them with tools and methods. Additionally, our study sheds light on novel aspects that practitioners might not be fully acquainted with, contributing to a deeper understanding of the subject. Method: We develop a survey study featuring 53 AI practitioners, in which we collect information on the practical prevalence, severity, and impact of AI technical debt issues affecting the code and the architecture other than the strategies applied by practitioners to identify and mitigate them. Results: The key findings of the study reveal the multiple impacts that AI technical debt issues may have on the quality of AI-enabled systems (e.g., the high negative impact that Undeclared consumers has on security, whereas Jumbled Model Architecture can induce the code to be hard to maintain) and the little support practitioners have to deal with them, limited to apply manual effort for identification and refactoring. Conclusion: We conclude the article by distilling lessons learned and actionable insights for researchers.
Rausa, Maria; Gaglio, Salvatore; Augello, Agnese; Caggianese, Giuseppe; Franchini, Silvia; Gallo, Luigi; Sabatucci, Luca
Enriching Metaverse with Memories Through Generative AI: A Case Study Proceedings Article
In: 2024 IEEE International Conference on Metrology for eXtended Reality, Artificial Intelligence and Neural Engineering (MetroXRAINE), pp. 371–376, 2024.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Metaverse, Modeling, Virtual Reality
@inproceedings{rausa_enriching_2024,
title = {Enriching Metaverse with Memories Through Generative AI: A Case Study},
author = {Maria Rausa and Salvatore Gaglio and Agnese Augello and Giuseppe Caggianese and Silvia Franchini and Luigi Gallo and Luca Sabatucci},
url = {https://ieeexplore.ieee.org/abstract/document/10796338},
doi = {10.1109/MetroXRAINE62247.2024.10796338},
year = {2024},
date = {2024-10-01},
urldate = {2025-01-08},
booktitle = {2024 IEEE International Conference on Metrology for eXtended Reality, Artificial Intelligence and Neural Engineering (MetroXRAINE)},
pages = {371–376},
abstract = {The paper introduces MetaMemory, an approach to generate 3D models from either textual descriptions or photographs of objects, offering dual input modes for enhanced representation. MetaMemory's architecture is discussed presenting the tools employed in extracting the object from the image, generating the 3D mesh from texts or images, and visualizing the object reconstruction in an immersive scenario. Afterwards, a case study in which we experienced reconstructing memories of ancient crafts is examined together with the achieved results, by highlighting current limitations and potential applications.},
keywords = {Artificial Intelligence, Metaverse, Modeling, Virtual Reality},
pubstate = {published},
tppubtype = {inproceedings}
}
The paper introduces MetaMemory, an approach to generate 3D models from either textual descriptions or photographs of objects, offering dual input modes for enhanced representation. MetaMemory's architecture is discussed presenting the tools employed in extracting the object from the image, generating the 3D mesh from texts or images, and visualizing the object reconstruction in an immersive scenario. Afterwards, a case study in which we experienced reconstructing memories of ancient crafts is examined together with the achieved results, by highlighting current limitations and potential applications.
Aversano, Lerina; Iammarino, Martina; Mancino, Ilaria; Montano, Debora
A systematic review on artificial intelligence approaches for smart health devices Journal Article
In: PeerJ Computer Science, vol. 10, pp. e2232, 2024, ISSN: 2376-5992.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Healthcare, Healthcare services
@article{aversano_systematic_2024,
title = {A systematic review on artificial intelligence approaches for smart health devices},
author = {Lerina Aversano and Martina Iammarino and Ilaria Mancino and Debora Montano},
url = {https://peerj.com/articles/cs-2232},
doi = {10.7717/peerj-cs.2232},
issn = {2376-5992},
year = {2024},
date = {2024-10-01},
urldate = {2025-10-22},
journal = {PeerJ Computer Science},
volume = {10},
pages = {e2232},
abstract = {In the context of smart health, the use of wearable Internet of Things (IoT) devices is becoming increasingly popular to monitor and manage patients’ health conditions in a more efficient and personalized way. However, choosing the most suitable artificial intelligence (AI) methodology to analyze the data collected by these devices is crucial to ensure the reliability and effectiveness of smart healthcare applications. Additionally, protecting the privacy and security of health data is an area of growing concern, given the sensitivity and personal nature of such information. In this context, machine learning (ML) and deep learning (DL) are emerging as successful technologies because they are suitable for application to advanced analysis and prediction of healthcare scenarios. Therefore, the objective of this work is to contribute to the current state of the literature by identifying challenges, best practices, and future opportunities in the field of smart health. We aim to provide a comprehensive overview of the AI methodologies used, the neural network architectures adopted, and the algorithms employed, as well as examine the privacy and security issues related to the management of health data collected by wearable IoT devices. Through this systematic review, we aim to offer practical guidelines for the design, development, and implementation of AI solutions in smart health, to improve the quality of care provided and promote patient well-being. To pursue our goal, several articles focusing on ML or DL network architectures were selected and reviewed. The final discussion highlights research gaps yet to be investigated, as well as the drawbacks and vulnerabilities of existing IoT applications in smart healthcare.},
keywords = {Artificial Intelligence, Healthcare, Healthcare services},
pubstate = {published},
tppubtype = {article}
}
In the context of smart health, the use of wearable Internet of Things (IoT) devices is becoming increasingly popular to monitor and manage patients’ health conditions in a more efficient and personalized way. However, choosing the most suitable artificial intelligence (AI) methodology to analyze the data collected by these devices is crucial to ensure the reliability and effectiveness of smart healthcare applications. Additionally, protecting the privacy and security of health data is an area of growing concern, given the sensitivity and personal nature of such information. In this context, machine learning (ML) and deep learning (DL) are emerging as successful technologies because they are suitable for application to advanced analysis and prediction of healthcare scenarios. Therefore, the objective of this work is to contribute to the current state of the literature by identifying challenges, best practices, and future opportunities in the field of smart health. We aim to provide a comprehensive overview of the AI methodologies used, the neural network architectures adopted, and the algorithms employed, as well as examine the privacy and security issues related to the management of health data collected by wearable IoT devices. Through this systematic review, we aim to offer practical guidelines for the design, development, and implementation of AI solutions in smart health, to improve the quality of care provided and promote patient well-being. To pursue our goal, several articles focusing on ML or DL network architectures were selected and reviewed. The final discussion highlights research gaps yet to be investigated, as well as the drawbacks and vulnerabilities of existing IoT applications in smart healthcare.
Aversano, Lerina; Bernardi, Mario Luca; Calgano, Vincenzo; Cimitile, Marta; Esposito, Concetta; Iammarino, Martina; Pisco, Marco; Spaziani, Sara; Verdone, Chiara
Using Machine Learning for Classification of Cancer Cells from Raman Spectroscopy Proceedings Article
In: pp. 15–24, 2024, ISBN: 978-989-758-584-5.
Abstract | Links | BibTeX | Tags: Classification, Healthcare, Machine Learning
@inproceedings{aversanoUsingMachineLearning2024,
title = {Using Machine Learning for Classification of Cancer Cells from Raman Spectroscopy},
author = {Lerina Aversano and Mario Luca Bernardi and Vincenzo Calgano and Marta Cimitile and Concetta Esposito and Martina Iammarino and Marco Pisco and Sara Spaziani and Chiara Verdone},
url = {https://www.scitepress.org/Link.aspx?doi=10.5220/0011142600003277},
doi = {10.5220/0011142600003277},
isbn = {978-989-758-584-5},
year = {2024},
date = {2024-10-01},
urldate = {2024-10-02},
pages = {15–24},
abstract = {Since cancer represents one of the leading causes of death worldwide, the development of approaches capable of discerning healthy from diseased cells would be of fundamental importance to support diagnostic and screening techniques. Raman spectroscopy is the most effective molecular analysis technique currently available and provides information on the molecular composition, bonds, chemical environment, phase, and crystalline structure of the samples under examination. This work exploits a combination of Raman spectroscopy and machine learning models to discriminate patients’ liver cells between tumor and non-tumor. The research uses real patient data, provided by the Center for Nanophotonics and Optoelectronics for Human Health (CNOS), which analyzed the cells of a patient with liver cancer. Specifically, the dataset has been built through a long data collection process, which first involved the analysis of the cells with Raman spectroscopy and then the training of two classifiers, Decision Tree and Random Forest. The results show good performance for the trained classifiers, especially those relating to the Random Forest, which reaches an accuracy of 90%.},
keywords = {Classification, Healthcare, Machine Learning},
pubstate = {published},
tppubtype = {inproceedings}
}
Since cancer represents one of the leading causes of death worldwide, the development of approaches capable of discerning healthy from diseased cells would be of fundamental importance to support diagnostic and screening techniques. Raman spectroscopy is the most effective molecular analysis technique currently available and provides information on the molecular composition, bonds, chemical environment, phase, and crystalline structure of the samples under examination. This work exploits a combination of Raman spectroscopy and machine learning models to discriminate patients’ liver cells between tumor and non-tumor. The research uses real patient data, provided by the Center for Nanophotonics and Optoelectronics for Human Health (CNOS), which analyzed the cells of a patient with liver cancer. Specifically, the dataset has been built through a long data collection process, which first involved the analysis of the cells with Raman spectroscopy and then the training of two classifiers, Decision Tree and Random Forest. The results show good performance for the trained classifiers, especially those relating to the Random Forest, which reaches an accuracy of 90%.
Aversano, Lerina; Bernardi, Mario Luca; Calgano, Vincenzo; Cimitile, Marta; Esposito, Concetta; Iammarino, Martina; Pisco, Marco; Spaziani, Sara; Verdone, Chiara
Using Machine Learning for Classification of Cancer Cells from Raman Spectroscopy Proceedings Article
In: pp. 15–24, 2024, ISBN: 978-989-758-584-5.
Abstract | Links | BibTeX | Tags: Classification, Healthcare, Machine Learning
@inproceedings{aversano_using_2024,
title = {Using Machine Learning for Classification of Cancer Cells from Raman Spectroscopy},
author = {Lerina Aversano and Mario Luca Bernardi and Vincenzo Calgano and Marta Cimitile and Concetta Esposito and Martina Iammarino and Marco Pisco and Sara Spaziani and Chiara Verdone},
url = {https://www.scitepress.org/Link.aspx?doi=10.5220/0011142600003277},
doi = {10.5220/0011142600003277},
isbn = {978-989-758-584-5},
year = {2024},
date = {2024-10-01},
urldate = {2024-10-02},
pages = {15–24},
abstract = {Since cancer represents one of the leading causes of death worldwide, the development of approaches capable of discerning healthy from diseased cells would be of fundamental importance to support diagnostic and screening techniques. Raman spectroscopy is the most effective molecular analysis technique currently available and provides information on the molecular composition, bonds, chemical environment, phase, and crystalline structure of the samples under examination. This work exploits a combination of Raman spectroscopy and machine learning models to discriminate patients’ liver cells between tumor and non-tumor. The research uses real patient data, provided by the Center for Nanophotonics and Optoelectronics for Human Health (CNOS), which analyzed the cells of a patient with liver cancer. Specifically, the dataset has been built through a long data collection process, which first involved the analysis of the cells with Raman spectroscopy and then the training of two classifiers, Decision Tree and Random Forest. The results show good performance for the trained classifiers, especially those relating to the Random Forest, which reaches an accuracy of 90%.},
keywords = {Classification, Healthcare, Machine Learning},
pubstate = {published},
tppubtype = {inproceedings}
}
Since cancer represents one of the leading causes of death worldwide, the development of approaches capable of discerning healthy from diseased cells would be of fundamental importance to support diagnostic and screening techniques. Raman spectroscopy is the most effective molecular analysis technique currently available and provides information on the molecular composition, bonds, chemical environment, phase, and crystalline structure of the samples under examination. This work exploits a combination of Raman spectroscopy and machine learning models to discriminate patients’ liver cells between tumor and non-tumor. The research uses real patient data, provided by the Center for Nanophotonics and Optoelectronics for Human Health (CNOS), which analyzed the cells of a patient with liver cancer. Specifically, the dataset has been built through a long data collection process, which first involved the analysis of the cells with Raman spectroscopy and then the training of two classifiers, Decision Tree and Random Forest. The results show good performance for the trained classifiers, especially those relating to the Random Forest, which reaches an accuracy of 90%.
Aversano, Lerina; Bernardi, Mario Luca; Calgano, Vincenzo; Cimitile, Marta; Esposito, Concetta; Iammarino, Martina; Pisco, Marco; Spaziani, Sara; Verdone, Chiara
Using Machine Learning for Classification of Cancer Cells from Raman Spectroscopy Proceedings Article
In: pp. 15–24, 2024, ISBN: 978-989-758-584-5.
Abstract | Links | BibTeX | Tags: Classification, Healthcare, Machine Learning
@inproceedings{aversano_using_2024-1,
title = {Using Machine Learning for Classification of Cancer Cells from Raman Spectroscopy},
author = {Lerina Aversano and Mario Luca Bernardi and Vincenzo Calgano and Marta Cimitile and Concetta Esposito and Martina Iammarino and Marco Pisco and Sara Spaziani and Chiara Verdone},
url = {https://www.scitepress.org/Link.aspx?doi=10.5220/0011142600003277},
doi = {10.5220/0011142600003277},
isbn = {978-989-758-584-5},
year = {2024},
date = {2024-10-01},
urldate = {2024-10-02},
pages = {15–24},
abstract = {Since cancer represents one of the leading causes of death worldwide, the development of approaches capable of discerning healthy from diseased cells would be of fundamental importance to support diagnostic and screening techniques. Raman spectroscopy is the most effective molecular analysis technique currently available and provides information on the molecular composition, bonds, chemical environment, phase, and crystalline structure of the samples under examination. This work exploits a combination of Raman spectroscopy and machine learning models to discriminate patients’ liver cells between tumor and non-tumor. The research uses real patient data, provided by the Center for Nanophotonics and Optoelectronics for Human Health (CNOS), which analyzed the cells of a patient with liver cancer. Specifically, the dataset has been built through a long data collection process, which first involved the analysis of the cells with Raman spectroscopy and then the training of two classifiers, Decision Tree and Random Forest. The results show good performance for the trained classifiers, especially those relating to the Random Forest, which reaches an accuracy of 90%.},
keywords = {Classification, Healthcare, Machine Learning},
pubstate = {published},
tppubtype = {inproceedings}
}
Since cancer represents one of the leading causes of death worldwide, the development of approaches capable of discerning healthy from diseased cells would be of fundamental importance to support diagnostic and screening techniques. Raman spectroscopy is the most effective molecular analysis technique currently available and provides information on the molecular composition, bonds, chemical environment, phase, and crystalline structure of the samples under examination. This work exploits a combination of Raman spectroscopy and machine learning models to discriminate patients’ liver cells between tumor and non-tumor. The research uses real patient data, provided by the Center for Nanophotonics and Optoelectronics for Human Health (CNOS), which analyzed the cells of a patient with liver cancer. Specifically, the dataset has been built through a long data collection process, which first involved the analysis of the cells with Raman spectroscopy and then the training of two classifiers, Decision Tree and Random Forest. The results show good performance for the trained classifiers, especially those relating to the Random Forest, which reaches an accuracy of 90%.
Mennella, Ciro; Esposito, Massimo; Pietro, Giuseppe De; Maniscalco, Umberto
Promoting fairness in activity recognition algorithms for patient’s monitoring and evaluation systems in healthcare Journal Article
In: Computers in Biology and Medicine, vol. 179, pp. 108826, 2024, ISSN: 00104825.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Bias, Deep Learning, Motion analysis, Rehabilitation, Time-series
@article{mennellaPromotingFairnessActivity2024,
title = {Promoting fairness in activity recognition algorithms for patient’s monitoring and evaluation systems in healthcare},
author = {Ciro Mennella and Massimo Esposito and Giuseppe De Pietro and Umberto Maniscalco},
url = {https://linkinghub.elsevier.com/retrieve/pii/S0010482524009119},
doi = {10.1016/j.compbiomed.2024.108826},
issn = {00104825},
year = {2024},
date = {2024-09-01},
urldate = {2024-07-21},
journal = {Computers in Biology and Medicine},
volume = {179},
pages = {108826},
abstract = {Researchers face the challenge of defining subject selection criteria when training algorithms for human activity recognition tasks. The ongoing uncertainty revolves around which characteristics should be considered to ensure algorithmic robustness across diverse populations. This study aims to address this challenge by conducting an analysis of heterogeneity in the training data to assess the impact of physical characteristics and soft-biometric attributes on activity recognition performance.
The performance of various state-of-the-art deep neural network architectures (tCNN, hybrid-LSTM, Transformer model) processing time-series data using the IntelliRehab (IRDS) dataset was evaluated. By intentionally introducing bias into the training data based on human characteristics, the objective is to identify the characteristics that influence algorithms in motion analysis.
Experimental findings reveal that the CNN-LSTM model achieved the highest accuracy, reaching 88%. Moreover, models trained on heterogeneous distributions of disability attributes exhibited notably higher accuracy, reaching 51%, compared to those not considering such factors, which scored an average of 33%. These evaluations underscore the significant influence of subjects’ characteristics on activity recognition performance, providing valuable insights into the algorithm’s robustness across diverse populations.
This study represents a significant step forward in promoting fairness and trustworthiness in artificial intelligence by quantifying representation bias in multi-channel time-series activity recognition data within the healthcare domain.},
keywords = {Artificial Intelligence, Bias, Deep Learning, Motion analysis, Rehabilitation, Time-series},
pubstate = {published},
tppubtype = {article}
}
Researchers face the challenge of defining subject selection criteria when training algorithms for human activity recognition tasks. The ongoing uncertainty revolves around which characteristics should be considered to ensure algorithmic robustness across diverse populations. This study aims to address this challenge by conducting an analysis of heterogeneity in the training data to assess the impact of physical characteristics and soft-biometric attributes on activity recognition performance.
The performance of various state-of-the-art deep neural network architectures (tCNN, hybrid-LSTM, Transformer model) processing time-series data using the IntelliRehab (IRDS) dataset was evaluated. By intentionally introducing bias into the training data based on human characteristics, the objective is to identify the characteristics that influence algorithms in motion analysis.
Experimental findings reveal that the CNN-LSTM model achieved the highest accuracy, reaching 88%. Moreover, models trained on heterogeneous distributions of disability attributes exhibited notably higher accuracy, reaching 51%, compared to those not considering such factors, which scored an average of 33%. These evaluations underscore the significant influence of subjects’ characteristics on activity recognition performance, providing valuable insights into the algorithm’s robustness across diverse populations.
This study represents a significant step forward in promoting fairness and trustworthiness in artificial intelligence by quantifying representation bias in multi-channel time-series activity recognition data within the healthcare domain.
The performance of various state-of-the-art deep neural network architectures (tCNN, hybrid-LSTM, Transformer model) processing time-series data using the IntelliRehab (IRDS) dataset was evaluated. By intentionally introducing bias into the training data based on human characteristics, the objective is to identify the characteristics that influence algorithms in motion analysis.
Experimental findings reveal that the CNN-LSTM model achieved the highest accuracy, reaching 88%. Moreover, models trained on heterogeneous distributions of disability attributes exhibited notably higher accuracy, reaching 51%, compared to those not considering such factors, which scored an average of 33%. These evaluations underscore the significant influence of subjects’ characteristics on activity recognition performance, providing valuable insights into the algorithm’s robustness across diverse populations.
This study represents a significant step forward in promoting fairness and trustworthiness in artificial intelligence by quantifying representation bias in multi-channel time-series activity recognition data within the healthcare domain.
Vergallo, Roberto; Mainetti, Luca
Measuring the Effectiveness of Carbon-Aware AI Training Strategies in Cloud Instances: A Confirmation Study Journal Article
In: Future Internet, vol. 16, no. 9, pp. 334, 2024, ISSN: 1999-5903.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Carbon Awareness, Fintech, Sustainability
@article{vergalloMeasuringEffectivenessCarbonAware2024,
title = {Measuring the Effectiveness of Carbon-Aware AI Training Strategies in Cloud Instances: A Confirmation Study},
author = {Roberto Vergallo and Luca Mainetti},
url = {https://www.mdpi.com/1999-5903/16/9/334},
doi = {10.3390/fi16090334},
issn = {1999-5903},
year = {2024},
date = {2024-09-01},
urldate = {2024-10-02},
journal = {Future Internet},
volume = {16},
number = {9},
pages = {334},
abstract = {While the massive adoption of Artificial Intelligence (AI) is threatening the environment, new research efforts begin to be employed to measure and mitigate the carbon footprint of both training and inference phases. In this domain, two carbon-aware training strategies have been proposed in the literature: Flexible Start and Pause & Resume. Such strategies—natively Cloud-based—use the time resource to postpone or pause the training algorithm when the carbon intensity reaches a threshold. While such strategies have proved to achieve interesting results on a benchmark of modern models covering Natural Language Processing (NLP) and computer vision applications and a wide range of model sizes (up to 6.1B parameters), it is still unclear whether such results may hold also with different algorithms and in different geographical regions. In this confirmation study, we use the same methodology as the state-of-the-art strategies to recompute the saving in carbon emissions of Flexible Start and Pause & Resume in the Anomaly Detection (AD) domain. Results confirm their effectiveness in two specific conditions, but the percentage reduction behaves differently compared with what is stated in the existing literature.},
keywords = {Artificial Intelligence, Carbon Awareness, Fintech, Sustainability},
pubstate = {published},
tppubtype = {article}
}
While the massive adoption of Artificial Intelligence (AI) is threatening the environment, new research efforts begin to be employed to measure and mitigate the carbon footprint of both training and inference phases. In this domain, two carbon-aware training strategies have been proposed in the literature: Flexible Start and Pause & Resume. Such strategies—natively Cloud-based—use the time resource to postpone or pause the training algorithm when the carbon intensity reaches a threshold. While such strategies have proved to achieve interesting results on a benchmark of modern models covering Natural Language Processing (NLP) and computer vision applications and a wide range of model sizes (up to 6.1B parameters), it is still unclear whether such results may hold also with different algorithms and in different geographical regions. In this confirmation study, we use the same methodology as the state-of-the-art strategies to recompute the saving in carbon emissions of Flexible Start and Pause & Resume in the Anomaly Detection (AD) domain. Results confirm their effectiveness in two specific conditions, but the percentage reduction behaves differently compared with what is stated in the existing literature.
Vergallo, Roberto; Mainetti, Luca
Measuring the Effectiveness of Carbon-Aware AI Training Strategies in Cloud Instances: A Confirmation Study Journal Article
In: Future Internet, vol. 16, no. 9, pp. 334, 2024, ISSN: 1999-5903.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Carbon Awareness, Fintech, Sustainability
@article{vergallo_measuring_2024,
title = {Measuring the Effectiveness of Carbon-Aware AI Training Strategies in Cloud Instances: A Confirmation Study},
author = {Roberto Vergallo and Luca Mainetti},
url = {https://www.mdpi.com/1999-5903/16/9/334},
doi = {10.3390/fi16090334},
issn = {1999-5903},
year = {2024},
date = {2024-09-01},
urldate = {2024-10-02},
journal = {Future Internet},
volume = {16},
number = {9},
pages = {334},
abstract = {While the massive adoption of Artificial Intelligence (AI) is threatening the environment, new research efforts begin to be employed to measure and mitigate the carbon footprint of both training and inference phases. In this domain, two carbon-aware training strategies have been proposed in the literature: Flexible Start and Pause & Resume. Such strategies—natively Cloud-based—use the time resource to postpone or pause the training algorithm when the carbon intensity reaches a threshold. While such strategies have proved to achieve interesting results on a benchmark of modern models covering Natural Language Processing (NLP) and computer vision applications and a wide range of model sizes (up to 6.1B parameters), it is still unclear whether such results may hold also with different algorithms and in different geographical regions. In this confirmation study, we use the same methodology as the state-of-the-art strategies to recompute the saving in carbon emissions of Flexible Start and Pause & Resume in the Anomaly Detection (AD) domain. Results confirm their effectiveness in two specific conditions, but the percentage reduction behaves differently compared with what is stated in the existing literature.},
keywords = {Artificial Intelligence, Carbon Awareness, Fintech, Sustainability},
pubstate = {published},
tppubtype = {article}
}
While the massive adoption of Artificial Intelligence (AI) is threatening the environment, new research efforts begin to be employed to measure and mitigate the carbon footprint of both training and inference phases. In this domain, two carbon-aware training strategies have been proposed in the literature: Flexible Start and Pause & Resume. Such strategies—natively Cloud-based—use the time resource to postpone or pause the training algorithm when the carbon intensity reaches a threshold. While such strategies have proved to achieve interesting results on a benchmark of modern models covering Natural Language Processing (NLP) and computer vision applications and a wide range of model sizes (up to 6.1B parameters), it is still unclear whether such results may hold also with different algorithms and in different geographical regions. In this confirmation study, we use the same methodology as the state-of-the-art strategies to recompute the saving in carbon emissions of Flexible Start and Pause & Resume in the Anomaly Detection (AD) domain. Results confirm their effectiveness in two specific conditions, but the percentage reduction behaves differently compared with what is stated in the existing literature.
Mennella, Ciro; Esposito, Massimo; Pietro, Giuseppe De; Maniscalco, Umberto
Promoting fairness in activity recognition algorithms for patient’s monitoring and evaluation systems in healthcare Journal Article
In: Computers in Biology and Medicine, vol. 179, pp. 108826, 2024, ISSN: 00104825.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Bias, Deep Learning, Motion analysis, Rehabilitation, Time-series
@article{mennella_promoting_2024,
title = {Promoting fairness in activity recognition algorithms for patient’s monitoring and evaluation systems in healthcare},
author = {Ciro Mennella and Massimo Esposito and Giuseppe De Pietro and Umberto Maniscalco},
url = {https://linkinghub.elsevier.com/retrieve/pii/S0010482524009119},
doi = {10.1016/j.compbiomed.2024.108826},
issn = {00104825},
year = {2024},
date = {2024-09-01},
urldate = {2024-07-21},
journal = {Computers in Biology and Medicine},
volume = {179},
pages = {108826},
abstract = {Researchers face the challenge of defining subject selection criteria when training algorithms for human activity recognition tasks. The ongoing uncertainty revolves around which characteristics should be considered to ensure algorithmic robustness across diverse populations. This study aims to address this challenge by conducting an analysis of heterogeneity in the training data to assess the impact of physical characteristics and soft-biometric attributes on activity recognition performance.
The performance of various state-of-the-art deep neural network architectures (tCNN, hybrid-LSTM, Transformer model) processing time-series data using the IntelliRehab (IRDS) dataset was evaluated. By intentionally introducing bias into the training data based on human characteristics, the objective is to identify the characteristics that influence algorithms in motion analysis.
Experimental findings reveal that the CNN-LSTM model achieved the highest accuracy, reaching 88%. Moreover, models trained on heterogeneous distributions of disability attributes exhibited notably higher accuracy, reaching 51%, compared to those not considering such factors, which scored an average of 33%. These evaluations underscore the significant influence of subjects’ characteristics on activity recognition performance, providing valuable insights into the algorithm’s robustness across diverse populations.
This study represents a significant step forward in promoting fairness and trustworthiness in artificial intelligence by quantifying representation bias in multi-channel time-series activity recognition data within the healthcare domain.},
keywords = {Artificial Intelligence, Bias, Deep Learning, Motion analysis, Rehabilitation, Time-series},
pubstate = {published},
tppubtype = {article}
}
Researchers face the challenge of defining subject selection criteria when training algorithms for human activity recognition tasks. The ongoing uncertainty revolves around which characteristics should be considered to ensure algorithmic robustness across diverse populations. This study aims to address this challenge by conducting an analysis of heterogeneity in the training data to assess the impact of physical characteristics and soft-biometric attributes on activity recognition performance.
The performance of various state-of-the-art deep neural network architectures (tCNN, hybrid-LSTM, Transformer model) processing time-series data using the IntelliRehab (IRDS) dataset was evaluated. By intentionally introducing bias into the training data based on human characteristics, the objective is to identify the characteristics that influence algorithms in motion analysis.
Experimental findings reveal that the CNN-LSTM model achieved the highest accuracy, reaching 88%. Moreover, models trained on heterogeneous distributions of disability attributes exhibited notably higher accuracy, reaching 51%, compared to those not considering such factors, which scored an average of 33%. These evaluations underscore the significant influence of subjects’ characteristics on activity recognition performance, providing valuable insights into the algorithm’s robustness across diverse populations.
This study represents a significant step forward in promoting fairness and trustworthiness in artificial intelligence by quantifying representation bias in multi-channel time-series activity recognition data within the healthcare domain.
The performance of various state-of-the-art deep neural network architectures (tCNN, hybrid-LSTM, Transformer model) processing time-series data using the IntelliRehab (IRDS) dataset was evaluated. By intentionally introducing bias into the training data based on human characteristics, the objective is to identify the characteristics that influence algorithms in motion analysis.
Experimental findings reveal that the CNN-LSTM model achieved the highest accuracy, reaching 88%. Moreover, models trained on heterogeneous distributions of disability attributes exhibited notably higher accuracy, reaching 51%, compared to those not considering such factors, which scored an average of 33%. These evaluations underscore the significant influence of subjects’ characteristics on activity recognition performance, providing valuable insights into the algorithm’s robustness across diverse populations.
This study represents a significant step forward in promoting fairness and trustworthiness in artificial intelligence by quantifying representation bias in multi-channel time-series activity recognition data within the healthcare domain.
Melillo, Antonio; Rachedi, Sarah; Caggianese, Giuseppe; Gallo, Luigi; Maiorano, Patrizia; Gimigliano, Francesca; Lucidi, Fabio; Pietro, Giuseppe De; Guida, Maurizio; Giordano, Antonio; Chirico, Andrea
In: Games for Health Journal, pp. g4h.2023.0202, 2024, ISSN: 2161-783X, 2161-7856.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Labor, Pain, User study, Virtual Reality
@article{melilloSynchronizationVirtualReality2024,
title = {Synchronization of a Virtual Reality Scenario to Uterine Contractions for Labor Pain Management: Development Study and Randomized Controlled Trial},
author = {Antonio Melillo and Sarah Rachedi and Giuseppe Caggianese and Luigi Gallo and Patrizia Maiorano and Francesca Gimigliano and Fabio Lucidi and Giuseppe De Pietro and Maurizio Guida and Antonio Giordano and Andrea Chirico},
url = {https://www.liebertpub.com/doi/10.1089/g4h.2023.0202},
doi = {10.1089/g4h.2023.0202},
issn = {2161-783X, 2161-7856},
year = {2024},
date = {2024-06-01},
urldate = {2024-07-21},
journal = {Games for Health Journal},
pages = {g4h.2023.0202},
abstract = {Background: Labor is described as one of the most painful events women can experience through their lives, and labor pain shows unique features and rhythmic fluctuations.
Purpose: The present study aims to evaluate virtual reality (VR) analgesic interventions for active labor with biofeedback-based VR technologies synchronized to uterine activity.
Materials and Methods: We developed a VR system modeled on uterine contractions by connecting it to cardiotocographic equipment. We conducted a randomized controlled trial on a sample of 74 cases and 80 controls during active labor.
Results: Results of the study showed a significant reduction of pain scores compared with both preintervention scores and to control group scores; a significant reduction of anxiety levels both compared with preintervention assessment and to control group and significant reduction in fear of labor experience compared with controls.
Conclusion: VR may be considered as an effective nonpharmacological analgesic technique for the treatment of pain and anxiety and fear of childbirth experience during labor. The developed system could improve personalization of care, modulating the multisensory stimulation tailored to labor progression. Further studies are needed to compare the synchronized VR system to uterine activity and unsynchronized VR interventions.},
keywords = {Artificial Intelligence, Labor, Pain, User study, Virtual Reality},
pubstate = {published},
tppubtype = {article}
}
Background: Labor is described as one of the most painful events women can experience through their lives, and labor pain shows unique features and rhythmic fluctuations.
Purpose: The present study aims to evaluate virtual reality (VR) analgesic interventions for active labor with biofeedback-based VR technologies synchronized to uterine activity.
Materials and Methods: We developed a VR system modeled on uterine contractions by connecting it to cardiotocographic equipment. We conducted a randomized controlled trial on a sample of 74 cases and 80 controls during active labor.
Results: Results of the study showed a significant reduction of pain scores compared with both preintervention scores and to control group scores; a significant reduction of anxiety levels both compared with preintervention assessment and to control group and significant reduction in fear of labor experience compared with controls.
Conclusion: VR may be considered as an effective nonpharmacological analgesic technique for the treatment of pain and anxiety and fear of childbirth experience during labor. The developed system could improve personalization of care, modulating the multisensory stimulation tailored to labor progression. Further studies are needed to compare the synchronized VR system to uterine activity and unsynchronized VR interventions.
Purpose: The present study aims to evaluate virtual reality (VR) analgesic interventions for active labor with biofeedback-based VR technologies synchronized to uterine activity.
Materials and Methods: We developed a VR system modeled on uterine contractions by connecting it to cardiotocographic equipment. We conducted a randomized controlled trial on a sample of 74 cases and 80 controls during active labor.
Results: Results of the study showed a significant reduction of pain scores compared with both preintervention scores and to control group scores; a significant reduction of anxiety levels both compared with preintervention assessment and to control group and significant reduction in fear of labor experience compared with controls.
Conclusion: VR may be considered as an effective nonpharmacological analgesic technique for the treatment of pain and anxiety and fear of childbirth experience during labor. The developed system could improve personalization of care, modulating the multisensory stimulation tailored to labor progression. Further studies are needed to compare the synchronized VR system to uterine activity and unsynchronized VR interventions.
Generosi, Andrea; Bruschi, Valeria; Cecchi, Stefania; Dourou, Nefeli Aikaterini; Montanari, Roberto; Mengoni, Maura
An Innovative System for Driver Monitoring and Vehicle Sound Interaction Proceedings Article
In: 2024 IEEE International Workshop on Metrology for Automotive (MetroAutomotive), pp. 159–164, IEEE, Bologna, Italy, 2024, ISBN: 979-8-3503-8498-7.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Deep Learning, Human Computer Interaction
@inproceedings{generosi_innovative_2024,
title = {An Innovative System for Driver Monitoring and Vehicle Sound Interaction},
author = {Andrea Generosi and Valeria Bruschi and Stefania Cecchi and Nefeli Aikaterini Dourou and Roberto Montanari and Maura Mengoni},
url = {https://ieeexplore.ieee.org/document/10615427/},
doi = {10.1109/MetroAutomotive61329.2024.10615427},
isbn = {979-8-3503-8498-7},
year = {2024},
date = {2024-06-01},
urldate = {2024-12-28},
booktitle = {2024 IEEE International Workshop on Metrology for Automotive (MetroAutomotive)},
pages = {159–164},
publisher = {IEEE},
address = {Bologna, Italy},
abstract = {An important aspect of Advanced Driver-Assistance Systems is the real-time monitoring of the driver and the interaction with him/her. In this scenario, the proposed work is focused on the development of an innovative system capable of analyzing the driver’s state and interact with him/her in a innovative way. In particular, the driver monitoring is obtained through the implementation of a multimodal approach that exploits deep learning and data fusion techniques while the interaction is achieved through sound signals elaborated with digital signal processing algorithm for the creation of an immersive scenario.},
keywords = {Artificial Intelligence, Deep Learning, Human Computer Interaction},
pubstate = {published},
tppubtype = {inproceedings}
}
An important aspect of Advanced Driver-Assistance Systems is the real-time monitoring of the driver and the interaction with him/her. In this scenario, the proposed work is focused on the development of an innovative system capable of analyzing the driver’s state and interact with him/her in a innovative way. In particular, the driver monitoring is obtained through the implementation of a multimodal approach that exploits deep learning and data fusion techniques while the interaction is achieved through sound signals elaborated with digital signal processing algorithm for the creation of an immersive scenario.
Melillo, Antonio; Rachedi, Sarah; Caggianese, Giuseppe; Gallo, Luigi; Maiorano, Patrizia; Gimigliano, Francesca; Lucidi, Fabio; Pietro, Giuseppe De; Guida, Maurizio; Giordano, Antonio; Chirico, Andrea
In: Games for Health Journal, pp. g4h.2023.0202, 2024, ISSN: 2161-783X, 2161-7856.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Labor, Pain, User study, Virtual Reality
@article{melillo_synchronization_2024,
title = {Synchronization of a Virtual Reality Scenario to Uterine Contractions for Labor Pain Management: Development Study and Randomized Controlled Trial},
author = {Antonio Melillo and Sarah Rachedi and Giuseppe Caggianese and Luigi Gallo and Patrizia Maiorano and Francesca Gimigliano and Fabio Lucidi and Giuseppe De Pietro and Maurizio Guida and Antonio Giordano and Andrea Chirico},
url = {https://www.liebertpub.com/doi/10.1089/g4h.2023.0202},
doi = {10.1089/g4h.2023.0202},
issn = {2161-783X, 2161-7856},
year = {2024},
date = {2024-06-01},
urldate = {2024-07-21},
journal = {Games for Health Journal},
pages = {g4h.2023.0202},
abstract = {Background: Labor is described as one of the most painful events women can experience through their lives, and labor pain shows unique features and rhythmic fluctuations.
Purpose: The present study aims to evaluate virtual reality (VR) analgesic interventions for active labor with biofeedback-based VR technologies synchronized to uterine activity.
Materials and Methods: We developed a VR system modeled on uterine contractions by connecting it to cardiotocographic equipment. We conducted a randomized controlled trial on a sample of 74 cases and 80 controls during active labor.
Results: Results of the study showed a significant reduction of pain scores compared with both preintervention scores and to control group scores; a significant reduction of anxiety levels both compared with preintervention assessment and to control group and significant reduction in fear of labor experience compared with controls.
Conclusion: VR may be considered as an effective nonpharmacological analgesic technique for the treatment of pain and anxiety and fear of childbirth experience during labor. The developed system could improve personalization of care, modulating the multisensory stimulation tailored to labor progression. Further studies are needed to compare the synchronized VR system to uterine activity and unsynchronized VR interventions.},
keywords = {Artificial Intelligence, Labor, Pain, User study, Virtual Reality},
pubstate = {published},
tppubtype = {article}
}
Background: Labor is described as one of the most painful events women can experience through their lives, and labor pain shows unique features and rhythmic fluctuations.
Purpose: The present study aims to evaluate virtual reality (VR) analgesic interventions for active labor with biofeedback-based VR technologies synchronized to uterine activity.
Materials and Methods: We developed a VR system modeled on uterine contractions by connecting it to cardiotocographic equipment. We conducted a randomized controlled trial on a sample of 74 cases and 80 controls during active labor.
Results: Results of the study showed a significant reduction of pain scores compared with both preintervention scores and to control group scores; a significant reduction of anxiety levels both compared with preintervention assessment and to control group and significant reduction in fear of labor experience compared with controls.
Conclusion: VR may be considered as an effective nonpharmacological analgesic technique for the treatment of pain and anxiety and fear of childbirth experience during labor. The developed system could improve personalization of care, modulating the multisensory stimulation tailored to labor progression. Further studies are needed to compare the synchronized VR system to uterine activity and unsynchronized VR interventions.
Purpose: The present study aims to evaluate virtual reality (VR) analgesic interventions for active labor with biofeedback-based VR technologies synchronized to uterine activity.
Materials and Methods: We developed a VR system modeled on uterine contractions by connecting it to cardiotocographic equipment. We conducted a randomized controlled trial on a sample of 74 cases and 80 controls during active labor.
Results: Results of the study showed a significant reduction of pain scores compared with both preintervention scores and to control group scores; a significant reduction of anxiety levels both compared with preintervention assessment and to control group and significant reduction in fear of labor experience compared with controls.
Conclusion: VR may be considered as an effective nonpharmacological analgesic technique for the treatment of pain and anxiety and fear of childbirth experience during labor. The developed system could improve personalization of care, modulating the multisensory stimulation tailored to labor progression. Further studies are needed to compare the synchronized VR system to uterine activity and unsynchronized VR interventions.
Raikov, Aleksandr; Giretti, Alberto; Pirani, Massimiliano; Spalazzi, Luca; Guo, Meng
Accelerating human–computer interaction through convergent conditions for LLM explanation Journal Article
In: Frontiers in Artificial Intelligence, vol. 7, 2024, ISSN: 2624-8212, (Publisher: Frontiers).
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Causal Loop Dynamics, Cognitive semantics, Eigenform, Explainable AI, Hybrid reality, LLM
@article{raikovAcceleratingHumanComputer2024,
title = {Accelerating human–computer interaction through convergent conditions for LLM explanation},
author = {Aleksandr Raikov and Alberto Giretti and Massimiliano Pirani and Luca Spalazzi and Meng Guo},
url = {https://www.frontiersin.org/journals/artificial-intelligence/articles/10.3389/frai.2024.1406773/full},
doi = {10.3389/frai.2024.1406773},
issn = {2624-8212},
year = {2024},
date = {2024-05-01},
urldate = {2024-10-05},
journal = {Frontiers in Artificial Intelligence},
volume = {7},
publisher = {Frontiers},
abstract = {<p>The article addresses the accelerating human–machine interaction using the large language model (LLM). It goes beyond the traditional logical paradigms of explainable artificial intelligence (XAI) by considering poor-formalizable cognitive semantical interpretations of LLM. XAI is immersed in a hybrid space, where humans and machines have crucial distinctions during the digitisation of the interaction process. The author's convergent methodology ensures the conditions for making XAI purposeful and sustainable. This methodology is based on the inverse problem-solving method, cognitive modeling, genetic algorithm, neural network, causal loop dynamics, and eigenform realization. It has been shown that decision-makers need to create unique structural conditions for information processes, using LLM to accelerate the convergence of collective problem solving. The implementations have been carried out during the collective strategic planning in situational centers. The study is helpful for the advancement of explainable LLM in many branches of economy, science and technology.</p>},
note = {Publisher: Frontiers},
keywords = {Artificial Intelligence, Causal Loop Dynamics, Cognitive semantics, Eigenform, Explainable AI, Hybrid reality, LLM},
pubstate = {published},
tppubtype = {article}
}
<p>The article addresses the accelerating human–machine interaction using the large language model (LLM). It goes beyond the traditional logical paradigms of explainable artificial intelligence (XAI) by considering poor-formalizable cognitive semantical interpretations of LLM. XAI is immersed in a hybrid space, where humans and machines have crucial distinctions during the digitisation of the interaction process. The author's convergent methodology ensures the conditions for making XAI purposeful and sustainable. This methodology is based on the inverse problem-solving method, cognitive modeling, genetic algorithm, neural network, causal loop dynamics, and eigenform realization. It has been shown that decision-makers need to create unique structural conditions for information processes, using LLM to accelerate the convergence of collective problem solving. The implementations have been carried out during the collective strategic planning in situational centers. The study is helpful for the advancement of explainable LLM in many branches of economy, science and technology.</p>
Jamali, Reza; Generosi, Andrea; Villafan, Josè Yuri; Mengoni, Maura; Pelagalli, Leonardo; Battista, Gianmarco; Martarelli, Milena; Chiariotti, Paolo; Mansi, Silvia Angela; Arnesano, Marco; Castellini, Paolo
Facial Expression Recognition for Measuring Jurors’ Attention in Acoustic Jury Tests Journal Article
In: Sensors, vol. 24, no. 7, pp. 2298, 2024, ISSN: 1424-8220.
Abstract | Links | BibTeX | Tags: Artificial Intelligence, Computer Vision and Pattern Recognition, Deep Learning, Emotion Recognition
@article{jamali_facial_2024,
title = {Facial Expression Recognition for Measuring Jurors’ Attention in Acoustic Jury Tests},
author = {Reza Jamali and Andrea Generosi and Josè Yuri Villafan and Maura Mengoni and Leonardo Pelagalli and Gianmarco Battista and Milena Martarelli and Paolo Chiariotti and Silvia Angela Mansi and Marco Arnesano and Paolo Castellini},
url = {https://www.mdpi.com/1424-8220/24/7/2298},
doi = {10.3390/s24072298},
issn = {1424-8220},
year = {2024},
date = {2024-04-01},
urldate = {2024-12-28},
journal = {Sensors},
volume = {24},
number = {7},
pages = {2298},
abstract = {The perception of sound greatly impacts users’ emotional states, expectations, affective relationships with products, and purchase decisions. Consequently, assessing the perceived quality of sounds through jury testing is crucial in product design. However, the subjective nature of jurors’ responses may limit the accuracy and reliability of jury test outcomes. This research explores the utility of facial expression analysis in jury testing to enhance response reliability and mitigate subjectivity. Some quantitative indicators allow the research hypothesis to be validated, such as the correlation between jurors’ emotional responses and valence values, the accuracy of jury tests, and the disparities between jurors’ questionnaire responses and the emotions measured by FER (facial expression recognition). Specifically, analysis of attention levels during different statuses reveals a discernible decrease in attention levels, with 70 percent of jurors exhibiting reduced attention levels in the ‘distracted’ state and 62 percent in the ‘heavy-eyed’ state. On the other hand, regression analysis shows that the correlation between jurors’ valence and their choices in the jury test increases when considering the data where the jurors are attentive. The correlation highlights the potential of facial expression analysis as a reliable tool for assessing juror engagement. The findings suggest that integrating facial expression recognition can enhance the accuracy of jury testing in product design by providing a more dependable assessment of user responses and deeper insights into participants’ reactions to auditory stimuli.},
keywords = {Artificial Intelligence, Computer Vision and Pattern Recognition, Deep Learning, Emotion Recognition},
pubstate = {published},
tppubtype = {article}
}
The perception of sound greatly impacts users’ emotional states, expectations, affective relationships with products, and purchase decisions. Consequently, assessing the perceived quality of sounds through jury testing is crucial in product design. However, the subjective nature of jurors’ responses may limit the accuracy and reliability of jury test outcomes. This research explores the utility of facial expression analysis in jury testing to enhance response reliability and mitigate subjectivity. Some quantitative indicators allow the research hypothesis to be validated, such as the correlation between jurors’ emotional responses and valence values, the accuracy of jury tests, and the disparities between jurors’ questionnaire responses and the emotions measured by FER (facial expression recognition). Specifically, analysis of attention levels during different statuses reveals a discernible decrease in attention levels, with 70 percent of jurors exhibiting reduced attention levels in the ‘distracted’ state and 62 percent in the ‘heavy-eyed’ state. On the other hand, regression analysis shows that the correlation between jurors’ valence and their choices in the jury test increases when considering the data where the jurors are attentive. The correlation highlights the potential of facial expression analysis as a reliable tool for assessing juror engagement. The findings suggest that integrating facial expression recognition can enhance the accuracy of jury testing in product design by providing a more dependable assessment of user responses and deeper insights into participants’ reactions to auditory stimuli.