68994 research outputs found
Sort by
Learning-Based Data-Driven Control with Applications in Robotics: from Kernel Methods to Neuromorphic Approaches
No full textIn questa tesi si esplorano metodologie di progettazione del controllo nel paradigma del direct data-driven learning, con particolare attenzione a strategie iterative che migliorano la legge di controllo direttamente dai dati sperimentali, evitando l’identificazione esplicita del modello. Gli approcci proposti sono validati in applicazioni robotiche su robot mobili e manipolatori, con l’obiettivo di ottenere buone prestazioni in anello chiuso e un’elevata efficienza computazionale. Un primo contributo introduce un metodo basato su kernel per l’apprendimento di controllori feedforward in un framework di Iterative Learning Control (ILC) per sistemi non lineari. A differenza degli schemi ILC tradizionali, l’aggiornamento del controllo è puramente data-driven. Il metodo è confrontato con un ILC convenzionale e con un controllore ILC basato su rete neurale, mostrando in simulazione migliori prestazioni di tracking e convergenza in un problema di path-following di un robot uniciclo. I capitoli successivi presentano applicazioni di imitation learning, in cui un controllore esperto ad alte prestazioni, nello specifico un Model Predictive Controller non lineare, è utilizzato per generare dati di addestramento. Tali dati sono impiegati per addestrare un controllore Linear Parameter-Varying basato su kernel ed un controllore basato su Spiking Neural Networks (SNN), valutati in simulazione applicati al controllo di un manipolatore. Infine, la tesi descrive un’implementazione reale di un controllore SNN per il controllo della velocità delle ruote di un robot mobile, dimostrandone la trasferibilità su hardware reale. Nel complesso, i risultati mostrano che i metodi proposti rappresentano valide alternative data-driven ai controlli basati su modello, consentendo prestazioni comparabili e frequenze di controllo più elevate. Prospettive future includono analisi di stabilità in anello chiuso, valutazioni di robustezza formale ed estensioni verso controllori adattivi online. This thesis explores control design methodologies within the direct data-driven learning paradigm, with a particular focus on iterative strategies that improve the control law directly from experimental data, avoiding explicit model identification. The proposed approaches are validated in robotic applications involving mobile robots and manipulators, with the aim of achieving good closed-loop performance and high computational efficiency. A first contribution introduces a kernel-based method for learning feedforward controllers within an Iterative Learning Control (ILC) framework for nonlinear systems. Unlike traditional ILC schemes, the control update is purely data-driven. The method is compared with a conventional ILC approach and a neural-network-based controller, showing improved tracking performance and convergence in simulation on a path-following problem for a unicycle robot. The following chapters present imitation learning applications, where a high-performance expert controller, specifically a nonlinear Model Predictive Controller, is used to generate training data. These data are employed to train a kernel-based Linear Parameter Varying controller and a controller based on Spiking Neural Networks (SNNs), which are evaluated in simulation, being applied to control the pose of the end-effector of a manipulator. Finally, the thesis describes a real-world implementation of an SNN-based controller for wheel speed control of a mobile robot, demonstrating transferability to real hardware. Overall, the results show that the proposed methods represent valid data-driven alternatives to model-based control strategies, enabling comparable performance and higher control frequencies. Future perspectives include closed-loop stability analysis, formal robustness evaluation, and extensions toward adaptive online controllers.
Endovascular treatment of true visceral artery aneurysms: a decade of experience and key outcomes from a high-volume single center
No full textBackground: Endovascular approach has emerged as the first-choice treatment of visceral artery aneurysms (VAAs). However, most published series include heterogeneous populations combining true aneurysms and pseudoaneurysms. Our single-center retrospective study aimed to evaluate a large and homogenous series of true VAAs treated endovascularly. Demographic, clinical and procedural details of patients with VAAs treated at our Interventional Radiology unit from January 2014 to January 2025 were collected in a dedicated database. Intraprocedural and perioperative outcomes included technical success, complications, mortality and need for reintervention. At long-term follow-up, overall survival, aneurysm-related complications, freedom from reintervention, and aneurysmal sac reperfusion were evaluated. Results: Ninety-four consecutive patients (41 men; median age: 67.3 years; range 29–98 years) with 97 VAAs were included in the study. VAAs had a median diameter of 2.93 cm (range 6–9.2 cm). Treatment strategies included sac packing in 45 VAAs, sandwich technique in 35, covered stent placement in 13, afferent artery occlusion in 5 and stent-assisted coiling in 4 cases. Technical success rate was 100% with no perioperative deaths. Perioperative major complication and reintervention rates were 5.2% and 2.1%, respectively. At a mean follow-up of 44.7 ± 28.9 months, overall survival, freedom from aneurysm-related complications, and freedom from reintervention were 89.4%, 92.8%, and 93.8%, respectively. Aneurysm sac reperfusion was observed in 8(8.2%) cases. No aneurysm-related deaths occurred. Conclusion: Endovascular approach is a safe and effective first-line option in patients with true VAAs, with notable results in terms of complication and freedom from reintervention rates and an excellent long-term overall survival
Privacy-preserving federated learning for multi-regional disability employment matching: a comprehensive framework with differential privacy and blockchain anchoring
No full textEmployment inclusion for people with disabilities remains a persistent challenge across developed nations, with employment rates 24 percentage points below general population levels in Europe. Traditional centralized employment matching systems face privacy constraints, data sovereignty requirements, and limited collaborative learning capabilities. This paper presents a comprehensive privacy-preserving federated learning framework for multi-regional disability employment matching that enables collaborative model training while maintaining data locality and regulatory compliance. Our approach combines ensemble-based LightGBM federation with parameter-level MLP federation, achieving 0.9011 F1-score (LightGBM) and 0.7881 F1-score (MLP with privacy preservation) across five Italian employment centers in Veneto region. The privacy framework integrates differential privacy with RDP composition (ε = 1.0, δ = 10−6), Shamir’s 3-of-5 secret sharing for secure aggregation, and blockchain anchoring for long-term integrity verification. Empirical evaluation demonstrates minimal performance degradation under federated constraints (0.0005 F1-score loss) while providing formal privacy guarantees. The system reduces manual processing time from 30–60 min to under 5 min per candidate with sub-100 ms response times suitable for real-world deployment. Ongoing pilot deployment at partner employment centers validates practical applicability for European disability employment services. The research contributes to trustworthy AI frameworks for sensitive social applications requiring regulatory compliance, algorithmic fairness, and privacy preservation
Strawberry Fragaria x ananassa Duchesne (cv. Romina) methanolic extract attenuates hydrogen peroxide induced DNA damage in human peripheral blood mononuclear cells in vitro
No full textInvestment valuation of photovoltaic and energy storage systems for diverse energy communities: A real option approach
No full textThe use of energy storage devices and renewable energy sources has evolved over the past few decades from an individual to a community concept, whereby the energy produced and stored is distributed among energy community participants. In this context, there is a need to provide an investment strategy to decide whether and when to invest in a photovoltaic-battery system to cover the energy demand. Therefore, we propose an optimization methodology to assess the profitability of these systems to be installed in such energy communities while exploiting the variability of the demand curves from different types of users. Moreover, considering the future uncertainty of photovoltaic production, energy demand, and price of electricity, we evaluate the possibility of deferring the investment by employing a real option approach. We perform the analysis on several energy communities composed of different types of users. Results show that being part of a diverse energy community guarantees savings in terms of total costs and battery capacity, together with a reduction in investment time and an increase in the value of the option
Non‐epistemic Values and the Automation of Science
No full textIn our paper, we reassess the role of non-epistemic values in scientific practice by drawing lessons from machine learning and the automation of science. Due to several influential arguments (e.g., Rudner 1953, or Longino 1990), traditional philosophy of science has largely converged on the view that non-epistemic values are necessary for the justification of scientific claims. Recently, renewed support of this view has been made by appealing to the No Free Lunch theorems in the context of machine learning—that there is no universally optimal machine learning algorithm (Dotan 2020). The argument claims that the No Free Lunch theorems entail that epistemic values are insufficient for discriminating hypotheses. In the negative part of our paper, we critique Dotan’s argument. We argue that NFL theorems do not entail that ‘all hypotheses have the same average expected error’. We also discuss what NFL theorems do in fact say about inductive inference and theory choice. In the positive part, we argue that the possibility of ‘value-free science’ (understood as a science that involves only epistemic values in the context of justification) seems to be supported by developments in the automation of science. For example, Robot Scientist presented in King et al. (2004) can automatically propose hypotheses, design and conduct experiments, and interpret the results. We argue that in these systems, only epistemic values are involved in hypothesis construction and testing
A Convergent Approach to Investigate the Environmental Behavior and Importance of a Man-Made Saltwater Wetland
No full textMediterranean saline wetlands are significant ecological habitats defined by seasonal water availability and various biological communities, forming a unique ecotone that combines traits of both freshwater and marine environments. Moreover, they are regarded as notable natural and economic resources. Since the sustainable management of protected wetlands necessitates a multidisciplinary approach, the purpose of this study is to provide a comprehensive picture of the hydrological, hydrochemical, and ecological dynamics of a man-made groundwater dependent ecosystem (GDE) by combining remote sensing, hydrochemical data, geostatistical tools, and ecological indicators. The study area, called “Le Soglitelle”, is located in the Campania plain (Italy), which is close to the Domitian shoreline, covering a surface of 100 ha. The Normalized Difference Water Index (NDWI), a remote sensing-derived index sensitive to surface water presence, from Sentinel-2 was used to detect changes in the percentage of the wetland inundated area over time. Water samples were collected in four campaigns, and hydrochemical indexes were used to investigate the major hydrochemical seasonal processes occurring in the area. Geostatistical tools, such as principal component analysis (PCA) and independent component analysis (ICA), were used to identify the main hydrochemical processes. Moreover, faunal monitoring using waders was employed as an ecological indicator. Seasonal variation in the inundation area ranged from nearly 0% in summer to over 50% in winter, consistent with the severe climatic oscillations indicated by SPEI values. PCA and ICA explained over 78% of the total hydrochemical variability, confirming that the area’s geochemistry is mainly characterized by the saltwater sourced from the artesian wells that feed the wetland. The concentration of the major ions is regulated by two contrasting processes: evapoconcentration in summer and dilution and water mixing (between canals and ponds water) in winter. Cl−/Br− molar ratio results corroborated this double seasonal trend. The base exchange index highlighted a salinization pathway for the wetland. Bird monitoring exhibited consistency with hydrochemical monitoring, as the seasonal distribution clearly reflects the dual behaviour of this area, which in turn augmented the biodiversity in this GDE. The integration of remote sensing data, multivariate geostatistical analysis, geochemical tools, and faunal indicators represents a novel interdisciplinary framework for assessing GDE seasonal dynamics, offering practical insights for wetland monitoring and management
Insights from comparison of serum IgG4 between T2-eosinophilic asthma and eosinophilic granulomatosis with polyangiitis/idiopathic hypereosinophilic syndrome
No full textObjective: Eosinophilic granulomatosis with polyangiitis (EGPA) and idiopathic hypereosinophilic syndrome (iHES) are systemic hypereosinophilic diseases largely overlapping. Type 2 (T2)-eosinophilic asthma is documented in 100% and 44-95% of EGPA and iHES patients, respectively, probably representing the beginning of the systemic eosinophilic disorders, as suggested by mepolizumab, effective in both asthma and EGPA/iHES. In this respect, there are no predictive biomarkers of progression from T2-eosinophilic asthma into EGPA/iHES. Immunoglobulins G type 4 (IgG4) take part in T2-eosinophilic inflammation, and elevated serum IgG4 have been previously documented in asthma, EGPA and HES. The objective of this study was to compare serum IgG4 between T2-eosinophilic asthma and EGPA/iHES in order to identify significant differences that could represent a reasonable background for prospective studies on IgG4 as potential predictive biomarker of progression from asthma to EGPA/iHES. Methods In this retrospective/cross-sectional case-control study, patients affected by T2-eosinophilic asthma or EGPA/iHES were consecutively enrolled. All patients underwent blood tests for serum IgG4. Asthmatics and EGPA/iHES patients were stratified upon serum IgG4 values (normal or elevated). Results 62 patients were enrolled (27 asthmatics, 35 EGPA/iHES). The frequency of patients with elevated serum IgG4 was higher in the EGPA/iHES group than in the asthmatics (45.7% vs. 11.1%, p=0.003), as well as the mean serum IgG4 value (p=0.010). Conclusion Elevated serum IgG4 seemed to discriminate between T2-eosinophilic asthma “alone” and T2-eosinophilic asthma evolved into EGPA/iHES. This finding could represent a reasonable background for prospective long-term studies on asthmatic naive patients aimed at investigating IgG4 as a potential predictive biomarker of progression from asthma to EGPA/iHES
Arthroscopic Medial Meniscus Posterior Root Repair and Centralization With Medial Capsulodesis Using a Knotless Soft Anchor
No full textMeniscal extrusion after posterior root tears of the medial meniscus disrupts the knee hoop-stress mechanism and increases joint contact pressure, accelerating osteoarthritis. Even with an anatomic root repair, extrusion often persists. Meniscal centralization has been proposed as a complementary strategy to restore joint biomechanics and reduce stress on the repaired root. This technique combines transtibial pull-out repair of the medial meniscus posterior root with meniscal centralization and medial capsulodesis. A tibial tunnel is created for root fixation using a FiberWire suture, and centralization is performed using an outside-in technique with 2 spinal needles and a single 1.8-mm knotless anchor. This allows simultaneous stabilization of the medial capsule and the meniscotibial ligament. This approach offers improved anchoring of the medial meniscus for effective centralization through reinforcement of capsule and meniscotibial ligament with reduced need for expensive or bulky instruments. The use of spinal needles lowers the risk of iatrogenic cartilage injury. This combined method is simple, efficient, and cost-effective. It enhances biomechanical stability, minimizes extrusion, and reduces surgical complexity potentially leading to better outcomes in treating posterior medial meniscal root tears
Coral health status before and after the tourism halt caused by the COVID-19 pandemic in Koh Tao (Thailand)
No full textBlue tourism is often the primary income for small tropical islands, but uncontrolled over tourism can cause irreversible detrimental effects to coral reefs. Among impacts of tourism growth, recreational diving can physically damage corals through direct contact or sediment resuspension, while nutrient enrichment, marine litter and increased sediment loads into the sea can reduce corals' immunocompetence and increase their susceptibility to disease. This study assessed the effects of decreased human activity during the COVID-19 pandemic on Koh Tao' coral reefs, (Thailand), comparing the coral health status in the pre- (2016-2019), during- (2021), and post-pandemic (2022-2024) periods. Underwater surveys were conducted from 2016 to 2024 across five bays by considering seven main coral health status categories and multiple subcategories. Results showed a general worsening of coral health conditions from the pre- to the post-pandemic period; diversity declined, with three coral families disappearing in one of the sites, while an overall increase in Acroporidae and severe reduction in Fungiidae were observed everywhere. Bleaching and White Syndrome were the primary disorders affecting corals, while Skeletal Eroding Band, Yellow, Brown, and Black Band diseases appear to develop as secondary infections. The combined effects of extreme temperature and intense rainfall together with decennial unregulated coastal development have caused a dramatic decline in reef health and biodiversity. This suggests that a brief tourism halt was insufficient to reverse chronic disturbances caused by climatic and anthropogenic stressors. Coordinated conservation actions are urgently needed to protect coral reefs in Koh Tao and in small tropical islands with high levels of unregulated tourism in general