Scientific Publications of the University of Toulouse II Le Mirail
Not a member yet
92205 research outputs found
Sort by
Sampling Nonsmooth Log-Concave Densities: A Comparative Study of Primal-Dual Based Proposal Distributions
No full textInternational audienceSampling from a real-valued distribution, whose density is nonsmooth and log-concave, is a computational issue that often arises in Machine Learning and Statistics. Langevin-based Hastings-Metropolis methods were proposed: they extend the Unadjusted Langevin Algorithm by using proximal methods to define a smoothed version of the density of interest. We consider the case when these extensions do not apply: the involved proximal operators do not have closed forms and the density is defined on a subset of the real numbers. We derive new Gaussian proposal mechanisms in a Metropolis Adjusted Langevin Algorithm, which use first-order information about the density function. We numerically compare these strategies and discuss the benefits of a change of geometry. The gain in using partial updates instead of global parameter updates is also illustrated
Quel accès aux savoirs endogènes ? Le rôle des relations sociales par-delà des contextes Nord et Suds
No full textInternational audienceCet article explore le rôle des relations sociales dans l’accès aux savoirs endogènes qui favorisent la transition agroécologique en France et au Bénin. Il vise à articuler la littérature sur les savoirs endogènes et celle sur les relations sociales, nous permettant de mettre en évidence les connexions entre ces notions. L’étude utilise une analyse quantitative de données d’entretiens réalisés dans ces deux territoires. Cette analyse montre que les agriculteurs et les collectifs mobilisent une diversité de savoirs endogènes qui contribuent à des degrés variés à la transition agroécologique. Trois profils d’agriculteurs ont été identifiés, soulignant tous l’importance des relations familiales, professionnelles et communautaires dans l’accès à ces savoirs. L’article illustre par ailleurs que, malgré des contextes différents, certains savoirs, identifiés comme des leviers de transition agroécologique se retrouvent de manière similaire dans les deux contextes. Le rôle des relations sociales pour accéder à ces savoirs dans les deux contextes est aussi mis en évidence. Enfin, l’article souligne l’intérêt des méthodes mixtes pour analyser les dynamiques de transition variées
Sharp quantitative stability of the M\"obius group among sphere-valued maps in arbitrary dimension
No full text23 pagesInternational audienceIn this work we prove a sharp quantitative form of Liouville's theorem, which asserts that, for all , the weakly conformal maps of with degree are M\"obius transformations. In the case this estimate was first obtained by Bernand-Mantel, Muratov and Simon (Arch. Ration. Mech. Anal. 239(1):219-299, 2021), with different proofs given later on by Topping, and by Hirsch and the third author. The higher-dimensional case requires new arguments because it is genuinely nonlinear: the linearized version of the estimate involves quantities which cannot control the distance to M\"obius transformations in the conformally invariant Sobolev norm. Our main tool to circumvent this difficulty is an inequality introduced by Figalli and Zhang in their proof of a sharp stability estimate for the Sobolev inequality
Social interaction dynamics of exiled Syrian adults in France: Insights from the SOFRA and CLASS projects
No full textInternational audienc
Derivation and numerical resolution of 2D shallow water equations for multi-regime flows of Herschel–Bulkley fluids
No full textInternational audienc
Covariance parameter estimation of Gaussian processes with approximated functional inputs
No full textInternational audienc
Le profil des étudiants de l'albigeois, Étude réalisée pour la Communauté d'agglomération de l'albigeois (C2A)
No full textLes territoires culturels vécus des jeunes. Spatialiser la diversité des ancrages et des pratiques dans la petite ville de Foix.
No full textInternational audienc
Hierarchical Bayesian Estimation of COVID-19 Reproduction Number
No full textInternational audienceAssessing the intensity of a epidemic, such as the COVID-19 pandemic, during the epidemic outbreak, constitutes a significant technical challenge with high societal stakes. Elaborating on classical epidemiological models, this work aims to define a hierarchical Bayesian model that permits the robust estimation of the temporal evolution of the pandemic intensity despite highly corrupted daily new infection counts. It also outputs uncertainty assessment, in the form of credibility intervals robust to the priors choice, accounting for uncertainties on model parameters. The estimation is performed by carefully designed Monte Carlo samplers. The relevance of the proposed estimation procedure is illustrated on real COVID-19 pandemic data for several countries and periods, made available from the Johns Hopkins University repository