Portail des publications scientifiques IMT Mines Alès
Not a member yet
5198 research outputs found
Sort by
Using novel ion-imprinting acrylamide/vinylphosphonic acid-based sorbent to improve thorium recovery - Application on the environmentally acidic solution
No full textInternational audienc
Environmental challenges and perspectives in the development of nanocomposites for enhanced flame-retardant properties (chap. 12)
No full textInternational audienc
Unravelling individual rhythmic abilities using machine learning
No full textInternational audienceHumans can easily extract the rhythm of a complex sound, like music, and move to its regular beat, for example in dance. These abilities are modulated by musical training and vary significantly in untrained individuals. The causes of this variability are multidimensional and typically hard to grasp with single tasks. To date we lack a comprehensive model capturing the rhythmic fingerprints of both musicians and non-musicians. Here we harnessed machine learning to extract a parsimonious model of rhythmic abilities, based on the behavioral testing (with perceptual and motor tasks) of individuals with and without formal musical training ( n = 79). We demonstrate that the variability of rhythmic abilities, and their link with formal and informal music experience, can be successfully captured by profiles including a minimal set of behavioral measures. These profiles can shed light on individual variability in healthy and clinical populations, and provide guidelines for personalizing rhythm-based interventions
Fire behavior of biobased concretes
No full textInternational audienceNowadays, the construction field is responsible for a quarter of french emissions of greenhouse gases and it represents the economic sector that consumes the most energy in France. A part of the solution is to promote the use of local and biobased materials with a low carbon impact. These biobased resources can be mixed with a binder to produce light (150 – 500 kg/m3), medium (500 – 1200 kg/m3), or high density (1200 – 1800 kg/m3) biobased concretes. The fire performance of these materials is poorly documented and therefore deserves further investigations. Consequently, it is hard for building professionals to demonstrate that building fire regulations can be fulfilled when using these materials. Thus, in order to study the fire behavior of these biobased concretes, at first, the flammability of the biobased resources was studied through tests carried out in the Pyrolysis Combustion Flow Calorimetry (PCFC), under aerobic and anaerobic conditions, and in the bomb calorimeter, to quantify their gross heat of combustion (GHC). Then, cone calorimeter experiments were carried out to study the ignition occurrence of biobased concretes. In the present study, more than 150 different formulations were studied. Results highlight that only the lightest biobased concretes ignite and, more precisely, it appears that fire ignition criterion relies on two parameters: the density of energy which bioconcrete can release and the energy required to heat the material surface, taking into account endothermic processes. Finally, a model to predict the ignition of bioconcretes was propose
Ontologie de Maintenance des Bâtiments et Capacités des Larges Modèles de Langage (LLM) pour le Peuplement
No full textInternational audienceLes données de maintenance des bâtiments proviennent de diverses sources, notamment de prestataires de services tels que les ascensoristes, les chauffagistes, ou des professionnels multiservices, ainsi que des clients pouvant être des gestionnaires immobiliers, des villes ou des acteurs du secteur tertiaire. La nature hétérogène de ces données, en raison de la diversité des sources, complique le processus de partage des données. Cet article propose une ontologie de domaine pour représenter ces données, explorant l’utilisation des LLMs pour peupler automatiquement l’ontologie. Les résultats indiquent une bonne performance de ChatGPT et TextCortex dans la génération d’instances à partir de données CSV semi-structurées. Cette approche vise à améliorer l’efficacité du peuplement de l’ontologie malgré la diversité des données de maintenance
Meaning in life and Parkinson’s disease in the UK Biobank
No full textInternational audienceHighlights•Meaning in life is an aspect of eudaimonic well-being associated with brain health.•In the large UK Biobank, meaning in life is associated with lower risk of incident PD over five years.•The association was independent of demographic, behavioral, clinical, and genetic risk factors.•The association generalized across age, sex, education, deprivation, and genetic risk.Introduction: Meaning in life is an aspect of eudaimonic well-being associated with lower dementia risk. This research examines whether this protective association extends to Parkinson’s disease (PD).Methods: Participants (N = 153,569) from the UK Biobank reported on their meaning in life. Cases of PD were identified through health records.Results: Meaning in life was associated with a 50 % lower likelihood of prevalent PD (OR = 0.68, 95 % CI = 0.59–0.78). Over the 5-year follow-up, meaning was associated with a 35 % lower risk of incident PD (HR = 0.74, 95 % CI = 0.65–0.83), an association robust to sociodemographic characteristics, depression, history of seeking mental health care, smoking, physical activity, and genetic risk and not moderated by age, sex, education, deprivation, or genetic risk.Conclusions: Meaning in life is associated with lower risk of incident PD, an association independent of other major risk factors and generalizable across sociodemographic groups. Meaning is a promising target of intervention for common neurodegenerative diseases
Définition d'un mode d'analyse et d'une méthodologie outillée de modélisation de la gestion des données / informations / connaissances
No full textCritical Infrastructure (CI) engineering projects are characterised by the high complexity of the systems of interest. They are often lengthy projects with major budgets, involving many stakeholders from different disciplines, professions and business sectors. As a result, these projects are now being managed according to the principles and processes of systems engineering (SE), with the aim of transitioning from document-based to model-based engineering based to Model-Based Systems Engineering (MBSE) approach. This is particularly the case today in the context of the engineering of Basic Nuclear Facilities (BNF), where regulations, constraints and requirements to be integrated are evolving rapidly. Similarly, managing, accessing and even simply using the data, information and knowledge (DIK) generated and/or manipulated throughout these projects is becoming more complex. This increase in complexity results, among other things, from the volume, variety, variability, speed of evolution and even veracity of this DIK, which hampers or even prevents its proper use, sharing and traceability. These projects therefore require sophisticated DIK management adapted to the context. The aim is to facilitate and improve the exchange and sharing of DIK between stakeholders, at the right time, according to their roles and responsibilities and, more generally, according to their specific needs depending on the activity they have to carry out. To address these issues, this thesis aims to establish a link between the operational domain of Model-Based Systems Engineering and that of Data Science in order to define a system for searching, recommending and evaluating data, information and knowledge to support the stakeholders involved in such projects. The resulting contribution is a tool-based method called GENIUS-CIME (Guided ENgineering by an Information User-centered recommendation System, developed as part of the CIME Critical Infrastructures Model-based systems Engineering Chair). This method studies and integrates the characteristics of IS processes and provides tools and approaches for recommending data, information and knowledge to project stakeholders by promoting the use of ontologies.Les projets d’ingénierie d’Infrastructures Critiques (IC) sont des projets caractérisés par une haute complexité des systèmes d’intérêts. Ce sont des projets longs, à budgets conséquents, qui impliquent de nombreuses parties prenantes de différents domaines disciplinaires, métiers et secteurs d’activité. De fait, ces projets sont aujourd’hui gérés en suivant les principes et les processus de l’ingénierie système (IS), visant à basculer d’une ingénierie basée sur documents à une ingénierie basée modèles portée par l’approche d’Ingénierie système Basée Modèles (ISBM). C’est aujourd’hui particulièrement le cas dans le contexte de l’ingénierie d’Installations Nucléaires de Base (INB) où les réglementations, les contraintes et les exigences à intégrer évoluent rapidement. De même, la gestion, l’accès, ou encore l’utilisation pure et simple des données, informations et connaissances (DIC) générées et/ou manipulées tout au long de ces projets deviennent plus complexes. Cette complexification résulte, entre autres, des caractéristiques de volume, de variété, de variabilité, de vitesse d’évolution et même de véracité de ces DIC qui gênent voire entravent leur bonne utilisation, leur partage et leur traçabilité. Ces projets requièrent donc une gestion des DIC élaborée et adaptée au contexte. Cela vise à faciliter et améliorer leur échange et partage entre les acteurs métier, au bon moment, selon leurs rôles et responsabilités et, plus globalement, selon leurs besoins ponctuels en fonction de l’activité qu’ils doivent mener à bien. Pour pallier ces enjeux, ces travaux de thèse visent à établir un lien entre le domaine opérationnel de l’Ingénierie Système Basée Modèle et celui des Sciences des Données pour définir un système de recherche, recommandation et évaluation de données, informations et connaissances venant en support des acteurs métier impliqués dans de tels projets. La contribution résultante est une méthode outillée baptisée GENIUS-CIME (Guided ENgineering by an Information User-centered recommendation System, développée dans le cadre de la Chaire CIME Critical Infrastructures Model-based systems Engineering). Cette méthode étudie et intègre les caractéristiques des processus d’IS et fournit les outils et démarches pour la recommandation de données, informations et connaissances aux parties prenantes d’un projet en promouvant l’utilisation des ontologies