HAL Portal Cnam (French National Conservatory for Arts and Crafts)
Not a member yet
31010 research outputs found
Sort by
Contrôle en boucle fermée robuste aux perturbations d'amplitude finie : application à des modèles réduits de la transition sous-critique
No full textInternational audienc
Deep Active Learning-driven Acceleration of Smart Grids Security Assessment
No full textInternational audienceReliable and efficient operation of smart grids (SGs) is contingent on fast and accurate security assessment, which is increasingly difficult because of the growing complexity and uncertainty of modern power grids. As traditional power flow (PF) simulations are computationally intensive, machine learning (ML)-based approaches have gained traction. However, such models typically require large amounts of labelled data, which is time-consuming to acquire via PF oracles. To mitigate this, we propose a deep active learning (DAL)-driven framework that actively selects the most informative operational points (OPs) for labelling, thereby reducing reliance on exhaustive PF computations. We evaluate multiple DAL query strategies—including Monte Carlo (MC) dropout and batch active learning by diverse gradient embeddings (BADGE)—on a binary classification task to detect congestion in the IEEE European low voltage test network (ELVTN). Results show that DAL methods significantly reduce the number of training labelled samples required over the random baseline on the considered SG security assessment dataset. Our findings suggest that DAL is a promising avenue for accelerating the training of ML-models in SGs by reducing dependence on costly PF-based labelling. The code used to reproduce the results presented in this paper is publicly available at https://gitlab.com/satie.sete/dal_accel_sg_pf
SDG-KG: A Framework to Compute SDG Indicators with Open Data
No full textInternational audienceMonitoring Sustainable Development Goal (SDG) indicators requires integrating heterogeneous open datasets from sources such as relational databases, NoSQL stores, and APIs. While SDG indicators follow standardized definitions, open data sources are often fragmented, schema-less, and inconsistent, making both integration and computation challenging. In this demonstration, we present SDG-KG , a spatio-temporal Knowledge Graph (KG) framework designed to structure metadata, guide data retrieval, and formalize indicator computation workflows. Our approach leverages graph-based modeling to construct a Metadata Graph, apply conflict resolution techniques when multiple sources provide overlapping data, and dynamically generate query-driven execution plans. Through an interactive interface, users can explore United Nations specifications, inspect data provenance and the generated KG, and visualize the computed indicators
Comptes-rendus de la conférence CIE France Midterm 2025
No full textNational audienceCette conférence rassemble une sélection des présentations des laboratoires française à la conférence Midterm de la Commission Internationale de l'Eclairage. Les exposés sont en français et sont à destination de la communauté de l'éclairage en France
Supporting interprofessional collaboration: boundary work carried out by nurse managers in French public hospital
No full textInternational audienc
Archéologie des pratiques astronomiques urbaines
No full textInternational audienceCette communication présente un projet de recherche-création portant sur l’histoire des pratiques astronomiques urbaines aux XIXe et XXe siècles. Dans ce cadre, nous nous intéressons plus particulièrement à l’usage de sites situés en hauteur dans le paysage urbain : toits, tours, terrasses… Notre travail se caractérise également par une importante dimension participative, avec la contribution de membres de la commission « Histoire » de la Société astronomique de France, d’un groupe d’auditeurs.rices du Cnam suivant un parcours de formation à la médiation culturelle des sciences et techniques, ainsi que du photographe Alain Cornu qui photographie les toits de Paris
Categorical Data Encoding: from H.O. Hirschfeld to Machine Learning
No full textInternational audienceThe need for Machine Learning algorithms to convert qualitative data into numerical data has led to many (and sometimes incongruous) proposals or to the rediscovery of work by statisticians dating back almost a century. The optimal coding methods developed by Hirschfeld (aka H.O. Hartley), R.A.Fisher, L.Guttmann, C.Hayashi and many others are at the origin of correspondence analysis. The links with the search for transformations to normal distributions were established by H.O.Lancaster, M.Kendall and A. Stuart. This paper starting with pioneering work in the 1930s, will provide an overview of the various approaches up to the current revival inspired by the need to process high-dimensional categorical data
Designing clinical trials for the comparison of single and multiple quantiles with right-censored data
No full textBased on the test for equality of quantiles originally introduced by Kosorok (1999), we propose new power formulas for the comparison of one quantile between two treatment groups, as well as for the comparison of a collection of quantiles. Under the null hypothesis of equality of quantiles, the test statistic follows asymptotically a normal distribution in the univariate case and a chi-squared with J degrees of freedom in the multivariate case, with J the number of quantiles compared. The variance of the test statistic depends on the estimation of the probability density function of the distribution of failure times at the quantile being tested. In order to apply the test on real data, we propose to estimate this quantity using a resampling-based method, as an alternative to Kosorok's original kernel density estimator. The whole procedure provides a practical tool for designing and analyzing data arising from clinical trials using quantiles of survival as an endpoint. Simulation studies are performed to show the appropriateness of the power formulas. We illustrate the proposed test in a phase III randomized clinical trial where the proportional hazards assumption between treatment arms does not hold
Association between dietary environmental pressures and major chronic diseases: assessment from the prospective NutriNet-Santé cohort
No full textInternational audienceBackgroundPlant-based diets offer co-benefits for human health and the environment, but assessments often consider only specific aspects. This study comprehensively examines the links between diet-related environmental pressures and risk of chronic diseases as well as mortality.MethodsData from a population study of 34,077 participants to the NutriNet-Santé French cohort were used. Dietary data were collected using a food frequency questionnaire, distinguishing between organic and conventional foods, and were merged with food production environmental indicators. The associations between greenhouse gas emissions (GHGe), energy demand, land occupation (LO), ecological infrastructures (EI), water use, and pesticide treatment frequency and a synthetic environmental pressures index (EPI) and incidence of cancer, cardiovascular diseases (overall, coronary and cerebrovascular diseases), type 2 diabetes and mortality were estimated using weighted multivariable cox proportional risk model.FindingsOver a mean median follow-up of 8.39 years (IQR = 5.62, 256,891 person-year), the diet's overall environmental pressures (EPI) was positively associated with the risk of all tested chronic diseases except stroke. The HR for 1 SD increment ranging from 1.15 (95% CI = 1.03–1.28) for cancer (all locations) to 1.50 (95% CI = 1.29–1.73) for coronary heart disease and type 2 diabetes, but no association with stroke or death was detected.InterpretationDiets with low overall environmental pressures are associated with important health benefits, suggesting that food systems with lower environmental impacts could be key drivers of both environmental and health sustainability.FundingData were collected in the context of the BioNutriNet and TRANSFood projects supported by the French National Research Agency (ANR-13-ALID-0001 and ANR-21-CE21-0011-01)
Comportement homogénéisé d'un microvia pour la modélisation mécanique d'un circuit imprimé avec composants enfouis.
No full textInternational audienceThe boom of the e-technology is leading to new designs for printed circuit boards in response to theincrease in component density. The solution found is to integrate active components into the unoccupiedvolume of the printed circuit board. To estimate the printed circuit board life time, the response of theinterconnection layer to thermo-mechanical loads is studied. In fact, this area is prone to various failures.As this kind of design with microvias interconnected chips become usual in industry, the large number ofmicrovias present can be managed by a finite element software. To overcome this problem, a numericalhomogenization is implemented to replace the microvia with its homogeneous equivalent material inlarge printed circuit board structure. This work also highlights the accuracy of the average response ofthe homogeneous equivalent material but also the necessity to keep at least a fully detailed unit cell to analyze the local stresses. This will enable the detection of crack initiation zones in microvia interconnect layers.L’essor des technologies du numérique conduit à la restructuration des circuits imprimés dans le butd’enfouir des composants actifs pour répondre à la densification de ceux-ci. Afin d’estimer la durée devie de ces circuits imprimés, la réponse des couches d’interconnexion des composants enfouis après sollicitations mécaniques est étudiée. En effet, ces zones sont connues pour être propices à des défaillancesde différentes catégories. Alors que dans l’industrie, ce type d’architecture avec des puces interconnectées par microvias se développent, le très grand nombre de microvias en présence est difficilementgérable par calculs par éléments finis. Pour passer outre ce problème, une méthode d’homogénéisation est mise en place afin de remplacer le microvia par son matériau homogène équivalent dans desstructures de circuit imprimé à grande échelle. Le travail réalisé dans cet article met en évidence la précision de la réponse moyenne du matériau homogène équivalent mais aussi, la nécessité de conserverau moins un motif parfaitement détaillé permettant ainsi l’analyse locale des efforts et la détection deszones d’initiation de fissures