HAL Portal UTC Université de Technologie de Compiègne
Not a member yet
11652 research outputs found
Sort by
Modeling and Quantification of Cyber-to-Safety Impact Propagation in Railway Signalling Systems
No full textInternational audienceThe digitalization of safety-critical systems, particularly in the railway sector, increases exposure to cyber threats such as jamming and spoofing, which can escalate into hazardous operational failures. Although current standards acknowledge the interdependence between safety and security, they lack quantitative methods to assess how cyber attacks propagate into safety degradation. This paper presents the Security Safety Stochastic Game Framework (S3G), a two layer quantitative model that links cybersecurity decisions with safety performance in complex cyber physical systems. The inner stochastic game captures dynamic attacker-defender interactions, while the outer Stackelberg layer determines the defender configuration, redundancy, transmission power, etc., that minimizes worst-case safety loss. To demonstrate its effectiveness, the framework is instantiated on the ERTMS/ETCS Level 2 signalling architecture. Results show that S3G identifies configurations that enhance safety resilience, mitigate integrity compromise, and accelerate recovery from cyber disruptions, while maintaining acceptable operational overhead. This confirms the relevance of integrating safety semantics into cybersecurity optimization to achieve resilient equilibria in safety-critical domains
Reducing Aleatoric and Epistemic Uncertainty through Multi-modal Data Acquisition
No full textTo generate accurate and reliable predictions, modern AI systems need to combine data from multiple modalities, such as text, images, audio, spreadsheets, and time series. Multi-modal data introduces new opportunities and challenges for disentangling uncertainty: it is commonly assumed in the machine learning community that epistemic uncertainty can be reduced by collecting more data, while aleatoric uncertainty is irreducible. However, this assumption is challenged in modern AI systems when information is obtained from different modalities. This paper introduces an innovative data acquisition framework where uncertainty disentanglement leads to actionable decisions, allowing sampling in two directions: sample size and data modality. The main hypothesis is that aleatoric uncertainty decreases as the number of modalities increases, while epistemic uncertainty decreases by collecting more observations. We provide proof-of-concept implementations on two multi-modal datasets to showcase our data acquisition framework, which combines ideas from active learning, active feature acquisition and uncertainty quantification
Transformation numérique et innovation dans les secteurs traditionnels d'activité: Une approche critique des mutations du travail des artisans d'art
No full textInternational audienceThe crafts sector is emblematic of traditional activities deeply rooted in their local communities, where work practices have historically relied on manual skills, the transmission of techniques, and small-scale organizational structures. However, in recent years, the increasing adoption of digital technologies has profoundly challenged the methods of production, cooperation, and creation within these organizations. This paper adopts a critical approach to digitalization and aims to address the following question: how does the adoption of digital technology transform the work of craftspeople?Using a territorial perspective, this research is based on a qualitative study conducted in the village of Pernes-les-Fontaines, renowned for its high concentration of artists and craftspeople. The methodological approach is based on long-term participant observation and semi-structured interviews conducted with artists and craftspeople, allowing us to understand their actual practices, their perceptions of digital technology, and the tensions they experience on a daily basis.L’artisanat d’art constitue un secteur emblématique des activités traditionnelles fortement ancréesterritorialement, où les pratiques de travail reposent historiquement sur des savoir-faire manuels, untransmission des gestes et des formes d’organisation de petite taille. Pourtant, ces dernières années,l’adoption croissante des technologies numériques interroge profondément les modalités deproduction, de coopération et de création au sein de ces organisations. Cette communications’inscrit dans une approche critique de la digitalisation et vise à répondre à la problématiquesuivante : en quoi l’adoption du numérique transforme-t-elle l’activité des artisans d’art ?Mobilisant une perspective territorialisée, cette recherche s’appuie sur une étude qualitative menéedans le village de Pernes-les-Fontaines, reconnu pour la densité de ses artistes et artisans d’art. Ledispositif méthodologique repose sur une observation participante de longue durée et sur desentretiens semi-directifs conduits auprès d’artistes et d’artisans d’art, permettant de saisir lespratiques effectives, les représentations du numérique et les tensions vécues au quotidien
Scanner spectral à comptage photonique
No full textComputed tomography imaging is based on the production of X-rays, their interaction with biological tissues, and their detection in order to generate tomographic images used in clinical practice. Conventional CT scanners, followed by spectral imaging, have enabled significant advances, particularly by improving tissue characterization and material differentiation through energy analysis ; however, these technologies remain limited by energy-integrating detectors defaults : electronic noise, spectral information overlap, and important dose requirement. In response to these limitations, photon-counting CT has emerged as a major technological evolution, relying on the direct and individual detection of X-ray photons with precise energy measurement, enabling truly multi-energy imaging. This technology provides a significant improvement in spatial resolution, enhanced tissue characterization, a substantial reduction in ionizing radiation dose, and a decrease in the amount of administered contrast medium. It opens important clinical perspectives, particularly in pediatric radiology, digestive oncology, and cardiovascular imaging, while integrating into adapted working environments and workflows through advanced reconstruction tools. Despite medico-economic, regulatory, and technical constraints related to its installation and complexity, photon-counting CT represents a major step toward more precise and safer medical imaging.L’imagerie scanner repose sur la production de rayons X, leur interaction avec les tissus biologiques et leur détection afin de former des images tomodensitométriques utilisées en pratique clinique. Les scanners conventionnels, puis l’imagerie spectrale, ont permis des avancées notables en améliorant la caractérisation tissulaire et la différenciation des matériaux grâce à l’analyse énergétique, mais ces technologies restent limitées par les détecteurs à intégration d’énergie qui entraînent un bruit électronique, une superposition des informations spectrales et une forte exposition des patients. Face à ces limites, le scanner à comptage photonique s’impose comme une évolution majeure, reposant sur la détection directe et individuelle des photons X avec une mesure précise de leur énergie, permettant une imagerie véritablement multi-énergétique. Cette technologie offre une amélioration significative de la résolution spatiale, une meilleure caractérisation des tissus, une réduction importante de la dose de rayonnements ionisants et une diminution du volume de produit de contraste administré. Elle ouvre des perspectives cliniques importantes, en particulier en radio-pédiatrie, en cancérologie digestive et en imagerie cardiovasculaire, tout en s’intégrant dans des environnements de travail et des workflows adaptés grâce à des outils de reconstruction avancés. Malgré des contraintes médico-économiques, réglementaires et techniques liées à son installation et à sa complexité, le scanner à comptage photonique représente une avancée vers une imagerie plus précise et plus sûre
Predicting from Nilometer Data in Ancient Egypt: A Source of Interest for Statistical Learning in Sustainability Governance
No full textWith the emergence of artificial intelligence in sustainability governance, statistical and machine learning algorithms play an important role in prediction and forecasting tasks, whether in environmental, social and economic dimensions. These approaches are often seen as very promising but also raise concerns regarding trust, ethics, and the role of humans in these models. In order to provide a complementary perspective face to these questions, a relevant and important approach is to look at the past in order to understand how older similar instruments were used, their benefits and the aspects that needed to be improved. While the field of statistical and machine learning algorithms is quite recent, the paradigm of prediction based on observed data existed long before the emergence of statistical learning theory and certain data-based predictive instruments like Nilometers in ancient Egypt have already played a major role in sustainable governance for millennia. In order to provide a complementary reflection regarding these two aspects, and based on several different sources whether in the fields of archaeology, history of mathematics, accounting or geography, the objective of this paper is to present a survey gathering the recovered pieces of information regarding Nilometers in ancient Egypt and their utility as data-based predictive instruments. This paper aims moreover to highlight the reflection that Nilometers can be viewed as an ancient root of predictive models from data (including in statistical learning) and that their longevity and sustainable engineering remain sources of interest for the field of sustainability governance
Chemical fate in vitro: A physiological biokinetic (PBK) model for cell-based assays
No full textInternational audienceIn vitro toxicology assays are widely used in the context of toxicology to study in vivo outcomes and to improve mechanistic understanding of toxic responses. For this purpose, a better characterization of actual exposure and the fate of chemicals within the cellular test system is required. The present study aimed to develop a novel model, INSIGHT (In Silico Guide for Harmonized in vitro Testing), that integrates physiological and physi-cochemical parameters to better describe chemical fate in vitro and to guide assay design. The newly devel-oped model, integrating the dynamic features of the Virtual Cell Based Assay model with the partitioning framework of the Virtual In Vitro Distribution model, was calibrated using both a large literature dataset and original experimental data, comprising a total of 42 chemicals and 7 commonly used cell lines: HepaRG, HepG2, 3T3 Balb/c, PC12, MCF-7, RTgill-W1, and HEK293. These cell lines were selected for their diverse tissue and species origins, metabolic capacities, and potential for functional transport mechanisms, known to influence chemical kinetics. The INSIGHT model demonstrated flexibility and robustness across a range of cell lines when the parameters driving their metabolic activity or functional transport were informed. The present study underscores the pivotal function of logPow determination and the necessity for accurate calibration of partition coefficients, permeability, and metabolic processes to account for variability across cell lines and tested chemicals. This approach supports and facilitates enhanced experimental design and advancing quan-titative in vitro-to-in vivo extrapolation (qIVIVE) in toxicology and strengthens next generation risk assessment (NGRA) workflows
Virtual Element methods for non-Newtonian shear-thickening fluid flow problems
No full textIn this work, we present a comprehensive theoretical analysis for Virtual Element discretizations of incompressible non-Newtonian flows governed by the Carreau-Yasuda constitutive law, in the shear-thickening regime ( r > 2) including both degenerate (delta = 0) and non-degenerate (delta > 0) cases. The proposed Virtual Element method features two distinguishing advantages: the construction of an exactly divergence-free discrete velocity field and compatibility with general polygonal meshes. The analysis presented in this work extends a previous work, where only shear-thinning behavior (1 < r < 2) was considered. Indeed, the theoretical analysis of the shear-thickening setting requires several novel analytical tools, including: an inf-sup stability analysis of the discrete velocity-pressure coupling in non-Hilbertian norms, a stabilization term specifically designed to address the nonlinear structure as the exponent r > 2; and the introduction of a suitable discrete norm tailored to the underlying nonlinear constitutive relation. Numerical results demonstrate the practical performance of the proposed formulation
Embedding Birth-Death Processes within a Dynamic Stochastic Block Model
No full textStatistical clustering in dynamic networks aims to identify groups of nodes with similar or distinct internal connectivity patterns as the network evolves over time. While early research primarily focused on static Stochastic Block Models (SBMs), recent advancements have extended these models to handle dynamic and weighted networks, allowing for a more accurate representation of temporal variations in structure. Additional developments have introduced methods for detecting structural changes, such as shifts in community membership. However, limited attention has been paid to dynamic networks with variable population sizes, where nodes may enter or exit the network. To address this gap, we propose an extension of dynamic SBMs (dSBMs) that incorporates a birth-death process, enabling the statistical clustering of nodes in dynamic networks with evolving population sizes. This work makes three main contributions: (1) the introduction of a novel model for dSBMs with birth-death processes, (2) a framework for parameter inference and prediction of latent communities in this model, and (3) the development of an adapted Variational Expectation-Maximization (VEM) algorithm for efficient inference within this extended framework
Analyse des méthodes de gestion des risques des fabricants de DM et proposition d’un outil d’aide à l’application du guide FD CEN ISO/TR 24971
No full textRisk management is a central element of medical device safety and regulatory compliance. It is governed by standard NF EN ISO 14971:2019 and supplemented by application guide FD CEN ISO/TR 24971, which aims to facilitate its operational implementation.The aim of this project is to analyse the actual risk management practices of medical device manufacturers and to identify the limitations encountered in the application of the standard requirements. A field survey, conducted in the form of a questionnaire among manufacturers of various sizes, assessed the level of mastery of the standard, the risk analysis methods used and the use of the FD CEN ISO/TR 24971 guide.The results show that the ISO 14971 standard is generally applied satisfactorily, but that its application guide is still only used to a limited extent. The most commonly used methods remain FMEA, criticality matrices and internal checklists, which are mainly based on qualitative approaches. Respondents expressed a need for educational and operational tools to harmonise practices and improve understanding of regulatory requirements.Based on these findings, this work proposes the development of an interactive tool to assist in the application of the FD CEN ISO/TR 24971 guide, designed to support manufacturers in structuring and improving their risk management processes.La gestion des risques est un élément central de la sécurité et de la conformité réglementaire des dispositifs médicaux. Elle est encadrée par la norme NF EN ISO 14971:2019 et complétée par le guide d’application FD CEN ISO/TR 24971, qui vise à faciliter sa mise en œuvre opérationnelle. Ce projet a pour objectif d’analyser les pratiques réelles de gestion des risques chez les fabricants de dispositifs médicaux et d’identifier les limites rencontrées dans l’application des exigences normatives. Une enquête terrain, réalisée sous forme de questionnaire auprès de fabricants de tailles variées, a permis d’évaluer le niveau de maîtrise de la norme, les méthodes d’analyse des risques utilisées ainsi que l’usage du guide FD CEN ISO/TR 24971. Les résultats montrent une application globalement satisfaisante de la norme ISO 14971, mais une utilisation encore partielle de son guide d’application. Les méthodes les plus employées restent l’AMDEC, les matrices de criticité et les check-lists internes, majoritairement basées sur des approches qualitatives. Les répondants expriment un besoin d’outils pédagogiques et opérationnels pour harmoniser les pratiques et améliorer la compréhension des exigences réglementaires. Sur la base de ces constats, ce travail propose le développement d’un outil interactif d’aide à l’application du guide FD CEN ISO/TR 24971, destiné à accompagner les fabricants dans la structuration et l’amélioration de leur processus de gestion des risques
Copper Oxides and Related Materials: Properties, Synthesis, and Applications
No full textInternational audienceCopper Oxides and Related Materials: Properties, Synthesis, and Applications provides a thorough description of the physical, structural, electrophysical, chemical, electronic, and biological properties of copper oxides in one convenient volume. The book focuses on human and environmental aspects, paying particular attention to health, environmental remediation, electronic applications, energy storage, and solar energy conversion.Copper oxides are widely used in many applications such as antibacterial, antimicrobial, and anticancer treatments; as biocide agents against phytopathogens or hospital-acquired diseases; and, they are also a standard material used for superconductivity. This volume thoroughly explores the most exciting and relevant applications of copper oxides, a class of material with a sometimes bewilderingly wide array of potential uses