Archive ouverte de Sorbonne Université
Not a member yet
324019 research outputs found
Sort by
Divergence-Free Diffusion Models for Incompressible Fluid Flows
No full textGenerative diffusion models are extensively used in unsupervised and self-supervised machine learning with the aim to generate new samples from a probability distribution estimated with a set of known samples. They have demonstrated impressive results in replicating dense, real-world contents such as images, musical pieces, or human languages. This work investigates their application to the numerical simulation of incompressible fluid flows, with a view toward incorporating physical constraints such as incompressibility in the probabilistic forecasting framework enabled by generative networks. For that purpose, we explore different conditional, score-based diffusion models where the divergence-free constraint is imposed by the Leray spectral projector, and autoregressive conditioning is aimed at stabilizing the forecasted flow snapshots at distant time horizons. The proposed models are run on a benchmark turbulence problem, namely a Kolmogorov flow, which allows for a fairly detailed analytical and numerical treatment and thus simplifies the evaluation of the numerical methods used to simulate it. Numerical experiments of increasing complexity are performed in order to compare the advantages and limitations of the diffusion models we have implemented and appraise their performances, including: (i) in-distribution assessment over the same time horizons and for similar physical conditions as the ones seen during training; (ii) rollout predictions over time horizons unseen during training; and (iii) out-of-distribution tests for forecasting flows markedly different from those seen during training. In particular, these results illustrate the ability of diffusion models to reproduce the main statistical characteristics of Kolmogorov turbulence in scenarios departing from the ones they were trained on
Light Dark Matter Search with 7.8 Tonne-Year of Ionization-Only Data in XENONnT
No full textInternational audienceWe report on a blinded search for dark matter (DM) using ionization-only (S2-only) signals in XENONnT with a total exposure of over 579 days in three science runs. Dedicated background suppression techniques and the first complete S2-only background model in XENONnT provide sensitivity to nuclear recoils of [0.5, 5.0] and electronic recoils of [0.04, 0.7] . No significant excess over the expected background is observed, and we set 90% confidence level upper limits on spin-independent DM--nucleon and spin-dependent DM--neutron scattering for DM masses between 3 and 8 , as well as on DM--electron scattering, axion-like particles, and dark photons, improving on previous constraints. For spin-independent DM--nucleon scattering, we exclude cross sections above cm at a DM mass of 5 , pushing the XENONnT sensitivity closer to the region where coherent elastic neutrino-nucleus scattering () becomes an irreducible background
Symbolically regressing dark matter halo profiles using weak lensing
No full textInternational audienceThe structure of dark matter haloes is often described by radial density profiles motivated by cosmological simulations. These are typically assumed to have a fixed functional form (e.g. NFW), with some free parameters that can be constrained with observations. However, relying on simulations has the disadvantage that the resulting profiles depend on the dark matter model and the baryonic physics implementation, which are highly uncertain. Instead, we present a method to constrain halo density profiles directly from observations. This is done using a symbolic regression algorithm called Exhaustive Symbolic Regression (ESR). ESR searches for the optimal analytic expression to fit data, combining both accuracy and simplicity. We apply ESR to a sample of 149 galaxy clusters from the HSC-XXL survey to identify which functional forms perform best across the entire sample of clusters. We identify density profiles that statistically outperform NFW under a minimum-description-length criterion. Within the radial range probed by the weak-lensing data ( h Mpc), the highest-ranked ESR profiles exhibit shallow inner behaviour and a maximum in the density profile. As a practical application, we show how the best-fitting ESR models can be used to obtain enclosed mass estimates. We find masses that are, on average, higher than those derived using NFW, highlighting a source of potential bias when assuming the wrong density profile. These results have important knock-on effects for analyses that utilise clusters, for example cosmological constraints on and from cluster abundance and clustering. Beyond the HSC dataset, the method is readily applicable to any data constraining the dark matter distribution in galaxies and galaxy clusters, such as other weak lensing surveys, galactic rotation curves, or complementary probes
The Physiology of Society’s Origins. Lucretius, De rerum natura Book V (925-1027)
No full textInternational audienceLucretius, De rerum natura Book V (925-1027) This article examines a pivotal section of Book V of Lucretius’ De Rerum Natura (v. 925–1027), arguing that the poem’s account of humanity’s social, political, and technological development is fully integrated into its broader physiological framework. The study proposes two interpretive hypotheses. First, it explores how the human body functions not merely as a passive object of description, but as an active narrative and philosophical instrument through which Lucretius articulates the conditions for social emergence. The analysis of v. 925–1010 highlights the strategic use of corporeal imagery to underscore the physiological and affective (pathē-related) processes driving early civilizational progress. Second, the article argues that the transition from a feral to a communal human condition (vv. 1011–1027) should be read as a rupture with the foedera naturae—the original natural laws—which is made possible by the uniquely human capacity for self-modification (Letter to Herodotus §75). This rupture, in turn, is interpreted through the lens of the clinamen (II, v. 225–260), Lucretius’ theory of atomic deviation, as an “anthropological swerve” that allows for the emergence of social life
In choro, ad occidentem spectans. Les peintures de Saint-Savin-sur-Gartempe d’un autre point de vue
No full textNational audienceThe abbey church of Saint-Savin-sur-Gartempe is renowned for its painted nave, a feature that has been widely debated but rarely examined in the context of the monastic enclosure. In the third bay, an internal partition – now only visible in a painted bestiary – defined the space reserved for monks, separating it from the area used by the laity. This threshold not only impacted the organization of the Old Testament cycle in the nave, but also how the monks perceived the paintings in the western part of the building when observing them from the middle of the nave. Taking this into account allows us to reconsider the distribution of the paintings and the function of the western porch tower, as well as focusing on certain little-studied paintings which are only visible when leaving after turning around to look west.L’abbatiale de Saint-Savin-sur-Gartempe est connue pour sa nef peinte, longuement commentée jusqu’à aujourd’hui, mais rarement étudiée en tenant compte de la clôture monastique. Dans la troisième travée, une partition interne, que seul un bestiaire peint matérialise aujourd’hui, définissait l’espace réservé aux moines, le séparant de celui des laïcs. Ce seuil de l’espace monastique a non seulement une incidence sur l’organisation du cycle vétérotestamentaire de la nef, mais aussi sur la perception des peintures de la partie occidentale de l’édifice par les moines qui les observent depuis le milieu de la nef. Sa prise en compte permet non seulement de renouveler la réflexion sur la répartition des peintures et la fonction de la tour-porche occidentale, mais aussi de s’intéresser à certaines peintures, peu étudiées, qui ne sont visibles qu’en sortant, après avoir effectué un demi-tour pour regarder en direction de l’ouest
Studying the repertory of Comédie-Italienne of Paris (1716-1779) throw the creation of a relational database
No full textInternational audienc
Euclid: Early Release Observations -- The extended stellar component of the IC10 dwarf galaxy
No full textInternational audienceWe present a detailed analysis of the old, extended stellar component of the Local Group dwarf galaxy IC 10 using deep resolved-star photometry in the VIS and NISP bands of the Euclid Early Release Observations. Leveraging Euclid's unique combination of wide field of view and high spatial resolution, we trace red giant branch (RGB) stars out to 8 kpc from the galaxy centre, reaching azimuthally-averaged surface brightness levels as faint as 29 mag arcsec. Our analysis reveals that IC 10's stellar distribution is significantly more extended than previously thought. After correcting for foreground extinction and subtracting contamination from Milky Way stars and background galaxies, we derive a radial stellar density profile from RGB star counts. The profile shows a marked flattening beyond 5 kpc, and is best fit by a two-component (Sersic + exponential) model, yielding a total stellar mass in old (age 1 Gyr) stars of . The origin of the outer stellar component is unclear. It may be accreted, even possibly associated with the counter-rotating HI gas in the outer regions of IC 10, or it may represent an ancient in-situ stellar halo. We tentatively detect two symmetric stellar overdensities at the edge of our imagery. These roughly align with the direction of IC 10's orbit around M31, suggesting that they may be signatures of tidal stripping. As part of our analysis, we derive a new distance to IC 10 based on the RGB tip, finding kpc and the distance modulus is
Lorentzian Path Integrals and Jackiw-Teitelboim wormholes with imaginary scalars
No full textInternational audienceThe Lorentzian path integral was recently used to argue that standard Euclidean axion wormholes do not dominate computations of connected AdS/CFT partition functions. We now apply similar methods to study the seemingly-analogous Jackiw-Teitelboim wormholes constructed by Garcia-Garcia and Godet using Jackiw-Teitelboim gravity with an imaginary-valued minimally-coupled massless scalar field. However, this time we find that these wormholes do dominate our path integral for the relevant connected partition function. This supports the suggestion by Garcia-Garcia and Godet that contributions from such wormholes parallel the physics of the Sachdev-Ye-Kitaev model at complex couplings. The result also illustrates the sensitivity of wormhole contributions to details of the relevant physics
A Censored Seasonal Trend Model for marine environmental time series with R-INLA
No full textInternational audienceIn the context of global change, the analysis of marine environmental time series is crucial for detecting trends, seasonality, and critical threshold exceedances, thereby supporting decision-making. However, these series often present quantification limits (i.e. censored data), irregular sampling, missing values, and multimodal seasonality. Substitution or ignoring censored values introduces significant biases in the subsequent analyses. This study introduces the Censored Trend Seasonal Model (CSTM), a state-space model specifically designed for the structural analysis of censored environmental time series. The CSTM combines a second-order random walk for the trend and trigonometric seasonality to decompose the series into trend, seasonality, and noise. The model assumes a log-normal distribution for observations and uses the R-INLA package for efficient Bayesian inference, enabling the direct integration of left-censored data without ad hoc imputation. Model fit is rigorously assessed using randomized quantile residuals (RQR), which normalizes the results in order to use classic goodness of fit tools. The CSTM was validated on both simulated data and a real-world dataset of \textit{Escherichia coli} concentrations in shellfish from the IFREMER's REMI monitoring network (France, 1997–2023). The results demonstrate the model's ability to accurately estimate concentrations --even below quantification limits-- and to detect trends and seasonal patterns. The CSTM's flexibility, computational efficiency, and robustness make it suitable for large-scale, automated analysis of environmental time series, providing valuable support for ecological research, regulatory monitoring, and public health decision-making. The model also opens new perspectives for the analysis of other censored environmental variables
Les formats d'encodage de la notation musicale à l'épreuve d'un objectif de transcription manuscrite automatique
No full textLe développement de la reconnaissance automatique de partitions (Optical Music Recognition, OMR) se heurte à une fragmentation des formats de représentation, partagée entre exigences éditoriales (MEI), impératifs de rendu graphique et sémantique (MusicXML) et analyse computationnelle (Humdrum **kern). Cette tension est particulièrement saillante pour les sources manuscrites, où la sémantique musicale se double d’une matérialité documentaire complexe. S’appuyant sur les récents progrès de l’ATR (Automatic Text Recognition), cet article évalue la capacité des standards actuels à servir de vérités de terrain. En prenant pour cas d’étude les manuscrits de piano de Claude Debussy, nous défendons une approche graphématique de la transcription. Nous justifions le détournement de MusicXML vers un modèle « structurel » capable de préserver, dans une certaine mesure, l’ancrage physique et les phénomènes non-standards de la source. Cette proposition vise à participer à la stabilisation des pratiques d’annotation pour la création de jeux de données mutualisables, condition essentielle au développement d’un OMR historique robuste