HAL Portal Univ-Toulon
Not a member yet
34537 research outputs found
Sort by
Muscle contraction type does not influence the etiology of the immediate crossover fatigue evidenced after submaximal unilateral exercises of knee flexor muscles
No full textInternational audienceMotor performance fatigue can occur in a non-exercised muscle following a unilateral exercise performed with the contralateral homologous muscle. While many studies have investigated this crossover fatigue phenomenon, the influence of muscle contraction type is unsettled. This study examined the occurrence and etiology of crossover fatigue after submaximal isometric (ISO), concentric (CONC) or eccentric (ECC) contractions of the knee flexors (KF). In separate sessions, thirteen participants performed unilateral ISO, CONC or ECC KF exercises inducing a 20 % reduction in maximal voluntary isometric contraction (MVIC) of the exercised limb (EL). Motor performance fatigue (i.e., MVIC, voluntary activation [VA], electrically evoked muscle responses [Pt 100Hz and Pt 10Hz -to-Pt 100Hz ratio]) and perceived muscle soreness (PMS) of both EL and nonexercised limb (NEL) were assessed before (PRE), immediately (POST), and 24 h post-exercise (POST24). Global perceived motor fatigue (PMF) was also measured. At POST, significant reductions in MVIC were observed for both EL (-27.7 %; p < 0.001) and NEL (-8.5 %; p < 0.01) indicating the presence of crossover fatigue. At POST, Pt 100Hz (-7.5 %; p < 0.05) was impaired in both limbs whereas VA was altered in the EL only (-2.6 %; p < 0.001). PMF was increased at POST only (p < 0.001). Finally, PMS scores were greater at POST (p < 0.01) and POST24 (p < 0.001) than PRE values regardless of the limb considered. At POST24, MVIC, Pt 100Hz , VA and PMF had returned to PRE values. Muscle contraction type did not influence crossover fatigue of the KF nor its etiology.</div
Commissaires de justice : panorama législatif et jurisprudentiel 2023/2024
No full textInternational audienceSi certaines évolutions concernent directement le statut des commissaires de justice, la majorité d’entre elles ont trait aux procédures civiles d’exécution (mesures conservatoires, comme mesures d’exécution forcée). Plutôt que de raisonner procédure par procédure, afin d’éviter un « effet catalogue », le parti a été pris de présenter les réformes législatives et interprétations jurisprudentielles au moyen de deux axes d’analyse, en l’occurrence celui des principaux acteurs (fonction et statut) et celui des sources constitutionnelles du droit de l’exécution
"Finn MacCool: héro légendaire de l'Irlande". Interview de Natacha Sumner, Harvard University, pour la Chaine YouTube Imaginaire celtique, créée par Frédéric Armao, Noémie Beck, et Gaël Hily. Sous-titrée en français. 15 février 2025. https://www.youtube.com/watch?v=yCzb0ub42V8
No full text"Finn MacCool: héro légendaire de l'Irlande". Interview de Natacha Sumner, Harvard University, pour la Chaine YouTube Imaginaire celtique, créée par Frédéric Armao, Noémie Beck, et Gaël Hily. Sous-titrée en français. 15 février 2025. https://www.youtube.com/watch?v=yCzb0ub42V
Prédiction de la trajectoire du patient : Intégration des notes cliniques aux transformers
No full textInternational audienceThe prediction of patient disease trajectories using Electronic Health Records (EHRs) ischallenging due to non-stationarity, the granularity of medical codes, and difficulties in integratingmultimodal data. Current models often overlook critical insights in unstructured data,primarily relying on structured diagnosis codes. In this paper, we propose a novel approachthat incorporates unstructured clinical notes into deep learning models for sequential diseaseprediction. By embedding clinical notes in Transformer-based models, we provide a richerrepresentation of patient histories, enhancing accuracy in predicting future diagnoses. Our experimentsshow significant improvements in predictive performance compared to traditionalmodels relying solely on structured codes.La prédiction des trajectoires de maladies à partir des dossiers médicaux électroniques (DME) pose des défis liés à la non-stationnarité des données, à la granularité des codes médicaux et à l'intégration de données multimodales. Les DME combinent données structurées, comme les codes de diagnostic, et données non structurées, telles que les notes cliniques, souvent sous-exploitées malgré leur richesse informative. Pour remédier à ces limites, nous proposons d'intégrer les notes cliniques non structurées dans des modèles de prédiction séquentielle des maladies basés sur les modèles Transformers, améliorant ainsi la précision des prédictions de diagnostics futurs. Les expérimentations menées sur les ensembles de données MIMIC-IV ont montré que l'approche proposée offre de meilleures performances en comparaison des modèles traditionnels reposant uniquement sur des données structurées.</div
Achieving measurement comparability in mercury speciation analysis in seawater: Key requirements and best practices
No full textInternational audienc
Part-Of-Speech Sensitivity of Routers in Mixture of Experts Models
No full textAccepted at COLING 2025International audienceThis study investigates the behavior of model-integrated routers in Mixture of Experts (MoE) models, focusing on how tokens are routed based on their linguistic features, specifically Part-of-Speech (POS) tags. The goal is to explore across different MoE architectures whether experts specialize in processing tokens with similar linguistic traits. By analyzing token trajectories across experts and layers, we aim to uncover how MoE models handle linguistic information. Findings from six popular MoE models reveal expert specialization for specific POS categories, with routing paths showing high predictive accuracy for POS, highlighting the value of routing paths in characterizing tokens
Investigating the quality of European silver eels by quantifying contaminants and parasite infestation in a French Mediterranean lagoon complex
No full textInternational audienceCoastal lagoons are diverse habitats with significant ecological gradients, which provide crucial ecosystem services but face threats from human activities such as invasive species and pollution. Among the species inhabiting the lagoons, the critically endangered European eel (Anguilla anguilla) is an emblematic species strongly impacted by contamination and parasitism. Several indicators were developed to assess the quality of eel at a large geographic scale. Most indicators are based on the concentration of individual pollutant and/or abundance of parasites separately without considering individual variations. This study assessed the quality of 59 eels captured at three different sites inside a Mediterranean lagoon complex (the Camargue, South of France), by integrating multiple degradation factors (POPs, TEs, and A. crassus infestation) and considering individual eel characteristics (length, age, growth rate, and sex). Using multivariate TOPSIS analysis including these degradation factors, this study found that eel quality decreased with age but did not significantly vary between sites. When focusing on each degradation factor, A. crassus infestation rates were lower in older eels, independently to the site, however, the POPs and TEs contaminations were lower in the Grandes Cabanes site compared to the Vaccarès and Fumemorte sites even if larger and younger eels were more contaminated by POPs. These findings reveal the finescale spatial variability in eel quality, with TOPSIS analysis providing a robust method to rank and score scenarios. This approach enhances the understanding of habitat degradation sources affecting eel contamination and parasitic infestation, supporting more effective strategies for sustainable habitat management
The golden triangle of place reputation for metropolitaninternational cities: the cases of London and Paris
No full textInternational audienceThis prescient book explores the relationship between city branding, reputation management and tourism development through a comprehensive collection of propositions to re-examine and refine city branding tools and ideas.Enhancing the existing literature and with contributions from cities across four continents, chapters set out how city branding relates to urban identity and image, city reputation and sustainable tourism development. Illustrating how city branding can contribute towards a more sustainable future for cities, the book transcends disciplinary boundaries and uses methodological propositions to provide an innovative assessment of the current status of reputation management and tourism development. Showcasing new understandings of tools and conceptualizations, this forward-thinking book develops this field of study in both theory and practice.City Branding is an excellent resource not only for scholars of urban studies and tourism management but also for those studying local development, planning, urban governance, placemaking and place management
Conceptualisation et modélisation du marketing relationnel responsable des e-marketplaces ?
No full textInternational audienceParmi les acteurs du e-commerce, les e-marketplaces ont acquis une place de choix. Pour la France, elles représentent, en 2023, 29% du volume d’affaire de la vente à distance d’après la FEVAD. La e-marketplace se définit comme une plateforme numérique d’intermédiation sur laquelle se rencontre des vendeurs et des acheteurs. Ce format s’appui d’après Abhishek et al. (2015), sur la théorie de l’agence. En fait, contrairement aux magasins en ligne traditionnels qui vendent leurs propres produits, la e-marketplace permet à des vendeurs tiers de proposer leurs produits ou services à travers une plateforme numérique. Celle-ci se charge généralement de l'infrastructure de transaction, y compris des paiements, de la logistique, et parfois même du marketing. Pour Bakos (1991), l’un des premiers auteurs à en avoir donné une définition, le principal avantage pour l’acheteur consiste en la réduction des coûts de recherche ; pour le vendeur, elle augmente l’efficacité des transactions. Le relationnel est, ainsi, au cœur de son business model en travaillant à faire rencontrer différentes parties prenantes : des acheteurs, des vendeurs sans oublier ses salariés qui « animent » la plateforme. Depuis quelques années, les acteurs des e-marketplaces cherchent à rendre compatible leur Business Model relationnel à la responsabilité sociale et environnementale (RSE) de manière à répondre à une conscientisation de plus en plus prenante des trois parties prenantes impliquées dans ce modèle d’affaire mais aussi de la société en générale. A partir de l’étude, via une netnographie, de deux cas d’e-marketplaces dotés d’une orientation RSE différente, un modèle de performance relationnelle compatible RSE est proposé. Ce modèle se conçoit en final comme un outil de réflexion managérial d’équilibre coopératif
Learning Permutations in Monarch Factorization
No full textInternational audienceIn order to reduce the quadratic cost of matrixvector multiplications in dense and attention layers, Monarch matrices have been recently introduced, achieving a sub-quadratic complexity. It consists in factorizing a matrix using fixed permutations and learned block diagonal matrices, at the price of a small performance drop. We propose a more general model where some permutations are learned. The optimization algorithm explores the space of permutations using a Straight-Through Estimator (STE) inspired by the support exploration algorithm designed for sparse support recovery. Our experimental results demonstrate performance improvement in the context of sparse matrix factorization and of end-to-end sparse learning