Higher Institute on Territorial Systems for Innovation
PORTO@iris (Publications Open Repository TOrino - Politecnico di Torino)Not a member yet
146173 research outputs found
Sort by
The longue durée of a flood: studying urban landscape transformations from a disaster resilience perspective
No full textDisasters are not sudden, unexpected ruptures; they are the result of long-term socio-environmental and technological transformations that generate exposure and vulnerability. To understand how risk is produced and sustained, it is essential to develop a historical perspective which draws on longue duree readings of territorial, infrastructural, and political change. Focusing on Moncalieri (metropolitan Turin) and the flood of 15–16 October 2000, this article develops a three-scale reading that combines archival sources with GIS-based spatial analysis. At the event scale, it reconstructs the emergency to show how infrastructures, decision points, and coordination practices shaped the distribution of damage. At the urban scale, it tracks the coupling of river works, energy production, canal diversions, and settlement growth along the Po, showing how development concentrated people, assets, and critical services in flood-prone areas. At the Po-basin scale, it situates Moncalieri within wider shifts in river management, highlighting how interventions and responsibilities displace risk across space and time. Rather than treating the flood as an anomaly, it reads historical and territorial production of risk. Read together, these scales foreground the processes through which exposure and vulnerability are historically assembled and argue for the necessity of a longue duree perspective to grasp their complexity
Methodological Frameworks for Computational Electrocatalysis: From Theory to Practice
No full textModeling electrocatalytic reactions at solid–liquid interfaces requires capturing both the quantum-mechanical processes at the electrode surface and the complex response of the surrounding electrochemical environment. This review examines the main theoretical frameworks and computational techniques used to describe such systems, focusing on first-principles approaches based on density functional theory (DFT). Key aspects include the treatment of reaction thermodynamics, electrode bias, solvation effects, electrolyte screening, and reaction kinetics. A broad range of methods is discussed, from thermochemical models, such as the computational hydrogen electrode, to potential-dependent formulations based on grand-canonical DFT and explicit calculation of kinetic barriers. The review also highlights recent machine-learning approaches for catalyst screening and the growing use of machine-learning-based force fields, which promise to enable efficient simulations of complex electrochemical environments over extended time and length scales with near-first-principles accuracy. The aim is not only to present the state of the art, but also to clarify the physical assumptions and approximations underlying each approach. The influence of modeling choices on reliability and computational cost is examined in detail. Alongside theoretical aspects, practical considerations are emphasized to support researchers in selecting appropriate methods and designing simulations that are both physically meaningful and computationally tractable
Toward data-driven machining: Prediction of surface roughness in high-entropy alloy coatings using a stacking ensemble machine learning model
No full textIn mechanical engineering, predictive tools are increasingly used to enable faster analysis. The study began with cold
spraying of high-entropy alloy coatings, Al0.1-0.5CoCrCuFeNi and MnCoCrCuFeNi coatings at nitrogen gas
temperatures of 650 C, 750 C, and 850 C. The surface roughness, Ra, as the output, was measured using profilometry and microscopy techniques. The experimental sample set consisted of 20 samples of different experiments. The stacking ensemble structure included three base learners: Linear Regression, Extreme Gradient Boosting, and Gaussian Process Regression with RidgeCV as the meta-learner. This ensemble was compared with Extreme Gradient Boosting, which is a powerful single machine learning approach. Results indicated that the stacking ensemble performed better than Extreme Gradient Boosting in all the regression metrics. Extreme Gradient Boosting reached RMSE of 0.22mm, MAPE of 3.32%, and R2 of 0.83, whereas the stacking ensemble achieved RMSE of 0.17 mm, MAPE of 2.73%, and R2 of 0.89. A sensitivity analysis was used to qualitatively assess the influence of input variables. The results suggest that stacking ensembles can improve predictive performance even in scenarios of small data
The future of mathematical oncology in the age of AI
No full textThis perspective article discusses emerging advances at the interface of mechanistic modeling and data-driven machine learning, highlighting opportunities for AI to accelerate discovery, improve predictive modeling, and enhance clinical decision-making. We address critical limitations of current AI approaches and propose a perspective on a future where AI augments mechanistic rigor, clinical relevance, and human creativity under the umbrella of a redefined understanding of Mathematical Oncology
UV-curable thiol-ene networks with intrinsic antioxidant functionality from eugenol-derived triallyl isocyanurate
No full textThe development of bio-based polymers with intrinsic antioxidant functionality offers a sustainable strategy to
enhance material longevity and stability while reducing reliance on migratory additives. Herein, a triallyl isocyanurate monomer (TEG) was synthesized via the reaction of eugenol, a renewable phenolic compound, with
thermally stable hexamethylene diisocyanate isocyanurate. TEG was subsequently employed to construct functional thiol-ene networks through UV-initiated thiol-ene photopolymerization with multifunctional thiols.
Additional formulations incorporating pristine eugenol were designed to preserve pendant phenolic hydroxyl
groups, thereby imparting enhanced radical-scavenging capacity. The photopolymerization kinetics, monitored
by real-time FTIR and photo-DSC, revealed rapid curing with high thiol conversion (≈78%) achieved within 2
min of irradiation. A slight retardation (conversion ≈72%) was observed in systems containing pendant phenolic
hydroxyl due to their free radical quenching effect. Dynamic mechanical analysis confirmed that increasing thiol
functionality led to higher crosslink densities and elevated glass transition temperatures (Tg = 16–38 ◦C), while
tensile testing demonstrated tunable stiffness (Et = 1.2–54.3 MPa) and elongation (εM = 47–63%). The resulting
networks exhibited excellent optical transparency (>80% transmittance at 500 nm), effective UV-shielding
(~0% transmittance below 320 nm), and outstanding thermal stability (Tmax ≈ 320 and 460 ◦C). DPPH assay
verified strong intrinsic antioxidant activity especially for phenolic hydroxyl containing formulations, achieving
up to 86% radical scavenging. Collectively, these findings establish a sustainable design approach for multifunctional thiol-ene networks that unite mechanical adaptability, UV protection, thermal robustness, and built-in
antioxidant functionality, suitable for applications in flexible coatings, packaging, and protective material
Artificial Intelligence in Gastrointestinal Disease Diagnosis: A Systematic Review of Endoscopy, Histology, and Radiology Applications
No full textGastrointestinal (GI) diseases remain among the leading causes of global mortality, with early detection directly linked to survival outcomes. While previous reviews have focused on single imaging modalities, this systematic review uniquely examines artificial intelligence applications across endoscopic, radiological, and histological approaches, reflecting actual clinical diagnostic pathways. This systematic review analyzes 76 high-quality studies (2016–2024) and provides the first comprehensive assessment of how AI performs across different imaging techniques for GI abnormality detection. This multi-modal perspective is particularly timely as healthcare systems move toward integrated diagnostic workflows. Our analysis reveals endoscopy as the most widely used modality (n = 44), particularly for Helicobacter pylori, colorectal polyps, and ulcerative colitis detection. Histological analysis emerges as the second most common approach (n = 25), especially for celiac disease and ulcerative colitis, while CT imaging (n = 10) primarily supports colorectal polyp detection. Deep learning methods significantly outnumber traditional machine learning techniques (68 vs. 8 studies), consistently achieving 90%–99% diagnostic accuracy across multiple disease categories. However, these systems face significant implementation barriers to clinical adoption. Most validation is still conducted in controlled, single-center settings using curated datasets that poorly reflect clinical complexity. Future studies must prioritize multicenter validation, standardized imaging protocols and preprocessing pipelines, and the integration of interpretable AI models capable of providing transparent diagnostic rationale. This review maps the current technical landscape while highlighting critical translational challenges that must be addressed to enable real-world impact
An Automated Diagram Generator of Reference Solutions for Modeling Educators
No full textUML class diagrams are a relevant modeling language in Software Engineering education since they can be used to teach students how to visualize and display the different entities that compose a system, with their functionalities and relationships. The definition of modeling exercises and their evaluation can be time-consuming for educators due to the need to consider possible semantic variations and alternative representations of the same system requirements. To facilitate teachers in this process, we present TIGRE (auTomated dIagram Generator of REference solutions), an online editor for the definition of UML modeling exercises where teachers can define reference solutions in the form of both diagrams and detailed structures to be used for automated evaluation. The tool is enhanced by the interaction with recent Large Language Models for the automated generation of reference solutions starting from text, facilitating the creation of early drafts. A proof-of-concept case study has been performed by having TIGRE generate reference solutions for two exercises: most of the relevant concepts have been represented correctly, but issues emerged in the form of unnecessary classes being included and incorrect understanding of associations
State of art in regularization methods for numerical analysis of structures with softening
No full textThis paper provides an extensive review of popular regularization methods utilized in numerical models to stabilize the structural response of materials exhibiting significant softening. The necessity for regularization is highlighted in cases of material softening, which is attributed to the loss of ellipticity in the governing differential equations. It discusses the advantages and disadvantages of the regularization methods most commonly employed in the scientific community. Furthermore, the paper highlights recent advancements, particularly in defining internal length within nonlocal models and characteristic element length in fracture energy regularization methods, as alternative solutions to address the limitations inherent in traditional approaches
Privacy-Preserving Container Attestation
No full textContainerization techniques have become essential to develop and deploy distributed applications in a Cloud Computing scenario. Containers' popularity continues to grow due to their flexibility, lightness, and availability. Despite their advantages, containers offer less isolation than virtual machines since they share the host's kernel. Therefore, attacks on a container could compromise other containers on the same node, or the host system itself. Trustworthiness in containers' operations is strictly related to demonstration of their software integrity and proper configuration, as these things are vital for early detection of tampering and breaches, and for fast response to attacks. The Trusted Computing paradigm offers techniques to attest the trustworthiness of a physical node, but they are not directly usable to attest containers due to the virtualization layer. Our work leverages the recently introduced Linux IMA namespace to achieve container attestation. Since attestation reveals the list of software components and configurations, the privacy of this operation is crucial in a multi-tenant scenario. Our solution ensures that a tenant authorized to attest a given container has access exclusively to the information of that container and its dependencies. We integrated this solution into an existing attestation framework to create a complete solution for privacy-preserving container integrity verification in a multi-tenant scenario. Our approach offers low latency for event measurement and a fast verification process, regardless of the number of containers or the containerization technology used