Portail HAL Nantes Université
Not a member yet
120169 research outputs found
Sort by
Réduction de dimension non linéaire pour l’approximation en grande dimension et les problèmes inverses
No full textThis thesis is concerned with two types of nonlinear dimension reduction.In a first part we focus on a model reduction framework for parameter dependent partial differential equations.In particular in Chapter 2 we introduce a dictionary-based approach to approximate a solution of such equation in the context of an inverse problem, where only few linear measurements on the solution are given.We introduce a new selection criterion based on random sketching for adaptively selecting a space generated by a few elements of the dictionary.We show near-optimality with high probability and we provide the practical details to ensure computational efficiency for both building and evaluating this new criterion.In Chapter 3 we consider an operator interpolation method recently introduced for preconditioning linear model reduction based on Galerkin projection.This is of particular interest for problems admitting good linear approximation but whose associated projection operator is ill-conditioned.The contribution of this thesis is to propose a practical construction of these preconditioners, illustrated with numerical experiments.In a second part we consider low dimensional featuring for approximating high dimensional functions.In particular in Chapter 4 we consider gradient-based construction of nonlinear feature maps, which leverages Poincar\'e inequalities on nonlinear manifolds to derive a computable objective function.However, optimizing the latter is in general a difficult non-convex problem.We thus introduce new quadratic surrogates to this objective function.Leveraging concentration inequalities, we provide sub-optimality results for a class of feature maps, including polynomials, and a wide class of input probability measures.In Chapter 5 we extend the approach from the previous chapter to dimension reduction for a family of high dimensional functions (collective dimension reduction).We then investigate structured forms of feature maps, aiming to leverage the aforementioned gradient-based method and surrogates to learn compositional function networks.For every chapter we provide open-source implementation of the methods presented therein.Cette thèse traite de deux types de réduction de dimension non linéaire.Dans une première partie, nous nous concentrons sur un cadre de réduction de modèle pour les équations différentielles partielles dépendantes des paramètres.En particulier, dans le Chapitre 2, nous introduisons une approche basée sur un dictionnaire pour approximer une solution d'une telle équation dans le contexte d'un problème inverse, où seules quelques mesures linéaires sur la solution sont connues.Nous introduisons un nouveau critère de sélection basé sur l'algèbre linéaire aléatoire pour sélectionner de manière adaptative un espace généré par quelques éléments du dictionnaire.Nous montrons une quasi-optimalité avec forte probabilité et nous fournissons les détails pratiques pour garantir l'efficacité computationnelle tant pour la construction que pour l'évaluation de ce nouveau critère.Dans le Chapitre 3, nous considérons une méthode d'interpolation d'opérateurs récemment introduite pour le préconditionnement de la réduction de modèle linéaire basée sur la projection de Galerkin.Cela présente un intérêt particulier pour les problèmes admettant une bonne approximation linéaire mais dont l'opérateur de projection associé est mal conditionné.La contribution de cette thèse est de proposer une construction pratique de ces préconditionneurs, illustrée par des expériences numériques.Dans une deuxième partie, nous considérons des caractéristiques de faible dimension pour approximer les fonctions en grande dimension.En particulier, dans le Chapitre 4, nous considérons la construction de caractéristiques non linéaires basée sur le gradient, qui exploite les inégalités de Poincaré sur les variétés non linéaires pour obtenir une fonction objective calculable.Cependant, l'optimisation de cette dernière est en général un problème non convexe difficile.Nous introduisons donc de nouveaux substituts quadratiques à cette fonction objective.En tirant parti des inégalités de concentration, nous fournissons des résultats de sous-optimalité pour une classe de caractéristiques, comprenant les polynômes, et une large classe de mesures de probabilité des variables d'entrée.Dans le Chapitre 5, nous étendons l'approche du chapitre précédent à la réduction de dimension pour une famille de fonctions en grande dimension (réduction de dimension collective).Nous étudions ensuite des formes structurées de caractéristiques, dans le but d'exploiter la méthode basée sur le gradient et les substituts mentionnés ci-dessus pour apprendre des réseaux de fonctions compositionnelles.Pour tous les chapitres nous fournissons une implémentation open source des méthodes qui y sont présentées
Tunneling nanotubes propagate a BMP-dependent preneoplastic state
No full textTunneling nanotubes (TNTs) are thin, actin-based structures allowing long-distance communication between cells by promoting the transfer of molecules and organelles. Well-known to contribute to cancer progression, we investigated their role in early tumor initiation using normal human mammary primary cells and an in-house human breast cell model recapitulating early transformation. We observed that TNTs become more abundant and elongated during this process, and preferentially connect transformed donor cells to non-transformed acceptor cells. We show that this long-range, directional communication involves the transfer of the signaling receptor BMPR1b. Within days, this transfer induces gene expression changes in acceptor cells, consistent with early transformation programs. Functional analyses of these acceptor cells revealed phenotypic changes, including anchorage-independent growth. In particular, the transferred BMPR1b receptor sensitized acceptor cells to BMP2 signals present in the microenvironment, amplifying their transformation potential. Hence, by tracking the earliest molecular responses in acceptor cells, we deciphered the initial steps of a transformation cascade triggered by TNT-mediated transfer. These findings uncover a BMP-dependent mechanism by which transformed cells propagate a preneoplastic state to adjacent and distant cells at the very onset of transformation, offering new perspectives on how epithelial transformation arises and spreads to neighboring cells
Cetaceans exhibit region-specific habitat preferences across tropical waters
No full textInternational audienc
Digital twin-based anomaly detection under concept drift: A comparison between iterative batch learning and incremental learning approaches
No full textInternational audienc
Le juge administratif au supermarché. Réflexions à partir d’une autre affaire Eurelec
No full textInternational audienc
Unlocking the power of non-coding RNAs: toward real-time cancer monitoring in precision oncology
No full textInternational audienceNo abstract availabl
Participatory democracy in question: The case of “the sea in debate”
No full textInternational audienceWhile participatory democracy invites all citizens to take part directly in the decision-making process, the selection of participants in public debates is a critical issue for the legitimacy of the resulting public choices. This paper examines this question in the context of the national public debate on offshore wind energy held in France in the first quarter of 2024. We study an original survey measuring spatial preferences for offshore wind energy in which both participants in the public debate and respondents from the general population were simultaneously surveyed. We find large differences between the two groups of respondents in terms of gender, age, and education, as well as in their spatial preferences for wind farm locations. Using an entropy balancing approach, we reject the hypothesis that these differences in spatial preferences are due to composition effects. These findings underscore the need for policymakers to exercise caution when interpreting the outcomes of public debates
Speaking Face-to-Face with a Virtual Avatar to Reduce Anxiety in Students Who Stutter: Tool Development and Pilot Study Results
No full textInternational audiencePurpose: Speaking in class is challenging for students who stutter. Cognitive-behavioral therapy (CBT) with exposure in virtual reality (VR) emerges as a promising intervention for treating speaking anxiety in pediatric populations. This pilot study tested if real-time avatar-based VR can elicit anxiety responses while remaining acceptable to youth who stutter. Method: Twelve students who stutter (aged 9-18) were randomly assigned to a single training session conducted either (1) in VR with a realistic avatar controlled live by their SLP, or (2) in role-play with their SLP, before facing a real actor. We assessed system acceptability, anxiety levels and perceived self-efficacy. Results: The VR system was well accepted and elicited physiological arousal comparable to reallife interactions. Although participants reported experiencing less anxiety during VR, skin conductance level showed higher arousal; suggesting a divergence between the subjective report and physiological response. Finally, one training session (either in VR or with the SLP) did not produce gains in self-efficacy or decrease in anxiety related to the final real-actor conversation. Conclusion: This study demonstrates evidence that the potential use of immersive VR could represent an acceptable and viable complementary strategy for SLP treatment, that could control exposure parameters while evoking physiological responses similar to real-life contexts. The differences between subjective and physiological measures suggest that VR is inducing anxiety responses differently than it was perceived. Further research could investigate the use of VR as anxiety interventions for students who stutter and should be explored across multi-session studies to understand their therapeutic effect.</div