Archive ouverte de Centrale Lyon
Not a member yet
32420 research outputs found
Sort by
Euclid preparation. Galaxy power spectrum modelling in redshift space
No full textInternational audienceAccurate modelling of redshift-space distortions (RSD) is essential for maximizing the cosmological information extracted from large galaxy redshift surveys. In preparation for the forthcoming analysis of the Euclid spectroscopic data, we investigate three approaches to modelling RSD effects on the power spectrum multipoles of mock H emission line galaxies. We focus on two one-loop perturbation theory models -- the effective field theory (EFT) and velocity difference generator () -- which differ in their treatment of the real-to-redshift space mapping on small scales, and a third approach, the BACCO emulator, which adopts a hybrid strategy combining perturbation theory with high-resolution N-body simulations. We assess the ability of these models to recover key cosmological parameters, including the expansion rate , the cold dark matter density parameter , and the scalar amplitude , across four redshift bins spanning . In each bin, we find that and BACCO outperform the EFT model across all scales up to . While BACCO saturates in constraining power at intermediate scales and higher redshift, the model continues to improve parameter constraints beyond . The EFT model, although robust on large scales, exhibits significant parameter biases for , limiting its applicability to Euclid-like H samples. Among the full perturbation theory-based models, the enhanced treatment of small-scale RSD effects in improves cosmological parameter constraints by up to a factor of two
Stratégies de Stabilisation pour les EDP linéarisées par des méthodes d'optimisation
No full textWe adapt an optimization technique, originally developed for quadratic matrix model updating problems, to construct optimal stabilizers for linearized partial differential equations (PDEs). The proposed approach is flexible, allowing the optimization of mass, damping, and stiffness operators either individually or simultaneously. It is also straightforward to implement within standard numerical frameworks. We analyze the numerical and practical properties of the method, highlighting its robustness even with loworder discretizations and its capacity to produce interpretable stabilization operators. The effectiveness and versatility of the approach are demonstrated through numerical experiments on several representative linearized PDEs, including the Klein-Gordon, Boussinesq, Benjamin-Bona-Mahony, and Korteweg-de Vries equations. Our results indicate that the method provides a unified, efficient, and modular tool for PDE stabilization.Nous adaptons une technique d'optimisation, initialement développée pour les problèmes de mise à jour de modèles matriciels quadratiques, afin de construire des stabilisateurs optimaux pour les équations aux dérivées partielles (EDP) linéarisées. L'approche proposée est flexible, elle permet l'optimisation des opérateurs de masse, d'amortissement et de rigidité individuellement ou simultanément. Elle est également facile à implémenter dans les environnements numériques standards. Nous analysons les propriétés numériques et pratiques de la méthode, en soulignant sa robustesse même avec des discrétisations d'ordre faible et sa capacité à produire des opérateurs de stabilisation interprétables. L'efficacité et la polyvalence de l'approche sont démontrées par des expériences numériques sur plusieurs EDP linéarisées représentatives, notamment les équations de Klein-Gordon, de Boussinesq, de Benjamin-Bona-Mahony et de Korteweg-de Vries. Nos résultats indiquent que la méthode constitue un outil unifié, efficace et modulaire pour la stabilisation des EDP
Hyperaldostéronisme primaire au cours de la grossesse et de la ménopause
No full textInternational audienc
Almost cohomology of finite-dimensional Lie rings
No full textWe introduce almost cohomology groups for Lie rings definable in finite-dimensional theory. In particular, we define the 0th and 1st almost cohomology groups of a Lie ring module. Moreover, we prove that the 1st almost cohomology group of a finite-dimensional definable Lie ring module is finite if the 0th almost cohomology group is finite
Analysis of Corona and Surface Discharge Signals from Different Non-Intrusive Sensors under HVDC
No full textInternational audienceThe use of high voltage direct current (HVDC) transmission technology has increased significantly in recent years due to its numerous advantages over traditional alternating current (AC) transmission. The aim of the present work is to investigate the efficiency of different non-intrusive sensors for partial discharge measurement in medium or high voltage direct current equipment such as air or gas insulated switchgear. In HVAC systems, partial discharge detection and quantification are well understood. However, under HVDC, it is a necessary to evaluate the efficiency of classic techniques and commercial non-intrusive sensors. This work presents a study of three different sensors for partial discharge measurement under HVDC: high frequency current transformer (HFCT), ultra-high frequency antenna (UHF), and transient earth voltage (TEV). The signals of the different sensors for corona and surface discharge detection are presented and compared in both temporal and frequency domains. The Pulse Sequence Analysis (PSA) technique is used for apparent charge and repetition rate study and the effect of the discharge type on the PSA patterns is investigated. The results show that the PSA patterns are sensitive to the discharge type and conditions. The increase of the applied voltage of corona leads to the increase of the apparent charge differences and the time duration between successive pulses. The spectrums of the discharge pulses of the three sensors are different and the maxima of the TEV and UHF PD signals are dependent on the apparent charge
LPV system identification with unknown scheduling variable
No full textA Linear Parameter Varying (LPV) system is a system whose parameters depend on an exogenous variable, the so-called scheduling variable. LPV systems are generally controlled using LPV controllers i.e., controller whose parameters also depend on the scheduling variable. In order to be able to use an LPV controller for an LPV system, it is generally assumed that the scheduling variable of the LPV system is either fully known or measurable. Here, we consider the case where the scheduling variable has to be identified based on data collected on the system. We do that in the case of LPV systems that are successively operated at different constant values of the scheduling variable. We show how to identify the maximum likelihood estimate of the unknown constant scheduling variable and we derive the statistical properties of this estimate. Moreover, we also develop an optimal experimental design framework in order to optimally design the experimental conditions of the identification experiment yielding the estimate of the scheduling variable
Construction automatique d'un graphe de connaissances géo-historiques à partir de textes encyclopédiques anciens
No full textNational audienceLes encyclopédies anciennes, comme celle de Diderot et d'Alembert (1751-1772), offrent une ressource précieuse pour étudier l'évolution des savoirs géographiques, mais leur ampleur complique toute analyse manuelle. Cet article présente une méthode automatique de construction d'un graphe de connaissances géo-historiques à partir de ces textes. Nous proposons des ontologies spatiale et de provenance adaptées au corpus et introduisons un gold standard de 2 750 articles géographiques. Le pipeline combine apprentissage supervisé et grands modèles de langage pour la classification d'articles, le typage d'entités et l'extraction de relations spatiales. Les performances atteignent F1 = 92% pour les relations et F1 > 97% pour la classification, aboutissant à un graphe RDF de 35 000 entités et 46 000 relations. Ce travail ouvre la voie à l'analyse computationnelle des savoirs géographiques anciens
On the relaxation towards mechanical equilibrium for two-pressure compressible flows
No full textWe introduce a symmetrization of a one-velocity two-pressures Baer-Nunziato type model for mixtures of barotropic compressible fluids. It allows us to justify the zero compaction viscosity limit and to recover a solution of the so-called Kapila model. On the other hand, the symmetrization highlights a pressure-induced stabilization mechanism which allows us to recover a global-in-time existence result for initial data close to constant states
STORM: Slot-based Task-aware Object-centric Representation for robotic Manipulation
No full textVisual foundation models provide strong perceptual features for robotics, but their dense representations lack explicit object-level structure, limiting robustness and contractility in manipulation tasks. We propose STORM (Slot-based Task-aware Object-centric Representation for robotic Manipulation), a lightweight object-centric adaptation module that augments frozen visual foundation models with a small set of semantic-aware slots for robotic manipulation. Rather than retraining large backbones, STORM employs a multi-phase training strategy: object-centric slots are first stabilized through visual–semantic pretraining using language embeddings, then jointly adapted with a downstream manipulation policy. This staged learning prevents degenerate slot formation and preserves semantic consistency while aligning perception with task objectives. Experiments on object discovery benchmarks and simulated manipulation tasks show that STORM improves generalization to visual distractors, and control performance compared to directly using frozen foundation model features or training object-centric representations end-to-end. Our results highlight multi-phase adaptation as an efficient mechanism for transforming generic foundation model features into task-aware object-centric representations for robotic control
On the parameterized complexity of the Maker-Breaker domination game
No full textSince its introduction as a Maker-Breaker positional game by Duchêne et al. in 2020, the Maker-Breaker domination game has become one of the most studied positional games on vertices. In this game, two players, Dominator and Staller, alternately claim an unclaimed vertex of a given graph G. If at some point the set of vertices claimed by Dominator is a dominating set, she wins; otherwise, i.e. if Staller manages to isolate a vertex by claiming all its closed neighborhood, Staller wins. Given a graph G and a first player, Dominator or Staller must have a winning strategy. We are interested in the computational complexity of determining which player has such a strategy. This problem is known to be PSPACE-complete on bipartite graphs of bounded degree and split graphs; polynomial on cographs, outerplanar graphs, and block graphs; and in NP for interval graphs. In this paper, we consider the parameterized complexity of this game. We start by considering as a parameter the number of moves of both players. We prove that for the general framework of Maker-Breaker positional games in hypergraphs, determining whether Breaker can claim a transversal of the hypergraph in k moves is W[2]-complete, in contrast to the problem of determining whether Maker can claim all the vertices of a hyperedge in k moves, which is known to be W[1]-complete since 2017. These two hardness results are then applied to the Maker-Breaker domination game, proving that it is W[2]-complete to decide if Dominator can dominate the graph in k moves and W[1]-complete to decide if Staller can isolate a vertex in k moves. Next, we provide FPT algorithms for the Maker-Breaker domination game parameterized by the neighborhood diversity, the modular width, the P4-fewness, the distance to cluster, and the feedback edge number