HAL-UJM
Not a member yet
95306 research outputs found
Sort by
Attribute recognition: A new method for grouping planetary images by visual characteristics, using the example of Mn-rich rocks in the floor of Gale crater, Mars
No full textInternational audienceClassifying images is particularly challenging when working with large datasets without predefined groups. We present a new method for grouping images by visual similarity using relatively simple terminology and apply this method to the process of grouping NASA Curiosity rover ChemCam target images into visually similar groups. This method is designed for offline use, rather than on-board applications where power constraints are a consideration. Given the large quantity of data from ChemCam, we narrow the scope of our study to consider only rock targets that are early-mission and contain elevated manganese. A standard list of visual attributes is assessed for each target, and for each attribute on the list, a 1 is recorded if the ChemCam target image exhibits the attribute, and a 0 otherwise. The binary number resulting from this analysis encodes the visual characteristics of each image and is also used to determine similarity between images. Images are modeled as nodes in a network, and similarities between images are modeled as edges between nodes in the network. We find that when using a conservative threshold for similarity and an undirected, unweighted graph to represent the network, visually similar images cluster effectively into disjoint connected components. To improve the geologic usefulness of the resulting target groupings, we define a metric for weak component connectivity and explore methods for automatically partitioning weakly connected components. We compare these results to weighted-graph approaches, as well as to control tests using random partitions. Starting with a dataset of 201 ChemCam Remote Micro Imager mosaics, we found that the “automatic partitioning” method divided these images into 13 groups and resulted in better intra-group visual coherence than the other methods assessed. These results may be applied to motivate machine learning models for automatic attribute recognition to expand data labeling, as well as future classification efforts, including citizen science endeavors
Effective models for generalized Newtonian fluids through a thin porous media following the Carreau law
No full textInternational audienceWe consider the flow of a generalized Newtonian fluid through a thin porous medium of thickness ε, perforated by periodically distributed solid cylinders of size ε. We assume that the fluid is described by the 3D incompressible Stokes system, with a non-linear viscosity following the Carreau law of flow index 1<r< +∞, and scaled by a factor ε^γ , where γ ∈ R. Generalizing (Anguiano et al., Q. J. Mech. Math., 75(1), 2022, 1-27), where the particular case r<2 and γ = 1 was addressed, we perform a new and complete study on the asymptotic behaviour of the fluid as ε goes to zero. Depending on γ and the flow index r, using homogenization techniques, we derive and rigorously justify different effective linear and non-linear lower-dimensional Darcy’s laws. Finally, using a finite element method, we study numerically the influence of the rheological parameters of the fluid and of the shape of the solid obstacles on the behaviour of the effective systems
Kinetically constrained models
No full text125 pages, 17 figuresThis is a preprint of the following work: Ivailo Hartarsky, Cristina Toninelli. Kinetically constrained models. Springer Nature Switzerland, 53, 2025, SpringerBriefs in Mathematical Physics. It is the version of the author’s manuscript prior to acceptance forpublication and has not undergone editorial and/or peer review on behalf of the Publisher (where applicable).The final authenticated version is available online at: http://dx.doi.org/10.1007/978-3-031-93115-4International audienceThe goal of this book is to provide an introduction to the mathematical theory of Kinetically constrained models developed in the last twenty years, intended for both mathematicians and physicists
Prendre place dans un monde multi-espèces : Pour une étude des socialisations enfantines aux animaux
No full textInternational audienceAu travers d’une synthèse de travaux en sciences humaines et sociales portant sur les relations entre enfants et animaux, cet article met en évidence les quatre axes principaux qui ont structuré la recherche sur ce thème depuis les années 1970 : le rôle de l’animal dans le bien-être de l’enfant, la défense d’une éducation en faveur de la cause animale, l’objectivation par enquête statistique des attitudes enfantines à l’égard des animaux et les espaces d’interaction entre enfants et animaux. Il invite, ensuite, à poursuivre l’examen de cet objet en puisant dans les outils théoriques de la sociologie des socialisations. Des questionnements essentiels pour comprendre la construction du rapport aux animaux durant l’enfance demeurent encore sans réponse. Il conviendrait d’approfondir, par exemple, la réception des discours et des supports sur les animaux par les enfants, l’articulation en termes de dispositions entre une diversité d’instances de socialisation ou le rôle joué par les pairs dans un monde de l’enfance saturé par la présence animale
A Bregman Proximal Viewpoint on Neural Operators
No full textInternational audienceWe present several advances on neural operators by viewing the action of operator layers as the minimizers of Bregman regularized optimization problems over Banach function spaces. The proposed framework allows interpreting the activation operators as Bregman proximity operators from dual to primal space. This novel viewpoint is general enough to recover classical neural operators as well as a new variant, coined Bregman neural operators, which includes the inverse activation operator and features the same expressivity of standard neural operators. Numerical experiments support the added benefits of the Bregman variant of Fourier neural operators for training deeper and more accurate models
Le défi préhistorique. Repenser l'histoire depuis l'art paléolithique
No full textInternational audienceEn révélant une ancienneté vertigineuse et sublime, la découverte d’un art préhistorique a bouleversé notre culture en profondeur. En raison des lacunes des vestiges, de l’absence de sources textuelles et dela paradoxale modernité artistique du Paléolithique, ce temps incommensurable aux cadres historiques traditionnels impose de repenser l’histoire. Quels concepts et modèles ont été élaborés pour faire uneplace à la préhistoire dans l’histoire ? Quelle est leur portée épistémologique ? Que nous disent-ils de l’art, de notre histoire, de notre culture ? Convoquant des grands noms de la préhistoire et de l’anthropologie (Gabriel de Mortillet, Henri Breuil, André Leroi-Gourhan) ainsi que des théoriciens de l’art aussi différents qu’Alois Riegl, Élie Faure, Carl Einstein ou George Kubler, l’ouvrage envisage l’art préhistorique comme une matrice philosophique pour interroger les liens entre art, histoire et humanité
Credal ensembling in multi-class classification
No full textInternational audienceIn this paper, we present a formal framework to (1) aggregate probabilistic ensemble members into either a representative classifier or a credal classifier, and (2) perform various decision tasks based on this uncertainty quantification. We first elaborate on the aggregation problem under a class of distances between distributions. We then propose generic methods to robustify uncertainty quantification and decisions, based on the obtained ensemble and representative probability. To facilitate the scalability of the proposed framework, for all the problems and applications covered, we elaborate on their computational complexities from the theoretical aspects and leverage theoretical results to derive efficient algorithmic solutions. Finally, relevant sets of experiments are conducted to assess the usefulness of the proposed framework in uncertainty sampling, classification with a reject option, and set-valued prediction-making
Néolibéralisme
No full textLe néolibéralisme est un ensemble de doctrines politiques envisageant la société comme un marché où des individus entrepreneurs doivent valoriser leur capital alors que l’État se trouve réduit à « ses seules fonctions régaliennes aptes à garantir la fluidité de la circulation des marchandises et l’exécution des contrats » (Bihr, 2011). Le néolibéralisme promeut l’extension à toutes les sphères d’activité du régime de la compétition fondé sur le dogme de l’efficacité du marché et de la responsabilité individuelle