HAL Portal Univ-Poitiers
Not a member yet
51771 research outputs found
Sort by
Simulation de la propagation de fissures de fatigue sous hydrogène en milieu hétérogène
No full textNational audienc
Hydrogen-assisted crack growth in polycrystalline aggregate using an anisotropic gradient damage model
No full textInternational audienc
Surgical vs Percutaneous Spinal Cord Stimulation for Persistent Spinal Pain Syndrome Type 2: A Digital Pain Mapping and Pain Coverage Quantitative Assessment Combined With Multidimensional Clinical Outcome Study
No full textInternational audienceBackground: Persistent spinal pain syndrome type 2 (PSPS-T2) remains a significant challenge in chronic pain management, particularly when refractory to conventional treatments. Spinal cord stimulation (SCS) is an established therapeutic option, but the comparative effectiveness of paddle leads (PLs) and cylindrical leads (CLs) remains debated.Objective: This study aimed to objectively compare PLs and CLs in terms of pain coverage mapping and clinical outcomes using a digital pain-mapping tool and a multidimensional response index (MCRI).Materials and methods: A retrospective, monocentric, real-world cohort study included 175 patients with refractory PSPS-T2 who were implanted with either PL (n = 35) or CL (n = 140) between 2018 and 2024. Pain coverage was assessed using the PRISMAP digital tool, whereas clinical outcomes (pain intensity, functional disability, quality of life, psychologic distress) were evaluated at baseline and at three, six, and 12 months after implantation. A propensity score-weighted analysis was conducted to mitigate selection bias.Results: In the weighted analysis, PL exhibited significantly superior back pain coverage to that of CL (88.0% vs 67.6%; p = 0.002) and greater improvements in MCRI, pain intensity (visual analog scale), and health-related quality of life (EuroQol 5 Dimension 5 Levels) at 12 months (p < 0.05). However, both lead types showed significant clinical benefits over time. CL was associated with a higher lead reimplantation rate (9.7%; p = 0.003).Conclusions: Although PL revealed superior spatial targeting and long-term clinical efficacy, both lead types provided meaningful pain relief in patients with PSPS-T2. However, given the relatively small sample size in the PL group and the substantial difference in follow-up duration between groups, these results should be interpreted cautiously owing to potential time-dependent confounding. The choice between PL and CL should be guided by patient-specific factors, including pain distribution, functional status, and surgical considerations. The integration of digital pain mapping offers valuable insights for optimizing SCS therapy and tailoring neuromodulation strategies to individual needs
Le cas du Marquis de Roccaverdina (1903) : genèse éditoriale d’une marginalisation « élitiste »
No full textRubrique : InterlanguesInternational audienceThis article aims to reconstruct the conditions that made possible the publication of L. Capuana’s Le Marquis de Roccaverdina (Fontemoing, 1903). In a specific publishing and literary situation, we want to trace the genesis of an “elitist” marginalisation, examining the roles, positions and interests of the three protagonists : the author Capuana, the publisher Fontemoing and the translator Douësnel. The case of Le Marquis raises the question of authorial intentionlity, which, thus broken down and recomposed into different discursive units, emerges in a new light.Cet article reconstruit les conditions qui ont rendu possible la publication du Marquis de Roccaverdina de L. Capuana, paru chez Fontemoing en 1903. Dans une conjoncture éditoriale et littéraire spécifique, il s’agit de retracer la genèse d’une marginalisation « élitiste », en examinant les rôles, les positionnements et les intérêts des trois acteurs principaux : l’auteur Capuana, l’éditeur Fontemoing et la traductrice Douësnel. Le cas du Marquis interroge l’intentionnalité auctoriale qui, ainsi décomposée et recomposée en différentes unités discursives, émerge sous un nouveau jour.Questo articolo ricostruisce le condizioni che hanno reso possibile la pubblicazione di Le Marquis de Roccaverdina di L. Capuana (Fontemoing, 1903). In una specifica congiuntura editoriale e letteraria, si vuole ripercorrere la genesi di una marginalizzazione “elitaria”, esaminando i ruoli, le posizioni e gli interessi dei tre protagonisti : l’autore Capuana, l’editore Fontemoing e la traduttrice Douësnel. Il caso del Marquis solleva la questione dell’intenzionalità autoriale che, così scomposta e ricomposta in diverse unità discorsive, emerge in una nuova luce
Comparing Text Augmentation by GPT-3.5 and Llama3 for Evaluating Student Responses
No full textInternational audienceWriting is a critical educational task because it encompasses so many skills necessary for the modern world, including vocabulary and grammar acquisition, critical thinking, adapting to different audiences, and determining how best to communicate one's ideas. However, written assignments are notoriously time-consuming for teachers to grade, and timely feedback is critical for students' learning. Automated evaluation can provide quick student feedback while easing the manual evaluation burden for teachers. Current machine learning-based methods of evaluating student textual responses have met with varying degrees of success. One main challenge in training these models is the scarcity of student-generated data. Large volumes of training data are needed to create accurate models, and few educational tasks are large enough. To overcome this data scarcity issue, text augmentation techniques have been used to balance and expand the data set so that classification models can be trained with higher accuracy, providing more useful feedback for teachers and students. This paper examines the performance of text augmentation using two Large Language Models (LLMs) to provide supplemental texts for training models for classifying student answers in English and French educational tasks. Our results show that text generation can dramatically improve model performance on small data sets over simple self-augmentation, especially when the LLM is set to generate more varied responses
Advanced theory of mind predicts adolescents’ evaluation of source benevolence while reading
No full textInternational audienceSourcing skills support online readers' assessment of relevance by establishing the credibility of a message. But many adolescents struggle to efficiently engage in the critical evaluation of the information they read online. This study investigated the relationship between adolescents' Advanced Theory of Mind (AToM), their executive and verbal skills, and their abilities to assess specific dimensions of source reliability (i.e‥, source expertise, benevolence, and level of editorial control). Seventy-six French eighth graders completed tasks assessing their AToM, lexical quality, Working Memory (WM) updating, and sourcing abilities. The data showed a statistically significant relation between AToM and source benevolence assessment, after controlling for WM and lexical quality, suggesting that AToM plays a role in inferring the intentions of sources. However, no statistically significant relationship was found between AToM and other source attributes such as expertise or editorial control. Additionally, no association was observed between WM and AToM or sourcing. These findings highlight the importance of considering AToM in understanding adolescents' sourcing abilities and suggest the need for specific perspective-taking training to support their critical reading
Coupling a gas gun with an X-pinch x-ray source to perform x-ray diffraction under shock loading
No full textInternational audienceX-ray diffraction is an appropriate technique to probe crystalline materials and better understand their response under shock loading, particularly when they experience phase transition. This technique was already used at various large-scale facilities. Here, we present an alternative way to perform x-ray diffraction under shock loading at the laboratory scale by coupling an X-pinch x-ray generator with a single stage gas gun. This x-ray source is capable of generating a single polychromatic x-ray flash shorter than 100 ns. Preliminary static diffraction tests gave promising results, and then, an experimental apparatus was set up to perform in situ x-ray diffraction in a shock-loaded material. X-ray diffraction is performed in reflection at the interface between the studied sample and an anvil window to ensure a homogeneous pressure state within the probed region. A specific target configuration was designed to synchronize the x-ray emission with the temporary shocked state. The synchronization is achieved by the use of a trigger chain whose adjustable delay is chosen prior to the experiment based on the expected travel time of the shock wave throughout the target. The technique was successfully used to investigate the solid–solid phase transition of tin between β and γ phases. Results indicate a satisfying synchronization between the shock wave arrival and the x-ray emission. Diffractograms under shock loading show a disappearance of the static ambient figure (parent phase) and the development of a new diffraction pattern (daughter phase)
Perioperative antimicrobial prophylaxis and surgical site infection prevention, what clinicians and researchers must know
No full textInternational audienceSurgical site infections (SSIs) are among the most common healthcare-associated infections, leading to increased morbidity, prolonged hospital stays, and significant healthcare costs. Surgical antimicrobial prophylaxis (SAP) is a critical strategy for SSI prevention, yet its effectiveness is threatened by antimicrobial resistance and variability in clinical practice.This narrative review provides an evidence-based update on the pathophysiology of SSIs, highlighting the interplay between endogenous microbiota, surgical stress, and perioperative factors such as hypoxia, immune modulation, and microbiome disruption. The current state-of-the-art in SAP is reviewed, including antibiotic selection, timing, dosing, intraoperative redosing, and the avoidance of unnecessary postoperative administration.Key intra- and postoperative measures to reduce the risk of SSI are covered, including glycaemic control, body temperature management, goal-directed fluid therapy, and skin antisepsis. A critical appraisal of the supporting evidence is included, with emphasis on areas of ongoing debate.The final section outlines future research priorities: optimizing dosing in obese patients, evaluating continuous infusion, tailoring prophylaxis to surgical site and microbiome, and addressing the management of patients colonized with multidrug-resistant organisms. Non-antibiotic strategies and rapid diagnostic tests are also discussed as promising avenues to enhance precision in infection prevention.By integrating current knowledge with emerging perspectives, this review aims to support the refinement of SSI prevention strategies and contribute to antimicrobial stewardship in modern surgical practice
Reduced-order modeling of experimental turbulent flows: from linear projection-based methods to autoencoders
No full textInternational audienceTurbulent flows, a longstanding challenge in classical physics, are characterized by their nonlinear, multi-scale, and chaotic behavior. While the Navier-Stokes equations provide a comprehensive mathematical foundation, solving them numerically remains computationally expensive. This poses a challenge for iterative applications, such as closed-loop control or design optimization, which require surrogate models for faster simulations. In this paper, reduced-order models (ROMs) are explored as a promising solution. ROMs are techniques designed to capture the essential dynamics of turbulent flows at a significantly reduced computational cost by identifying and exploiting features inherent in the flow field. Two main classes of ROMs can be defined: projection-based methods and operator-based methods. Projection-based models, while effective, are inherently intrusive, as they require the application of basis expansions and projections onto the full-order model's operators. In contrast, operator-based ROMs offer non-intrusive alternatives but depend on prior knowledge of the ROM structure. To address the limitations of both approaches, this work explores the potential of neural compression algorithms, i.e. the application of neural networks for unsupervised feature extraction to compress data into lower-dimensional representations. As a proof of concept, this study examines the flow developing around a one degree-of-freedom elastically-mounted cylinder experiencing vortex-induced vibrations. The dataset consists of time-resolved, two-component velocity snapshots obtained experimentally by Particle Image Velocimetry for different Reynolds numbers, ranging from 6000 to 16000, based on the cylinder diameter and the free-stream velocity , in the wake of the oscillating cylinder. This work makes several contributions: (i) it assesses the performance of both AE and VAE compared to POD in reconstructing unseen flow regimes; (ii) it develops a single model capable of representing all Reynolds numbers in the dataset with just three variables; and (iii) it proposes novel techniques to analyze and exploit the mapping learned by the neural network
β-Variational autoencoder and transformer-based data-driven modeling of near-wall turbulence
No full textInternational audienceWall-bounded turbulent flows are fundamental to numerous engineering applications, yet complexity of near-wall turbulence dynamics remains challenging to model. Recent advances in machine learning, particularly β-variational autoencoders (β-VAEs) and transformer-based methods, have shown promise in uncovering and predicting these dynamics. Building on this progress, this work presents a framework integrating a novel β--VAE architecture for unsupervised feature extraction and a unidirectional transformer for temporal evolution prediction in turbulent channel flows. Using a minimal channel configuration at a friction Reynolds number of 200, simulated with the spectral element method code SOD2D, our framework effectively captures and compactly represents near-wall turbulence features. Detailed analyses, including spanwise energy reconstructions and quadrant analyses, validate the approach's efficacy. Additionally, transformer-based predictions of chaotic dynamics are assessed using Lyapunov exponents and Poincaré maps. This study provides a foundational testbed for near-wall turbulence modeling, offering insights applicable to more complex, larger channels