HAL-INSA Toulouse
Not a member yet
34325 research outputs found
Sort by
Contractive transport maps from to nearly spherical surfaces with positive Ricci curvature
No full textWe prove that every nearly spherical, positively curved surface is the contractive, volumepreserving image of a round sphere. The proof combines three main tools: the Ricci flow on surfaces, the Kim-Milman construction, and a multiscale Bakry-Émery criterion.</div
Θ-Categories and Tannakian duality
No full textWe introduce a notion of Θ-categories, which is a renement of the notion of symmetric monoidal ∞-categories. We use this notion to prove a Tannakian duality statement, relating Θ-categories with fpqcstacks by means of a certain stack of ber functors in the context of Θ-categories. This provides, over a base ring of arbitrary characteristic, a strong link between Tannakian Θ-categories and the schematic homotopy types of [Toe06]. Contents</div
Finding the right regression testing method: a taxonomy-based approach
No full textInternational audienceWith numerous regression testing (RT) methods available in the literature, it is challenging to choose the right one for a specific context. Practitioners need support identifying suitable research. To this end, recent work has proposed a taxonomy. By mapping both the RT problem and existing solutions onto the taxonomy, practitioners should be able to determine which solutions are best aligned with their problem. Our work explores the practical relevance of this idea through an industrial case study. The context is the development of R&D projects at a major automotive company, in the domain of connected vehicles. We developed an RT problem solving approach based on the taxonomy. Following the approach, we characterized the RT problem, identified a set of 8 potentially relevant solutions from a set of 52 papers, and empirically evaluated their suitability. Our approach was successful, as we found effective RT methods among those selected using the taxonomy. One method, in particular, demonstrated remarkable robustness across various datasets, making it a strong recommendation for the industrial partner. However, this success came at the cost of difficulties due to unclear taxonomy elements, missing elements, and paper classification errors. We conclude that the taxonomy has practical value but would have to mature for easier applicability
Efficient ion re-acceleration in laboratory-produced interpenetrating collisionless shocks
No full textInternational audienceAlthough the origin of cosmic rays (CRs) remains an open question, collisionless magnetized shock waves are widely regarded as key sites for particle acceleration. Recent theories further suggest that shock-shock collisions in stellar clusters could provide the additional acceleration needed to explain the observed high-energy CR spectrum. Here, we investigate this hypothesis through a laser-based experiment that creates magnetized plasma conditions similar to astrophysical environments. Our results demonstrate that interpenetrating collisionless shocks can significantly boost the energy of ambient protons previously energized by the individual shocks, while also improving the overall acceleration efficiency. Numerical kinetic simulations corroborate these findings, revealing that protons are reaccelerated via their bouncing motion in the convective electric fields of the colliding magnetized flows. By allowing to highly energize ambient protons, our novel colliding-shock platform opens the prospect to test the long-discussed mechanism of diffusive shock acceleration in a controlled laboratory setting
Evaluating Embeddable Language Models in Verbalizing Rule-based Inferences through Justifications
No full textInternational audienceWhile Language Models have shown promising performance, they still struggle with limitations regarding reasoning and are very token-sensitive. In contrast, knowledgebased systems, such as ontologies, allow for provable logically valid reasoning and provide explicit justifications regarding newly inferred knowledge. However, those justifications can be hard to understand for non-expert users given their formal syntax and their length. We investigated if language models could be considered as reliable tools for verbalizing such explanations, thus increasing explainability over reasoning output. This paper presents a reference evaluation of a set of embeddable language models on a task of translation from ontology formatted inferences and justifications into natural language sentences. We show that the order of justifications significantly decreases performance, whereas adding the inference rule as additional context significantly improves performance, leading to more reliable results
Exponentially Stable Stubborn Observers for Discrete-Time Linear Systems (Extended Version)
No full textInternational audienceState estimation is crucial for control and monitoring of dynamical systems, but sporadic disturbances (outliers) can severely degrade the estimator's performance. This paper studies the stability of a so-called "stubborn" observer, designed to mitigate the effects of outliers, for discrete-time linear systems. We prove that the detectability of the system is both necessary and sufficient for global exponential stability of the observer, providing an improvement over existing results that rely on potentially infeasible LMI conditions. The proof is constructive, leading to practical guidelines for selecting the observer parameters. Numerical simulations demonstrate the observer's effectiveness in mitigating outliers and its superior performance compared to a classical Luenberger observer
E-MUSE - Data Management Plan
No full textTo understand a complex microbial ecosystem and identify levers to control and/or predict its evolution, biological, mathematical, statistical and machine learning tools must be developed. Innovative modelling methodologies have been developed. Data have been generated at multi-scale levels from single-cells to microbial populations and to macroscopic properties. Genome-scaled models have been constrained with the data produced. Multi-scale data have been integrated and correlated to design mathematical models and develop predictive models
On the crack propagation along brittle viscoelastic adhesive layer in DCB test
No full textInternational audienceThe mechanical behaviour of adhesives is known to be sensitiveto the deformation rate. The dependence of the critical StrainEnergy Release Rate (SERR) in bonded joints on the crack propagationrate has been widely reported in the literature.Typically, the evolution of the critical SERR with respect to thecrack propagation rate is often presented as a “master curve”,without specifying the necessary conditions that ensure thegeneralization and transferability of the findings. In this context,the present work provides a theoretical analysis of the DCB testin the presence of a brittle viscoelastic interface. A Euleriandescription is employed to describe the steady-state crack propagationregime. The results are then compared to thoseobtained from more comprehensive Lagrangian simulations ofthe DCB test. The analysis reveals that the evolution of thecritical SERR depends not only on the crack propagation ratebut also on the instantaneous boundary and loading conditions.This indicates that the critical SERR cannot be regarded as anintrinsic property of a viscoelastic interface. These results alsothe discussion on the specific treatment of the crack initiationand propagation regimes.<br /
Multi-omics data integration in constraint-based modeling of metabolic networks to study the metabolism of adipose-derived stem cells
No full textInternational audienceBackground: Genome-scale metabolic networks (GSMNs) are models representing all known metabolic processes occurring within a given organism as an interconnected network of metabolites, biochemical reactions, enzymes, and enzyme-coding genes. Through the integration of experimental data and the use of constraint-based modeling algorithms, these models can be used to simulate the metabolism of cells in various experimental conditions. Within the scope of the ASCending project, we aim to study the metabolic changes occurring in adipose-derived stem cells (ASCs) obtained from different donors over the course of a cell culture process. The final goal of this project is to adjust the culture parameters in order to optimize the proliferation and potency of the ASCs for the production of cellular therapy to treat medical conditions such as Alzheimer's disease. Results: For the purpose of this project, we used the previously published DEXOM algorithm [1] for con-straint-based modeling of metabolic networks, which was integrated into the larger OCMMED work-flow [2]. This workflow, which was originally designed for the reconstruction of cell-specific metabolic networks based on transcriptomics data, was adapted in order to add constraints on the model derived from the cell population doubling times, cell culture medium composition, and time-series exometabo-lomics data. We used these constraints to reconstruct metabolic networks for the cells in each experi-mental condition. We then examined the metabolic changes occurring over the course of the cell cul-ture process by comparing the contents of the different metabolic networks using both network-based approaches and methods for binary matrix comparisons.Conclusions: The simultaneous inclusion of different types of experimental data in our constraint-based modeling workflow presents a challenge, both due to the necessity of adapting our data processing strategy, and due to the stark increase in the number of constraints of differing natures on the meta-bolic model. However, these additional constraints allow for more precise modeling of cellular metabo-lism in various experimental conditions and better comparisons between different experimental condi-tions. With the adapted OCMMED workflow, we generate cell-specific models with which we can ex-amine the metabolic changes undergone by ASCs during their cell culture process, and which can then be used to predict metabolic reaction fluxes in simulated new medium conditions.References1.Rodríguez-Mier P, Poupin N, de Blasio C, Le Cam L, Jourdan F. DEXOM: Diversity-based enumeration of optimal context-specific metabolic networks. PLoS Comput Biol 2021;17:e1008730.2.Available from: https://forgemia.inra.fr/metexplore/cbm/ocmme
The long-term evolution of razor-thin galactic discs: Balescu–Lenard prediction and perspectives
No full textInternational audienceIn the last five decades, numerical simulations have provided invaluable insights into the evolution of galactic discs over cosmic times. As a complementary approach, developments in kinetic theory now also offer a theoretical framework to understand statistically their long-term evolution up to the onset of gravitational instability. The current state-of-the-art kinetic theory of isolated stellar systems is the inhomogeneous Balescu–Lenard equation. It can describe the long-term evolution of a self-gravitating razor-thin disc under the effect of resonant interactions between collectively amplified noise-driven fluctuations. In this work, comparing theoretical predictions to numerical simulations, we quantitatively show that kinetic theory indeed captures the average long-term evolution of cold stellar discs. Leveraging the versatility of kinetic methods, we then offer some new perspectives on this problem, namely (i) the crucial impact of collective effects in accelerating the relaxation; (ii) the role of (weakly) damped modes in shaping the disc’s orbital heating; (iii) the bias introduced by gravitational softening on long timescales when compared with non-softened theoretical predictions; (iv) the resurgence of strong stochasticity near marginal stability. These elements call for an appropriate choice of softening kernel when simulating the long-term evolution of razor-thin discs and for an extension of kinetic theory beyond the average evolution. Nevertheless, kinetic theory captures quantitatively the ensemble-averaged long-term response of such discs