HAL-Polytechnique
Not a member yet
51406 research outputs found
Sort by
C. Favre - Dynamical degree 4: Summer School 2025 - Cremona Group
No full textDynamical degrees were introduced by Russakovskii and Shiffman in 1997. They measure the growth of preimages of subvarieties of a fixed dominant rational map of a projective variety, and control many dynamical features of this map. In this first set of lectures, we shall define dynamical degrees using positivity properties of divisor classes, and explain how to compute them in low dimension and in some significant examples
Spatialized and Prospective LCA of Agrivoltaism in the Mediterranean basin under Climate Change
No full textInternational audienc
Optimal business model adaptation plan for a company under a transition scenario
No full textClimate stress-tests aim at projecting the financial impacts of climate change, covering both transition and physical risks under given macro scenarios. However, in practice, transition risk has been the main focus of supervisory and academic exercises, and existing tools to downscale these macroeconomic projections to the firm level remain limited. We develop a methodology to downscale sector-level trajectories into firm-level projections for credit risk stress-tests. The approach combines probabilistic modeling with stochastic control to capture firm-level uncertainty and optimal decision-making. It can be applied to any transition scenario or sector and highlights how firm-level characteristics such as initialintensity, abatement cost, and exposure to uncertainty shape heterogeneous firm-level responses to the transition. The model explicitly incorporates firm-level business uncertainty through stochastic dynamics on relative emissions and sales, which affect both optimal decisions and resulting financial projections. Firms’ rational behavior is modeled as a stochastic minimization problem, solved numerically through a method we call Backward Sampling. Illustrating our method with the NGFS transition scenarios and three types of companies (Green, Brown and Average), we show that firm-specific intensity reduction strategies yield significantly different financial outcomes compared to assuming uniformsectoral decarbonisation rates. Moreover, investing an amount equivalent to the total carbon tax paid at a given date is limited by its lack of a forward-looking feature, making it insufficient to buffer against future carbon shocks in a disorderly transition. This highlights the importance of firm-level granularity in climate risk assessments. By explicitly modeling firm heterogeneity and optimal decision-making under uncertainty, our methodology complements existing approaches to granular transition risk assessment and contributes to the ongoing development of scenario-based credit risk projections at the firm level
Sept œuvres pour penser 2035 : polyphonie critique et désir de management
No full textInternational audienceThis volume brings together seven contributions from the 2023–2024 national examination jury in management science, each starting from a work of art—a film, painting, or song—to explore the transformations of management looking toward 2035. Rather than engaging in conventional foresight, the collection offers a polyphony of sensitive, critical, and unruly approaches. Each piece shows how art can open breaches in managerial routines, reveal tensions, and sketch out unexpected paths for strategic thinking. Far from closed models, this project embraces the idea of management as something to be reinvented—rooted in sensation and creation.Ce recueil réunit sept contributions issues du jury d'agrégation en sciences de gestion 2023-2024, chacune prenant pour point de départ ou d'arrivée une œuvre artistique - film, tableau, chanson - afin de penser, à travers elle, les transformations du management à l'horizon 2035. Plutôt qu'un exercice de prospective conventionnelle, l'ensemble propose une polyphonie d'approches sensibles, critiques et indisciplinées. Chaque intervention donne à voir comment une œuvre peut ouvrir une brèche dans les routines gestionnaires, dévoiler des tensions, et esquisser des chemins inattendus pour la pensée managériale. Loin des modèles fermés, ce projet fait le pari d’un management à réinventer, ancré dans le sensible et la création
The intersection of technological trajectories : The case of medical instrumentation and artificial intelligence
No full textInternational audienceThis article examines the interaction between the technological trajectories of artificial intelligence (AI) and medical instrumentation (MI) from 2010 to 2020. Using patent data, we created a cumulative graph of technological codes and measured the direction and intensity of the two fields’ interactions. The results reveal three things. First, there was a rapid expansion of AI/MI interaction, with a tendency toward a center/periphery structure in which AI technologies are dominant. Second, cognitive flows are highly asymmetrical, with AI transferring many more new combinations to MI than vice versa. Third, our prediction method had a very high success rate at the beginning of the decade; however, its advantage eroded as the structure solidified, reflecting the transition from a logic of exploration to a logic of exploitation
Évaluation automatique de textes rédigés par des humains et générés par des machines
No full textWith the rapid expansion of digital content, automatic evaluation of textual information has become increasingly crucial. This thesis addresses the challenge of evaluating and enhancing the quality of both human-written and machine-generated texts. Beginning with human-written texts, we develop an argument extraction system to evaluate the substantiation of scientific peer reviews, providing insights into the quality of peer review in NLP conferences in recent years. Additionally, we assess the factuality of abstractive summarization datasets and propose a data refinement approach that enhances model performance while reducing computational demands. For machine-generated texts, the thesis focuses on the underexplored aspects of diversity and naturalness. We introduce a suite of metrics for measuring linguistic diversity and conduct a systematic evaluation across state-of-the-art LLMs, exploring how development and deployment choices influence their output diversity. We also investigate the impact of training LLMs on synthetic data produced by earlier models, demonstrating that recursive training loops lead to diminishing diversity. Finally, we explore the naturalness of multilingual LLMs, uncovering an English-centric bias and proposing an alignment method to mitigate it. These contributions advance evaluation methodologies for natural language generation and provide insights into the interaction between evaluation metrics, dataset quality and language model performance.Avec l'expansion rapide du contenu numérique, l'évaluation automatique de l'information textuelle est devenue de plus en plus essentielle. Cette thèse aborde le défi d’évaluer et d’améliorer la qualité des textes, qu’ils soient rédigés par des humains ou générés par des modèles. Pour les textes rédigés par des humains, nous développons un système d’extraction d’arguments afin d’évaluer la justification des évaluations scientifiques par les pairs, offrant ainsi un aperçu de la qualité du processus de relecture dans les conférences de TAL au cours des dernières années. De plus, nous analysons la factualité des jeux de données de résumé abstratif et proposons une approche d'affinement des données qui améliore les performances des modèles tout en réduisant les coûts computationnels. Pour les textes générés par des modèles, cette thèse explore des aspects encore peu étudiés, tels que la diversité et la naturalité linguistique. Nous introduisons un ensemble de métriques permettant de mesurer la diversité linguistique et menons une évaluation systématique sur des modèles de langage de pointe, en examinant comment les choix de développement et de déploiement influencent la diversité de leurs sorties. Nous étudions également l'impact de l'entraînement des modèles sur des données synthétiques générées par des modèles antérieurs, démontrant que les boucles d'entraînement récursives entraînent une diminution progressive de la diversité. Enfin, nous analysons la naturalité des modèles de langage multilingues, mettant en évidence un biais centré sur l'anglais et proposant une méthode d’alignement pour y remédier. Ces contributions font progresser les méthodologies d’évaluation de la génération de langage naturel et apportent un éclairage sur l’interaction entre les métriques d’évaluation, la qualité des jeux de données et les performances des modèles de langage
From the Rose–DuBois ansatz of hotspot fields to the instanton solution: A pedestrian presentation
No full textInternational audienceThis paper gives a pedestrian presentation of some technical results recently published in mathematical physics with non-trivial implications for laser–plasma interaction. The aim is to get across the main results without going into the details of the calculations or offering a specialist's user guide but by focusing conceptually on how these results modify the commonly held description—in terms of laser hotspot fields—of backscattering instabilities with a spatially smoothed laser beam. The intended readers are plasma physicists as well as graduate students interested in laser–plasma interaction. No prior knowledge of scattering instabilities is required. Step by step, we explain how the laser hot spots are gradually replaced with other structures, called instantons, as the amplification of the scattered light increases. In the amplification range of interest for laser–plasma interaction, instanton–hotspot complexes tend to appear in the laser field (in addition to the expected hot spots), with a non-negligible probability. For even larger amplifications and systems longer than a hotspot length, the hotspot field description is clearly invalidated by the instanton takeover
Curvature penalization of strongly anisotropic interfaces models and their phase-field approximation
No full textThis paper studies the effect of anisotropy on sharp or diffuse interfaces models. When the surface tension is a convex function of the normal to the interface, the anisotropy is said to be weak. This usually ensures the lower semicontinuity of the associated energy. If, however, the surface tension depends on the normal in a nonconvex way, this so-called strong anisotropy may lead to instabilities related to the lack of lower semicontinuity of the functional. We investigate the regularizing effects of adding a higher order term of Willmore type to the energy. We consider two types of problems. The first one is an anisotropic nonconvex generalization of the perimeter, and the second one is an anisotropic nonconvex Mumford-Shah functional. In both cases, lower semicontinuity properties of the energies with respect to a natural mode of convergence are established, as well as Γ-convergence type results by means of a phase field approximation. In comparison with related results for curvature dependent energies, one of the original aspects of our work is that, in the context of free discontinuity problems, we are able to consider singular structures such as crack-tips or multiple junctions
Territoires zéro chômeur de longue durée : de l’expérimentation à l’institutionnalisation
No full textLe projet expérimental Territoires zéro chômeur de longue durée (TZCLD) représente au moment de cette publication plus de 6 000 personnes sorties de la privation d’emploi en France, 83 territoires expérimentaux, 92 entreprises à but d’emploi et plus d’une cinquantaine de territoires qui souhaitent le rejoindre. Il entre dans la phase finale d’expérimentation et la question sur son avenir après 2026 demeure une question principale. Que peut apporter ce projet en matière de politiques publiques pour l’emploi en France et en Europe où le projet commence à essaimer ? Cet ouvrage a voulu proposer une réflexion autour des tensions existantes entre le développement territorial d’une innovation sociale comme TZCLD et son institutionnalisation. Le passage à l’échelle comportant un risque de dévoiement du projet par la perte de ses caractéristiques originelles. À travers différentes contributions de chercheurs et d’acteurs présentés lors d’une journée d’étude dédiée, l’ouvrage cherche à identifier les modalités politiques, stratégiques, juridiques et organisationnelles de l’institutionnalisation ainsi que les marges de manœuvre des acteurs agissant en sa faveur de façon à dépasser ces tensions
Nascent water waves induced by the impulsive motion of a solid wall
No full textInternational audienceIn the present study, we investigated the generation phase of laboratory-scale water waves induced by the impulsive motion of a rigid piston, whose maximum velocity and total stroke are independently varied, as well as the initial liquid depth . By doing so, the influence of two dimensionless numbers is studied: the Froude number , with the gravitational acceleration, and the relative stroke of the piston. During the constant acceleration phase of the vertical wall, a transient water bump forms and remains localised in the vicinity of the piston, for all investigated parameters. Experiments with a small relative acceleration , where , are well captured by a first-order potential flow theory established by \citet{1990_joo}, which provides a fair estimate of the overall free surface elevation and the maximum wave amplitude reached at the contact with the piston. For large Froude numbers, an unsteady hydraulic jump theory is proposed, which accurately predicts the time evolution of the wave amplitude at the contact with the piston throughout the generation phase. At the end of the formation process, the dimensionless volume of the bump evolves linearly with and the wave aspect ratio is found to be governed by the relative acceleration . As the piston begins its constant deceleration, the water bump evolves into a propagating wave and several regimes are then reported and mapped in a phase diagram in the (, ) plane. While the transition from waves to water jets is observed if the typical acceleration of the piston is close enough to the gravitational acceleration , the wave regimes are found to be mainly selected by the relative piston stroke while the Froude number determines whether the generated wave breaks or not