Scientific Publications of the University of Toulouse II Le Mirail
Not a member yet
92205 research outputs found
Sort by
Limit theorems for non linear (compound marked) Hawkes processes
No full textIn this article, we fill a gap in the literature on Hawkes processes. In particular, we derive a CLT for a non linear compound marked Hawkes process. We also provide an upper bound on the convergence rate using the functional 1-Wasserstein distance. This result is obtained by discretizing the time line and reducing the problem to the quantification of the distance between finite marginal vectors, as well as between the discretized process and the original one.</div
« Portraits de divas italiennes : Entre apparition et effacement (1560-fin xviiie siècle) »: Scènes émancipatrices, Clio. Femmes, Genre, Histoire, 2025/2 n° 62, 2025. p.41-65.
No full textInternational audienceLe texte explore l’effacement iconographique des comédiennes italiennes entre le xvie et le xviiie siècle malgré leur forte présence scénique. Si certaines comme Isabella Andreini (1562-1604) ont construit activement leur image, la plupart restent anonymes ou absentes des représentations picturales. Cette invisibilité résulte de facteurs religieux (Contre-Réforme), sociaux (misogynie), mais aussi artistiques (refus de promouvoir les comédiennes comme figures publiques). Le texte distingue sources directes (contrôlées par les artistes ou actrices) et indirectes (objets décoratifs, gravures), révélant une tendance à sexualiser ou figer les actrices. Au xviiie siècle, malgré la documentation littéraire de leurs rôles (notamment chez Goldoni), aucune iconographie ne leur survit. Cette absence de portraits souligne le paradoxe entre la visibilité scénique des actrices et leur effacement pictural. Il semblerait que l’image muette de la diva contredise la puissance expressive de son jeu, fondé sur la palette des passions. L’iconographie relative aux comédiennes dell’arte reproduit un théâtre figé, alors que leur répertoire était né avec la scène de folie fondatrice d’Isabella Andreini (La pazzia d’Isabella, Florence, 1589)
Évolution historique de la carte universitaire toulousaine de 1808 à 2024 : ruptures institutionnelles et spatiales.
No full textInternational audienc
Postdoc report (PO2.1): Détection dans l'Internet des Objets
No full textLe postdoc 2.1 du PEPR Superviz vise à travailler sur la détection d’intrusion dans le cadre de réseaux composés d’objets connectés. Ce travail comporte deux volets principaux, correspondant à deux tâches T1 et T2.— Tâche T1 : compléter les travaux d’une précédente thèse réalisée au LAAS-CNRS sur la détection d’intrusion embarquée dans des contrôleurs Bluetooth Low Energy (BLE).— Tâche T2 : proposer une approche alternative aux travaux précédents en considérant une implémentation des mécanismes de détection d’intrusion dans une pile BLE en source libre
Block-Additive Gaussian Processes under Monotonicity Constraints
No full textInternational audienceWe generalize the additive constrained Gaussian process framework to handle interactions between input variables while enforcing monotonicity constraints everywhere on the input space. The block-additive structure of the model is particularly suitable in the presence of interactions, while maintaining tractable computations. In addition, we develop a sequential algorithm, MaxMod, for model selection (i.e., the choice of the active input variables and of the blocks). We speed up our implementations through efficient matrix computations and thanks to explicit expressions of criteria involved in MaxMod. The performance and scalability of our methodology are showcased with several numerical examples in dimensions up to 120, as well as in a 5D real-world coastal flooding application, where interpretability is enhanced by the selection of the blocks
MPI malleability validation under replayed real-world HPC conditions
No full textInternational audienceDynamic Resource Management (DRM) techniques can be leveraged to maximize throughput and resource utilization in computational clusters. Although DRM has been extensively studied through analytical workloads and simulations, skepticism persists among end administrators and users regarding their feasibility under realworld conditions. To address this problem, we propose a novel methodology for validating DRM techniques, such as malleability, in realistic scenarios that reproduce actual cluster conditions of jobs and users by replaying workload logs on a High-performance Computing (HPC) infrastructure. Our methodology is capable of adapting the workload to the target cluster. We evaluate our methodology in a malleability-enabled 125-node partition of the Marenostrum 5 supercomputer. Our results validate the proposed method and assess the benefits of MPI malleability on a novel use case of a pioneer user of malleability (our "PhD Student"): parallel-efficiency-aware malleability reduced a malleable workload time by 27 % without delaying the baseline workload, although introducing queueing delays for individual jobs, but maintaining the resource utilization rate
Assignments and Strategies in Democratic Forking Analysed with
No full textInternational audienceWe consider a recent game theory paper, on democratic forking: suppose you want to go to the opera with a group of 10 friends, with a choice of two operas. Everyone has a preference for one of the operas, but everyone also prefers a large community over a small one. Is there an assignment of each friend to one opera that is stable? The paper answers the question affirmatively giving an algorithm, and also contains results concerning uniqueness of stable assignments and strategyproofness.In previous work, we have formalised the main result of that paper in the proof assistant Isabelle. This provides proofs of the results that are much more reliable, since computer-checked, than paper-and-pencil proofs.Here we continue this formalisation work: we prove that there is a unique stable assignment if voters are sufficiently loyal to some alternative. We show under which conditions an assignment assigning all voters to the same alternative can exist. We refute a theorem stated by previous authors stating that if more than one stable assignment exists, then it is necessarily possible to manipulate the voting process. Finally, we prove that if there is a unique stable assignment, then no manipulation is possible
Mixed integer quadratically constrained quadratic programming for neural network Lipschitz constant computation
No full textInternational audienceTo ensure or certify the robustness of a neural network, its Lipschitz constant plays a prominent role. However, its calculation is NP-hard and one can only expect to bound or estimate it when computing time is limited. By taking into account activation regions at each ReLU neural network layer as new constraints, we propose new quadratically constrained Mixed Integer Programming (MIP) formulations for the neural network Lipschitz constant computation problem. The solutions of these problems provide lower bounds and upper bounds and we show that these bounds coincide with the Lipschitz constant almost everywhere in the parameter space of the neural network. Using various benchmark architectures and datasets from the literature, we run numerical comparisons of the proposed approach with a polynomial optimization technique, a mixed integer programming technique and a branch and bound method. The results show that, when the computing time budget is limited, the quadratically constrained MIP formulation achieves tighter bounds than the other methods for the L2 and L∞-norms