HAL Portal de Univ. Gustave Eiffel
Not a member yet
72847 research outputs found
Sort by
Contrôle périmétrique avec des taux hétérogènes de franchissement du cordon : une approche basée sur l'apprentissage par renforcement et régularisé par la physique
International audiencePerimeter Control (PC) strategies have been proposed to address urban road network control in oversaturated situations by regulating the transfer flow of the Protected Network (PN) based on the Macroscopic Fundamental Diagram (MFD). The uniform metering rate for cordon signals in most existing studies overlooks the variance of local traffic states at the intersection level, which may cause severe local traffic congestion and degradation of the network stability. PC strategies with heterogeneous metering rates for cordon signals allow precise control for the perimeter but the complexity of the problem increases exponentially with the scale of the PN. This paper leverages a Multi-Agent Reinforcement Learning (MARL)-based traffic signal control framework to decompose this PC problem, which considers heterogeneous metering rates for cordon signals, into multi-agent cooperation tasks. Each agent controls an individual signal located in the cordon, decreasing the dimension of action space for the controller compared to centralized methods. A physics regularization approach for the MARL framework is proposed to ensure the distributed cordon signal controllers are aware of the global network state by encoding MFD-based knowledge into the action-value functions of the local agents. The proposed PC strategy detects the overall traffic state within the PN and distributes local instructions to cordon signal controllers in the MARL framework via the physics regularization. Through numerical tests in a microscopic traffic environment, the proposed PC strategy shows promising robustness on different demand patterns and transferability by inheriting the trained models to several distinct signals without retraining. It outperforms state-of-the-art feedback PC strategies and centrliazed RL methods in increasing network throughput, decreasing distributed delay for gate links, and reducing carbon emissions
“The Chambat Sardine Can” technique for the treatment of chronic quadriceps tendon rupture
International audienceRuptures of the quadriceps tendon (QT) are rare but serious injuries accounting for less than 2% of all tendon injuries around the knee. These injuries, often occurring in individuals over 40, are leading to a loss of active extension and a significant impact on knee function. While the treatment of acute QT ruptures through various reinsertion techniques has shown excellent outcomes, managing chronic injuries and failed primary repairs remains challenging due to tendon retraction and difficulties in repositioning the tendon stump. This study introduces a novel approach associating direct tendon reinsertion with metal frame reinforcement, aiming to effectively lower the retracted tendon to the proximal pole of the patella. This technique offers a promising alternative that addresses the limitations of traditional methods and potentially improves patient outcomes by providing a safe primary fixation and protection of the repair, enabling early rehabilitation and reducing the need for subsequent interventions.Level of evidenceIV; case series study
L'efficacité des stratégies de gamification et de personnalisation dans la phase d'entrée en relation en ligne avec un prospect
International audienceVery little research has focused on the initial contact with a prospect. Yet this first encounter between a brand and a prospect is crucial to the consumer's long-term commitment to the relationship. A between-subject experiment tested the effectiveness of two levers of entering into an online relationship with a prospect - personalisation and gamification - in encouraging the prospect's attitude towards the brand and his/her intention to continue the relationship with the brand. Compared to the "classic" version of the website, personalisation and gamification significantly increased the entertainment and relevance associated with this first contact with the brand. They also led to an increased level of perceived intrusiveness, without any impact on attitude towards the brand. Finally, personalisation proved to be a more effective technique than gamification for increasing commitment towards the brand.Peu de recherches se sont intéressées à la phase d’entrée en relation avec un prospect. Pour autant, cette première rencontre entre une marque et un prospect est déterminante pour engager durablement le consommateur dans la relation. Une expérimentation en inter-sujets teste l’efficacité de deux leviers d’entrée en relation en ligne avec le prospect– la personnalisation et la gamification – en vue de favoriser l’attitude du prospect à l’égard de la marque et sa volonté de poursuivre la relation avec la marque. Par rapport à la version « classique » du site web, la personnalisation comme la gamification augmentent de manière significative le divertissement et la pertinence associés à ce premier contact avec la marque. Elles conduisent également à un niveau accru d’intrusivité perçue, sans que celle-ci n’ait d’impact sur l’attitude à l’égard de la marque. Enfin, la personnalisation se révèle être une technique plus efficace que la gamification pour accroître l’engagement relationnel sur le long terme
« Des horizons proches et lointains : une exploration des représentations sanatoriales au prisme de la linguistique »
International audienceExpression of an architectural, geographical and medical thinking, in which the contemplation of the landscape is regarded as a therapeutic means (Luxembourg et al. 2023), the sanatorium occupies a distinctive position in the collective imagination. This exploratory article seeks to delineate the contours of this place by examining the contemporary representations that extend beyond its mere physical or geographical reality. Through a linguistic lens, our research addresses the following questions: Which terms belong to the semantic field of the sanatorium, and what proximity (or distance) do they have with our research object? How do these “keywords” contribute to shaping our understanding of sanatorium representations?We begin with a few etymological references to establish the general meaning of the concept of sanatorium, before exploring its semantic proximity by examining words that share similar or related meanings. Then, we conduct a detailed analysis of a corpus of scientific and journalistic articles published between 1991 and 2021, focusing on the most recurrent lexical forms that evoke contemporary representations of the sanatorium. Through this study, we aim to characterize specific facets of the sanatorium’s identity and trace its evolution over time.Expression d’une pensée architecturale, géographique et médicale, dans laquelle la contemplation du paysage est envisagée comme un moyen thérapeutique (Luxembourg et al. 2023), le sanatorium occupe une place singulière dans l’imaginaire collectif. Cet article exploratoire vise à esquisser les contours de ce lieu, en explorant les représentations contemporaines qui lui sont associées, et qui dépassent sa simple réalité physique ou géographique. Au prisme de la linguistique, notre recherche vise plus particulièrement à répondre aux questions suivantes : quels vocables font partie du champ sémantique du sanatorium, et quels liens de proximité (ou de distance), entretiennent-ils avec notre objet de recherche ? En quoi ces « mots-témoins » contribuent-ils à éclairer la compréhension des représentations sanatoriales ?Nous introduisons premièrement quelques repères étymologiques pour définir le sens général du concept de sanatorium, avant de prêter attention à sa proximité sémantique, via l’examen de certains mots qui lui sont plus ou moins proches du point de vue du sens. Ensuite, nous procédons à l’analyse plus détaillée d’un corpus d’articles scientifiques et journalistiques publiés entre 1991 et 2021, en nous penchant sur les formes lexicales les plus récurrentes qui évoquent les représentations sanatoriales contemporaines. Notre étude cherche ainsi à caractériser certaines facettes de l’identité du sanatorium et à documenter son évolution au cours du temps
Tiers-lieux dans la recherche : une analyse généalogique et scientométrique
International audienceThis article compares the notion of third place as it has institutionnalized in France with concepts mobilized in the international scientific literature to characterize places for open innovation (living labs, fablabs, and so on). The comparison is based on two successive approaches: a genealogy of the uses of the notion of third places in France and a scientometric study of the international scientific literature. We define the notion of "French-style third places" and conclude that they are part of an ecosystem of open laboratories for social innovation, at the intersection of living labs, fab labs, makerspaces, etc. Based on this, we provide programmatic recommendationsto better understand their role in transformative research and innovation processes.En quoi les tiers-lieux se différencient-ils d’autres pratiques d’innovation conceptualisées dans la littérature scientifique ? Quel est leur rôle dans les interactions entre sciences et sociétés, et secondairement, dans les transitions environnementales et sociales ? Cet article propose de comparer la notion de tiers-lieux telle qu’elle a émergé en France, avec des concepts issus de la littérature scientifique internationale. La comparaison s’appuie sur deux approches successives : une généalogie des usages de la notion de tiers-lieux en France et une étude scientométrique de la littérature scientifique internationale. Nous en concluons que les tiers-lieux « à la française » font partie d’un écosystème de laboratoires ouverts d’innovation sociale et de recherche, à la croisée des livings-labs/fablabs/makerspaces etc. Nous en tirons des recommandations pour l’orientation des recherches afin de mieux comprendre le rôle des tiers-lieux dans les processus de recherche et d’innovation à visée transformative
Automated mapping between SDG indicators and open data: An LLM-augmented knowledge graph approach
International audienceMeeting the Sustainable Development Goals (SDGs) presents a large-scale challenge for all countries. SDGs established by the United Nations provide a comprehensive framework for addressing global issues. To monitor progress towards these goals, we need to develop key performance indicators and integrate and analyze heterogeneous datasets. The definition of these indicators requires the use of existing data and metadata. However, the diversity of data sources and formats raises major issues in terms of structuring and integration. Despite the abundance of open data and metadata, its exploitation remains limited, leaving untapped potential for guiding urban policies towards sustainability. Thus, this paper introduces a novel approach for SDG indicator computation, leveraging the capabilities of Large Language Models (LLMs) and Knowledge Graphs (KGs). We propose a method that combines rule-based filtering with LLM-powered schema mapping to establish semantic correspondences between diverse data sources and SDG indicators, including disaggregation. Our approach integrates these mappings into a KG, which enables indicator computation by querying graph’s topology. We evaluate our method through a case study focusing on the SDG Indicator 11.7.1 about accessibility of public open spaces. Our experimental results show significant improvements in accuracy, precision, recall, and F1-score compared to traditional schema mapping techniques
A comprehensive methodology to assess human bone transversal toughness based on macroscopic specimens, the compliance method, and 3D bio-faithful numerical simulations
International audienceThis study proposes a method for assessing the transverse toughness of human long-bone cortical tissue. The method is based on a three-point bending test of pre-notched femur diaphysis segments, post-processed using the compliance method coupled with numerical simulations. Given the cracking nature of bone and if cracking processes remain confined to the crack tip, it is assumed that the compliance method can be used. Numerical simulations are based on a bio-faithful 3D reconstruction of the bones tested and a detailed consideration of the boundary and loading conditions of the mechanical test. The resulting toughness values obtained on embalmed bones range from Gc=4.3 to 7.1 N/mm. The assumptions made, the biofidelity of the simulations, and the ability of the method to determine an intrinsic toughness value of cortical bone, considered a heterogeneous material, are discussed. Although related to embalmed bones, and considering the limitations this state can induce, the toughness values obtained are consistent with data from the literature. Due to the larger specimen size, they are also more realistic, ensuring a complete description of the material’s crack extension resistance curve. They mainly characterize the medial and lateral quadrants of the bone transversal section. The study concludes that the proposed method provides a robust approach for assessing bone transversal toughness
No Annotations for Object Detection in Art through Stable Diffusion
8 pages, 6 figures, to be published in WACV 2025International audienceObject detection in art is a valuable tool for the digital humanities, as it allows for faster identification of objects in artistic and historical images compared to humans. However, annotating such images poses significant challenges due to the need for specialized domain expertise. We present NADA (no annotations for detection in art), a pipeline that leverages diffusion models' art-related knowledge for object detection in paintings without the need for full bounding box supervision. Our method, which supports both weakly-supervised and zero-shot scenarios and does not require any fine-tuning of its pretrained components, consists of a class proposer based on large vision-language models and a class-conditioned detector based on Stable Diffusion. NADA is evaluated on two artwork datasets, ArtDL 2.0 and IconArt, outperforming prior work in weakly-supervised detection, while being the first work for zero-shot object detection in art. Code is available at https://github.com/patrick-john-ramos/nad
QUANTITATIVE OBSERVABILITY FOR ONE-DIMENSIONAL SCHR ÖDINGER EQUATIONS WITH POTENTIALS
International audienceIn this note, we prove the quantitative observability with an explicit control cost for the 1D Schrödinger equation over R with real-valued, bounded continuous potential on thick sets. Our proof relies on different techniques for low-frequency and high-frequency estimates. In particular, we extend the large time observability result for the 1D free Schrödinger equation in Theorem 1.1 of Huang-Wang-Wang [20] to any short time. As another byproduct, we extend the spectral inequality of Lebeau-Moyano [27] for real-analytic potentials to bounded continuous potentials in the one-dimensional case