HAL-Polytechnique
Not a member yet
51406 research outputs found
Sort by
Analyse de sensibilité quantitative pour l'équation de Fokker-Planck par rapport à la distance de Wasserstein
No full textWe analyze the sensitivity of solutions to the Fokker-Planck equation with respect to some unknown parameter. Our main result is to provide quantitative upper bounds for the -Wasserstein distance between two solutions with different parameters, for every . We are able to give two proofs of this result, the first relying on synchronous coupling between two solutions of an SDE, and another one that relies on the differentiation of Kantorovitch dual formulation of optimal transport. We also provide more specific bounds in the case of the overdamped Langevin process, for which we are able to compare convergence to the invariant measure and sensitivity to the parameter.Nous analysons la sensibilité des solutions de l'équation de Fokker-Planck par rapport à un paramètre inconnu. Notre principal résultat est de fournir des bornes supérieures quantitatives pour la distance de Wasserstein d'ordre entre deux solutions correspondant à des paramètres différents. Nous donnons deux preuves de ce résultat, une première reposant sur un couplage synchrone entre deux solutions à une EDS, et une autre qui repose sur la différentiation de la formulation duale de Kantorovitch du transport optimal. Nous considérons aussi le processus de Langevin dans un second temps, pour lequel nous prouvons des bornes légerement différentes afin de comparer convergence vers une mesure invariante et sensibilité au paramètre
IPSL-Perm-LandN: improving the IPSL Earth System Model to represent permafrost carbon-nitrogen interactions
No full textInternational audienceAbstract. Permafrost soils have the potential to release large amounts of soil carbon to the atmosphere under climate change. However, in the Sixth Coupled Model Intercomparison Project (CMIP6), only two Earth System Models (ESM) represented permafrost carbon, both sharing the same land surface model. This makes future permafrost carbon dynamics highly uncertain and underscores the urgent need to include permafrost carbon in ESMs to enable more reliable future projections of climate change and remaining carbon budget estimates. Here, we present IPSL-Perm-LandN, an improved version of the Institut Pierre-Simon Laplace (IPSL) ESM (used for CMIP6) aiming at better representing high-latitude land ecosystems. The main developments are the inclusion of an explicit nitrogen cycle and of key permafrost physical and biogeochemical processes. The latent heat associated with soil water freeze/thaw is taken into account in the energy budget, as well as soil thermal insulation by soil organic matter and a surface organic layer (e.g. litter or moss). Soil organic carbon and nitrogen are vertically resolved with depth-dependent decomposition dynamics, a key feature for representing the effect of gradual permafrost thaw on soil biogeochemistry. Cryoturbation is represented as a diffusion process that buries organic matter in the deeper soil layers. Compared to the previous version of the model used for CMIP6, we show that the extent of the permafrost region has improved significantly and that the simulated active layer thickness in the Arctic is in better agreement with observations. Permafrost soil carbon stocks have increased 20-fold to reach 1006 PgC in the top 3 m of soil, which is consistent with observation-based estimates. We simulate that the permafrost region has been a net carbon sink over the past 150 years (+0.32 ± 0.04 PgC yr−1 on average between 2005 and 2014), primarily due to carbon uptake from boreal forests. This is comparable with recent pan-Arctic carbon balance estimates, when accounting for unrepresented processes in our model (fire and riverine carbon losses). Overall, the inclusion of permafrost processes has improved the response of the model to anthropogenic perturbations in high latitudes over the past century, marking a step forward in the representation of Arctic ecosystems
Asymmetric conformal prediction with penalized kernel sum-of-squares
No full textConformal prediction (CP) is a distribution-free method to construct reliable prediction intervals that has gained significant attention in recent years. Despite its success and various proposed extensions, a significant practical feature which has been overlooked in previous research is the potential skewed nature of the noise, or of the residuals when the predictive model exhibits bias. In this work, we leverage recent developments in CP to propose a new asymmetric procedure that bridges the gap between skewed and non-skewed noise distributions, while still maintaining adaptivity of the prediction intervals. We introduce a new statistical learning problem to construct adaptive and asymmetric prediction bands, with a unique feature based on a penalty which promotes symmetry: when its intensity varies, the intervals smoothly change from symmetric to asymmetric ones. This learning problem is based on reproducing kernel Hilbert spaces and the recently introduced kernel sum-of-squares framework. First, we establish representer theorems to make our problem tractable in practice, and derive dual formulations which are essential for scalability to larger datasets. Second, the intensity of the penalty is chosen using a novel data-driven method which automatically identifies the symmetric nature of the noise. We show that consenting to some asymmetry can let the learned prediction bands better adapt to small sample regimes or biased predictive models
SCAR : a self-consistent recurrent cell for real-time finite strain elastoplastic simulations
No full textComplex fabrication and forming processes operating under finite strains could benefit significantly from optimization loops of process parameters, which are often hindered by the prohibitive computational costs of process modeling. Neural networks present a promising solution to derive fast and accurate surrogate models, thereby enabling such optimizations. Furthermore, many processes involve substantial inherent variability, and hence often require manual process control. Neural networks could also provide real-time predictions that would greatly assist in decision-making. Although recursive neural networks have been applied in mechanics, their use in modeling elastoplastic behavior at finite strains remains underexplored. This paper introduces a new family of self-consistent recurrent cells, referred to as SCAR. These cells are specifically designed to address history-dependent problems, such as elastoplasticity, and ensure compliance with key properties required for such applications. To evaluate the SCAR cells, a generic architecture named PlastiNN, featuring a spatially resolved neural decoder, is employed. This approach results in faster training times and more accurate predictions in comparison to commonly used architectures. Additionally, PlastiNN can accommodate a series of successive loads on a workpiece, which is critical for most fabrication and forming processes. The effectiveness of this strategy is demonstrated by comparing SCAR cells to other recurrent cells within the PlastiNN architecture through a comprehensive benchmark including two datasets of 1D and 3D simulations, ranging from challenging toy applications to more realistic industrial test cases. Results highlight the superiority of the proposed recurrent neural network architecture for modeling elastic-plastic behavior at finite strains in engineering processes
Schwarzian quantum corrections to shear correlators of the near-extremal Reissner-Nordström-AdS black hole
No full textInternational audienceNear-AdS spacetimes are controlled by a Schwarzian effective dual theory. The Kaluza-Klein reduction of higher-dimensional black holes shows that the Schwarzian generates a logarithmic contribution to the entropy, thereby resolving a long-standing puzzle in near-extremal black hole thermodynamics. Here, we leverage exact results for quantum-corrected, Schwarzian scalar correlation functions in order to evaluate the impact of bulk quantum fluctuations on the low-temperature shear correlators of the state dual to Reissner-Nordström-AdS black holes with a flat, compact horizon. In the hydrodynamic regime, we find that quantum fluctuations tend to increase the shear viscosity away from , thereby preserving the Kovtun-Son-Starinets bound. Outside the hydrodynamic regime, quantum fluctuations lift the zero temperature, classical gapless modes reported in previous literature
Measurement of solar neutrino interaction rate below 3.49 MeV in Super-Kamiokande-IV
No full textInternational audienceSuper-Kamiokande has observed solar neutrino elastic scattering at recoil electron kinetic energies () as low as 3.49 MeV to study neutrino flavor conversion within the sun. At SK-observable energies, these conversions are dominated by the Mikheyev-Smirnov-Wolfenstein effect. An upturn in the electron neutrino survival probability in which vacuum neutrino oscillations become dominant is predicted to occur at lower energies, but radioactive background increases exponentially with decreasing energy. New machine learning approaches provide substantial background reduction below 3.49 MeV such that statistical extraction of solar neutrino interactions becomes feasible. This article presents an analysis of the solar neutrino interaction rate at < 3.49 MeV with the full SK-IV period, using data from a wideband intelligent trigger when available and with a boosted decision tree for event selection. A solar neutrino signal is observed between 2.99 MeV < < 3.49 MeV with significance and a data to unoscillated MC ratio of . This additional low energy data has a negligible effect on the intervals of the fits to the solar neutrino energy spectrum but has a noticeable effect on the best fit when using the exponential parameterization
Characterization of the quantum state of top quark pairs produced in proton-proton collisions at = 13 TeV using the beam and helicity bases
No full textInternational audienceMeasurements of the spin correlation coefficients in the beam basis are presented for top quark-antiquark () systems produced in proton-proton collisions at = 13 TeV collected by the CMS experiment in 20162018, and corresponding to an integrated luminosity of 138 fb. The system is reconstructed from final states containing an electron or muon, and jets. Together with the previously reported results in the helicity basis, these measurements are used to decompose the system into the Bell and spin eigenstates in various kinematic regions. The spin correlation coefficients are also used to evaluate properties of the quantum state, such as the purity, von Neumann entropy, and entanglement. All results are consistent with standard model predictions
Hard exclusive photoproduction of photon-meson pairs: pseudoscalar channels π, η and η'
No full textInternational audienc
Autoregressive Multiplier Bootstrap for In-situ Error Estimation and Quality Monitoring of Finite Time Averages in Turbulent Flow Simulations
No full textInternational audienceIn Computational Fluid Dynamics (CFD), and particularly within Direct Numerical Simulation (DNS) and Large Eddy Simulation (LES), the computational cost is largely dictated by the effort required to obtain statistically converged quantities such as time-averaged fields and higher-order moments. Despite the importance of accurately quantifying statistical uncertainty in unsteady simulations, no continuous and cost-effective, on-line method currently exists for monitoring the convergence quality of such statistics during runtime. This work introduces a novel, fully on-line bootstrapping approach to estimate the variance of finite-time averages without requiring the estimation of the flows Auto-Correlation Function (ACF). Unlike existing methods that rely on ACF estimation, which are often impractical due to excessive storage demands in large-scale simulations, or require off-line processing or a priori modeling assumptions, our method operates entirely during the simulation and incurs minimal overhead. The proposed technique employs a recursive update of bootstrap replicates of the time average, using correlated random weights generated via an autoregressive model. This formulation is computationally efficient: the update cost scales linearly with the number of bootstrap replicates and the dimensionality of the flow field, and the autoregressive model is inexpensive to evaluate. The method only requires storage of a small number of fields, making it suitable for large-scale CFD applications. We demonstrate the effectiveness of the approach on synthetic data from the Ornstein-Uhlenbeck process and on two canonical LES cases: a turbulent pipe flow and a round jet. We further discuss the methods applicability to simulations with non-uniform time stepping, highlighting its flexibility and robustness
Formation of gradients of atomic oxygen in nanosecond plasma for plasma-assisted detonation: experimental and numerical study
No full textInternational audienceThis work aims at producing a gradient of atomic oxygen on a scale of 10 cm in a plane-to-plane nanosecond discharge in 150 mbar of air with a varying gap size for applications in combustion and ignition of detonation waves. Local measurements of atomic oxygen density along the discharge span, at varying heights between high-voltage and grounded electrode, are performed with Xe calibrated O-TALIF and validated by 2D numerical modelling. They both show existence of a gradient of atomic density of oxygen along the span. Reduced electric field is measured with two experimental techniques: optical emission spectroscopy by a spectral band intensity ratio of the first negative system and the second positive system of nitrogen, and E-FISH. It is also compared with numerical modelling. All techniques show existence of a gradient of reduced electric field along the span. This distribution of reduced electric field, in combination with the non-uniform energy deposition in the plasma, is shown to explain the measured gradient of density of atomic oxygen