HAL Paris Dauphine-PSL
Not a member yet
21932 research outputs found
Sort by
Comprendre l'ampleur des catastrophes naturelles par l'écoute des réseaux sociaux
No full textInternational audienc
On the Krein-Rutman theorem and beyond
No full textIn this work, we revisit the Krein-Rutman theory for semigroups of positive operators in a Banach lattice framework and we provide some very general, efficient and handy results with constructive estimates about: the existence of a solution to the first eigentriplet problem; the geometry of the principal eigenvalue problem; the asymptotic stability of the first eigenvector with possible constructive rate of convergence. This abstract theory is motivated and illustrated by several examples of differential, integro-differential and integral operators. In particular, we revisit the first eigenvalue problem and the asymptotic stability of the first eigenvector for: some parabolic equations in a bounded domain and in the whole space; some transport equations in a bounded or unbounded domain, including some growth-fragmentation models and some kinetic models; the kinetic Fokker-Planck equation in the torus and in the whole space; some mutation-selection models
Consumptive Work in Coworking: Using Consumption Strategically for Work
No full textInternational audienceConsumption has always been part of the workplace, yet it has traditionally been seen as nonwork—an activity that depletes rather than creates value. In the knowledge and digital economy, however, consumption and work are becoming increasingly intertwined, calling for a relational perspective on consumption’s productive role. We develop this perspective through a four-year ethnography of coworking spaces across Paris and London, supplemented by post-pandemic archival data. We introduce consumptive work as the instrumentalization of consumption activities in the workplace to generate productive value. Consumptive work emerges within a postindustrial societal context where workplace culture is shaped by consumer ideology, leading to 1) customer entitlement in the workplace, 2) consumer desire toward the workplace, and 3) consumer lifestyle aspirations toward work. Consumptive work is characterized by inconspicuousness, boundarilessness, and communal and market exchange. While it can be empowering, it also fosters neo-normative alienation, particularly through performative play and leisure, and the pursuit of productive wellness. Ultimately, consumptive work reinforces evolving consumer desires and aspirations about office work and workplaces. This study advances interdisciplinary research on consumption and consumption ideology in the workplace, workplace alienation, new ways of working, and consumer research connecting work, home, and leisure
Three-dimensional reconstruction and fracture segmentation based on X-ray and computed tomography paired dataset
No full textInternational audienceIn some orthopedic surgeries, the use of three-dimensional (3D) computed tomography (CT) scanning technology is not feasible due to scene limitations, leaving doctors to rely on two-dimensional (2D) X-ray images for real-time diagnosis. However, X-ray images lack 3D information, making accurate diagnosis challenging. Developing an algorithm to convert 2D X-ray images into 3D CT images, while simultaneously combining high-quality 3D reconstruction with precise fracture segmentation, offers a promising solution to the problem. In this study, we propose a novel artificial intelligence (AI)-driven framework named 3D reconstruction and segment anything model (3DRecSAM). The reconstruction image enhancer (RIE) is designed to achieve high-precision 3D reconstruction and provide high-quality feature initialization for fracture segmentation. Meanwhile, the mamba segment anything model (MSAM), based on the segment anything model (SAM) architecture, is developed for accurate fracture segmentation. We introduce a Kolmogorov-Arnold network (KAN)-based attention fusion module (KAF), which facilitates the joint optimization of the RIE reconstruction network and the MSAM segmentation network. Furthermore, the selective scanning mamba with KAN (SKM) is incorporated to enhance feature extraction for both RIE and MSAM. Mamba efficiently captures long-range dependencies and sequential patterns, while KAN's learnable activation functions facilitate adaptive feature fusion and non-linear representation. To train and evaluate 3DRecSAM, we introduce the real X-ray and CT paired dataset (XCPData), which is publicly available on GitHub: https://github.com/YuanGao1201/XCPData
France’s Territorial Frameworks for Public Health Policy
No full textInternational audienceThe aim of this chapter is to retrace the three key periods in the history of the French state that have shaped the fabric of contemporary healthcare territorialization. More specifically, it aims to understand the conditions of their transformation, by examining the ways in which actors are organized, their frames of reference for shaping policies and their resistance to change, which bear the imprint of their past history on today’s territorial divisions
Le règlement européen sur l'IA et au-delà : Quelle régulation de l'IA ?
No full textInternational audienceThe European AI Regulation adopted in June 2024 proposes an original legal framework for AI systems and general-purpose AI models, the implementation of which began in part on 2 February 2025. This complex text is based on a comprehensive approach founded on the risks associated with the use of these services and products.This book analyses two innovative cross-cutting themes proposed by the regulation: the attempt to standardise fundamental rights in order to enable the circulation of safe AI systems within the European Union, and the establishment of regulatory sandboxes to ensure a renewed form of governance for AI regulation in the European Union.The book also examines the impact of the regulation on AI in various key sectors: health, insurance, financial services, asylum, migration and border management.These sectoral analyses are coupled with reflections aimed at understanding the relationship between the AI Regulation and, on the one hand, product safety regulations within the internal market and, on the other hand, digital copyright law.Finally, the book addresses comparative aspects to give readers a better understanding of the specific nature of the European model for regulating AI in relation to the approaches of other countries such as the United Kingdom, Singapore, the United States, Japan, South Korea and Brazil, as well as in relation to the approach of the Council of Europe with its recent framework convention on AI.Le règlement européen sur l’IA adopté en juin 2024 propose un encadrement juridique original des systèmes d’IA et des modèles d’IA à usage général dont la mise en application a partiellement commencé le 2 février 2025. Ce texte complexe repose sur une approche globale fondée sur les risques liés aux usages de ces services et produits.Cet ouvrage analyse deux thématiques transversales novatrices proposées par le règlement : la tentative de standardisation des droits fondamentaux afin de permettre la circulation de systèmes d’IA sûrs au sein de l’Union européenne et l’institution de bacs à sable règlementaire afin d’assurer une forme renouvelée de gouvernance de la régulation de l’IA dans l’ Union européenne.L’ouvrage examine également l’impact du règlement sur l’IA dans différents secteurs clés : la santé, les assurances, les services financiers, l’asile, les migrations et la gestion des frontières.Ces analyses sectorielles sont couplées avec des réflexions visant à comprendre l’articulation du règlement sur l’IA avec , d’une part, les réglementations sur la sécurité des produits au sein du marché intérieur, et d’autre part, le droit d’auteur dans le numérique.L’ouvrage aborde enfin des aspects comparatifs afin de permettre aux lecteurs de mieux comprendre la spécificité du modèle européen d’encadrement de l’IA par rapport aux approaches d’autres États tels que le Royaume-Uni et Singapour, les États-Unis, le Japon, la Corée du Sud et le Brésil, mais aussi par rapport à l’approche du Conseil de l’Europe avec sa récente convention-cadre sur l’IA
A note on Apresjan Hierarchies and Linkage-Based Hierarchical Clustering
No full textThe model of an Apresjan cluster is a formal account of what the intuitive notion of cluster amounts to. The set of all Apresjan clusters forms a hierarchy, called the Apresjan hierarchy. In this note, we identify a general class of linkage functions for which the Hierarchical Agglomerative Clustering (HAC) algorithm generates hierarchies containing the Apresjan hierarchy. We deduce that, if the HAC algorithm is applied with a standard linkage function not based on centroids, then every generated hierarchy contains the Apresjan hierarchy. We also present preliminary results related to the HAC algorithm and arbitrary linkage functions, e.g. the characterization of all subsets belonging to all hierarchies generated by the HAC algorithm when this algorithm is applied to a given dataset and used with an arbitrarily chosen linkage function
Cesser le feu: Crudivorisme, science et contre-connaissance
No full textCesser le feuCrudivorisme, science et contre-connaissance
The archeology of Google Inc. from foundation to IPO 1
No full textInternational audienceGoogle is certainly one of the world's best-known company, but the path that takes a start-up from a Silicon Valley garage to Wall Street remains largely unknown. We still don't know, for example, how much the two founders Sergei Brin and Larry Page initially put into the business. How did they manage to retain more than 50% of the voting rights? What were the financial arrangements between the initial shareholders, and what capital gains did they make on the IPO? Using unpublished documents on the company's articles of incorporation and the composition of its capital between 1998 and 2004, we show how the chances of profit are shared beforehand between the various protagonists: founders, shareholders, managers, employees. The hierarchical position they occupy in relation to each other determines the amount of profit they receive at IPO. The case of Google, while singular in the scale of its technological and financial success, is exemplary of the way American capitalism has functioned since its origins. 1 We would like to warmly thank Olivier Alexandre and Benjamin Loveluck for giving us the opportunity to present a preliminary version of this article in their seminar: "Digital Capitalism and Ideologies" in February 2021. We would also like to extend our heartfelt thanks to Philippe Humeau, start-up founder, who helped us decipher the intricacies of fund-raising, Marion Flécher and Mathilde Krill, PhD students at University Paris Dauphine, for their invaluable bibliographical indications, and of course Dominique Maret for his attentive proof-reading. 2 Alphabet, the company that succeeded in 2014, has a market capitalization of 85 to $190. Over the years, Google's shares have become one of the world's leading market capitalizations 2 ,</div