Scientific Publications of the University of Toulouse II Le Mirail
Not a member yet
92205 research outputs found
Sort by
Les représentations sociales sur les ateliers de théâtre dans le vieillissement.
No full textInternational audienceSi le développement d’intervention non médicamenteuse basé sur le théâtre commence à se développer auprès des personnes âgées force est de constater une grande hétérogénéité dans la mise en œuvre de ce type d’intervention au sein des institutions accueillant ce type de population.L’objectif de cette étude est de comprendre les représentations sociales pouvant exister interventions théâtrales un public âgé. En effet, ces ateliers sont souvent perçus comme étant difficiles à mettre en place auprès d’un public âgé et sont donc actuellement peu développés au sein des Etablissements d’Hébergement pour Personnes Âgées Dépendantes (EHPAD). Le concept des représentations sociales proposé par Moscovici (1961) sert de référentiel à ce travail de recherche.Cent-vingt et une personnes âgées résidant en EHPAD ont participé à cette étude. Le protocole est constitué d’une tâche d’association verbale libre, hiérarchisée et avec un contexte de substitution.Les cinq mots les plus cités dans le groupe personne âgée sont « pièce de théâtre », « difficile », « regarder », « bien » et « trouble cognitif ». Il y a une tendance significative à utiliser des mots à valence neutre dans le contexte libre pour l’ensemble des participants comparativement au contexte de substitution qui est plus positif (U = 2646 ; p < 0,001). Cependant, pour l’ensemble des participants la valence des mots a tendance à être plus positive lors de la tâche libre comparativement au contexte de substitution (W = -3,26 ; p < 0,001) qui est plus négatif.Ces données préliminaires laissent présumer des éventuelles représentations à tendance neutre les interventions théâtrales auprès d’un public âg
Performance Isolation in Multi Tenant Cloud Data Centers
No full textInternational audienceEnsuring network performance isolation in multitenant cloud data centers is critical for maintaining Service Level Agreement (SLA) compliance while supporting diverse and dynamic workloads. Traditional approaches predominantly emphasize flow-level fairness, which, although effective in certain scenarios, often fails to address the broader needs of tenant-based fairness essential for SLA-driven environments. My research introduces a novel framework leveraging control theory and AI-enhanced decision-making to achieve adaptive and intelligent network resource management. By shifting the focus to tenantbased fairness, the proposed solution ensures that resources are allocated equitably among tenants while meeting SLA requirements under fluctuating conditions. Preliminary investigations demonstrate the potential of this approach to enhance reliability, scalability, and efficiency, addressing the limitations of fairnesscentric methods focused solely on flows. This work lays the groundwork for intelligent, automated resource management in cloud data centers, advancing the capabilities of multi-tenant infrastructure to meet future demands
Evaluation of the biodistribution of different anti-TNF agents in combination with Immunotherapy in Melanoma model using ImmunoTEP
No full textInternational audienc
Predictive QoS for tele-operated driving over 5G SA networks: an experimental study
No full textInternational audienceVehicular connectivity is becoming an integral part of the automotive industry, pushed forth by edge computing and smart cities. However, in urban environments, wireless links are prone to physical disturbances. To maintain their service with these conditions, vehicular networking architectures need the ability to anticipate phenomena and adapt proactively. The multitude of variables involved in this problem calls for deep learning approaches. As such, predictive Quality-of-Service has shown promising contributions to the performance of low-latency infrastructures, but a field study of appropriate recurrent neural networks has yet to be conducted for vehicular networks. Moreover, the inherent heterogeneity of urban networks can be leveraged for service continuity purposes, by developing a radio access technology selection algorithm based on the predictions made for each technology. In that regard, while standalone 5G has been publicized as a key technology for very demanding applications like tele-operated driving, previous works have not studied its contribution in a real-life environment. This work thereby proposes a predictive Quality-of-Service infrastructure with proactive radio access technology selection for urban vehicular networks, evaluated by field trial and involving 5G SA, C-V2X and DSRC. Experimental observations show the enhanced contribution of 5G SA, with a minimal latency of 16 milliseconds and higher overall reliability than device-to-device communications. These measurements pave the way to cellular vehicular communications in applications with high service level requirements. Moreover, our system is able to ensure 99.9% success rates from 72% of the vehicle travel time to more than 90% of the time. Average reliability increases from 95.7% to 99.4%
Design of colloidal vectors for active targeting via complexation of biotinylated copolymers with gadolinium ions
No full textInternational audienceThis study presents the design of biotin-functionalized hybrid polyionic complexes (HPICs) using RAFT polymers for MRI applications. The synthesized colloidal vectors, complexed with gadolinium ions, demonstrate high stability, relaxivity, and biotin accessibility for active targeting. Preliminary in vivo studies confirm their potential for improved MRI contrast and pharmacokinetic tracking
Reflecting on 25+ Years of Devolution in Scotland and Wales: Has It Truly “Kill[ed] Nationalism Stone Dead?”
No full textInternational audienc
Optimal bounds for the Kobayashi distance near -smooth boundary points
No full text13 pagesIt is shown that the optimal upper and lower bounds for the Kobayashi distance near -smooth strongly pseudoconvex boundary points obtained in L. Kosinski, N. Nikolov, A.Y. Okten: "Precise estimates of invariant distances on strongly pseudoconvex domains", Adv. Math. 478 (2025), 110388, remain true in the general strongly pseudoconvex setting. In fact, the upper bound is extended to the general -smooth case. We also give upper and lower bounds for the Kobayashi distance near non-semipositive boundary points
Refinements on higher order Weil-Oesterl\'e bounds via a Serre type argument
No full text17 pages, 2 figuresWeil's theorem gives the most standard bound on the number of points of a curve over a finite field. This bound was improved by Ihara and Oesterl\'e for larger genus. Recently, Hallouin and Perret gave a new point of view on these bounds, that can be obtained by solving a sequence of semi-definite programs, and the two first steps of this hierarchy recover Weil's and Ihara's bounds. On the other hand, by taking into account arithmetic constraints, Serre obtained a refinement on Weil's bound. In this article, we combine these two approaches and propose a strengthening of Ihara's bound, based on an argument similar to Serre's refinement. We show that this generically improves upon Ihara's bound, even in the range where it was the best bound so far. Finally we discuss possible extensions to higher order Weil-Oesterl\'e bounds
SMOS: A streamlined model-based optimization approach for super-resolution ultrasound localization microscopy
No full textUltrasound Localization Microscopy (ULM) represents a significant advancement over conventional ultrasound (US) imaging, for reconstructing microvascular structures with micrometer-scale precision. However, existing model-based ULM approaches often rely on a complex five-step processing framework, which complicates parameter tuning and tissue signal suppression. Additionally, many of these methods either insufficiently account for the point spread function (PSF) - thereby compromising the accuracy of the final reconstruction - or are computationally intensive. While end-to-end data-driven methods show promise, they often require substantial training data and reliable ground-truth annotations. In this study, we introduce a novel approach named Streamlined Model-Based Optimization Approach for Super-Resolution Ultrasound Localization Microscopy (SMOS) which addresses these limitations. Our method improved the existing ULM process by replacing the first three steps with a single unified model. Experimental results on both in silico and in vivo datasets demonstrate the enhanced performance of our approach compared to existing benchmarks
Biased PromptORE: Enhancing Relation Extraction in Gendered Languages and Complex Texts The Case of Spanish Documents from the 16th Century
No full textECIR 2025 7-11 April Lucca, ItalyInternational audienc