Portail HAL de l'Institut Polytechnique de Paris
Not a member yet
28935 research outputs found
Sort by
Machine-learning techniques for model-independent searches in dijet final states
No full textInternational audienceAnomaly detection methods used in a recent search for new phenomena by CMS at the CERN LHC are presented. The methods use machine learning to detect anomalous jets produced in the decay of new massive particles. The effectiveness of these approaches in enhancing sensitivity to various signals is studied and compared using data collected in proton-proton collisions at a center-of-mass energy of 13 TeV. In an example analysis, the capabilities of anomaly detection methods are further demonstrated by identifying large-radius jets consistent with Lorentz-boosted hadronically decaying top quarks in a model-agnostic framework
Bayesian optimization for re-analysis and calibration of extreme sea state events simulated with a spectral third-generation wave model
No full textAccurate hindcasting of extreme sea state events is essential for coastal engineering, risk assessment, and climate studies. However, the reliability of numerical wave models remains limited by uncertainties in physical parameterizations and model inputs. This study presents a novel calibration framework based on Bayesian Optimization (BO), leveraging the Tree structured Parzen Estimator (TPE) to efficiently estimate uncertain sink term parameters, specifically bottom friction dissipation, depth induced breaking, and wave dissipation from strong opposing currents, in the ANEMOC-3 hindcast wave model. The proposed method enables joint optimization of continuous parameters and discrete model structures, significantly reducing discrepancies between model outputs and observations. Applied to a one month period encompassing multiple intense storm events along the French Atlantic coast, the calibrated model demonstrates improved agreement with buoy measurements, achieving lower bias, RMSE, and scatter index relative to the default sea-state solver configuration. The results highlight the potential of BO to automate and enhance wave model calibration, offering a scalable and flexible approach applicable to a wide range of geophysical modeling problems. Future extensions include multi-objective optimization, uncertainty quantification, and integration of additional observational datasets
A year-long observational analysis of atmospheric trace gases and particulate matter in Kathmandu
No full textInternational audienceKathmandu Valley is one of the most densely populated cities in Nepal, facing significant air quality challenges. This study presents a comprehensive analysis based on twelve months of continuous air quality measurement data collected at the Khumaltar Air Quality Monitoring Station (ICIMOD AQMS), which is located in the southern part of the Kathmandu Valley. The study investigates particulate matter (PM10, PM2.5 , PMC (PM10-PM2.5)) alongside trace gases (O3, NOx, SO2, and CO), focuses on their seasonal, diurnal variations, interspecies correlations, potential sources, and the impact of regional atmospheric transport. The results reveal that the annual PM2.5 concentration (49 µg m-3) largely exceeds the WHO air quality guideline of 5 µg m -3 . Pollutant concentrations show clear seasonal variation, with peaks during the winter and pre-monsoon seasons, and a noticeable decline during the monsoon. A strong correlation (r = 0.82, p = 0.00) between PM2.5 and NOx indicates that fossil fuel combustion is a dominant source of fine particulate pollution. During the pre-monsoon season, O3 concentrations occasionally exceed 100 ppb (8-hr running average), with seasonal diurnal concentration observed close to 90 ppb in the afternoon. The study found that temperature and relative humidity significantly influence coarse PM levels, with a strong negative correlation (r = -0.89, p = 0.00) between coarse PM and the relative humidity, suggesting less resuspension under moist conditions. Furthermore, this study highlights the severe air pollution in the Kathmandu Valley throughout the year, driven primarily by local emissions, with regional atmospheric transport further contributing to poor air qualit
Homogenizing elastic lattices with mechanisms
No full textInternational audienceWe propose an asymptotic method for homogenizing periodic elastic latticesthat works in the presence of mechanisms, both of the macroscopic type(strain-producing modes) and of the microscopic type (internal modes). When amicroscopic mechanism is present, the unit-cell problem produced by classicalhomogenization is singular. It can be fixed by including the amplitude~ of the mechanism as an additional macroscopic degree offreedom (enrichment variable) contributing to the effective energy via itsgradient . When a macroscopic mechanism ispresent, homogenization delivers a degenerate effective energy at leadingorder, which can be regularized by accounting for the strain gradient. Weintroduce an asymptotic second-order homogenization scheme that integratesthese two features: it delivers an effective energy capturing both thestrain-gradient effect relevantto macroscopic mechanisms, and the regularization relevant to microscopic mechanisms, if any is present. Theversatility of the approach is illustrated with a selection of latticesdisplaying a variety of effective behaviors. It follows a unified pattern thatleads to a classification of these effective behaviors. Whereas the proceduredelivers known effective models for elastic lattices without mechanisms, itcan generate novel effective models for lattices possessing mechanisms
Accuracy of the Ensemble Kalman Filter in the Near-Linear Setting
No full textInternational audienceThe filtering distribution captures the statistics of the state of a dynamical system from partial and noisy observations. Classical particle filters provably approximate this distribution in quite general settings; however they behave poorly for high dimensional problems, suffering weight collapse. This issue is circumvented by the ensemble Kalman filter which is an equal-weight interacting particle system. However, this finite particle system is only proven to approximate the true filter in the linear Gaussian case. In practice, however, it is applied in much broader settings; as a result, establishing its approximation properties more generally is important. There has been recent progress in the theoretical analysis of the algorithm, establishing stability and error estimates in non-Gaussian settings, but the assumptions on the dynamics and observation models rule out the unbounded vector fields that arise in practice and the analysis applies only to the mean field limit of the ensemble Kalman filter. The present work establishes error bounds between the filtering distribution and the finite particle ensemble Kalman filter when the model exhibits linear growth
Search for new particles decaying into top quark-antiquark pairs in proton-proton collisions at = 13 TeV
No full textInternational audienceA search for new particles decaying to top quark-antiquark pairs is performed using proton-proton collision data at a centre-of-mass energy of 13 TeV. The data set recorded with the CMS detector between 2016 and 2018 is used, corresponding to an integrated luminosity of 138 fb. Final states with 0, 1, and 2 leptons are analyzed, covering all decay modes of the top quark-antiquark pairs. Heavy Z' bosons with relative widths of 1, 10, and 30% are excluded for masses in the ranges 0.44.8, 0.46.2, and 0.47.4 TeV, respectively. A KaluzaKlein gluon in the RandallSundrum model and a dark-matter mediator are excluded for masses between 0.55.5 and 1.04.2 TeV, respectively. These results set the most stringent limits to date for the considered models in the final state. In addition, in the two-Higgs-doublet models, upper limits are set on the coupling strength modifier for scalar and pseudoscalar Higgs bosons with relative widths of 2.5, 10, and 25% in the mass range of 0.51.0 TeV
A pan-European map of shallow aquifer transmissivity in crystalline headwater catchments inferred from wetland and stream networks
No full textInternational audienceGroundwater systems in headwater catchments are poorly represented at continental scales due to model resolution constraints and limited observations available to characterize the wide diversity of catchments. Yet, low-order headwater streams accounting for a major fraction of the global river network. This is particularly true in upland crystalline bedrock regions with dense drainage networks, where the lithology has long been considered impermeable, without aquifers, and thus has received limited hydrogeological attention.We present a new continental-scale assessment of effective transmissivity for 3,333 European crystalline headwater catchments (median ≈35 km²), underlain by unconfined, shallow hard-rock aquifers where subsurface-surface interactions strongly shape hydrological connectivity. Catchments including dams, glaciers, and extensive permafrost were excluded.The methodology represents lateral hillslope groundwater flow within shallow subsurface systems, capturing the spatial patterns of saturated areas at the catchment scale. This framework of physically based groundwater flow models enables steady-state simulation of perennial surface water networks (springs, streams, wetlands), whose length and structure are highly sensitive to shallow aquifer transmissivity (Abhervé et al., 2023). Transmissivity was inferred through optimization of simulated seepage areas against observed wetland and stream networks, using constant recharge estimates from an independent land surface model and assuming dominant superficial subsurface flow in the upper 50 m. Across all calibrated models, the simulated networks closely replicate the available European-scale extended wetland ecosystem layer and stream network from the EU-Hydro database.Estimated transmissivity ranges from 10⁻⁸ to 10⁻² m² s⁻¹ (mean ≈10⁻⁴ m² s⁻¹), with pronounced spatial variability across geological provinces, massifs, or regions sharing similar tectonic framework legacies. The broad transmissivity range demonstrates the method’s sensitivity and its ability to resolve catchment-scale effective hydraulic properties across diverse climatic, topographic, and geological contexts. Values are consistent with textbook estimates for the studied lithologies and with hydraulic test data (pumping and slug tests) from regional or global datasets. Both measurements and estimates follow a log-normal distribution. Hydraulic conductivity was also derived from transmissivity using independent aquifer thickness datasets, including global depth-to-bedrock and regolith thickness maps.Our results provide the first EUropean crystalline bedRock hydrogeological HEADwater map of transmissivitY (EURHEADY), explicitly accounting for groundwater flows at the catchment scale. All calibrated simulations are provided as a georeferenced dataset, complemented by physiographic, climatic, hydrologic, pedologic, geologic, hydrogeologic, and anthropogenic attributes. This approach addresses a critical gap in estimating hydrogeological properties, a long-standing challenge for the critical zone community, and opens new opportunities for large-scale hydro(geo)logical modeling with improved representation of groundwater contributions.Reference:Abhervé, R., Roques, C., Gauvain, A., Longuevergne, L., Louaisil, S., Aquilina, L., & de Dreuzy, J. (2023). Calibration of groundwater seepage against the spatial distribution of the stream network to assess catchment-scale hydraulic properties. Hydrology and Earth System Sciences, 27(17), 3221–3239. https://doi.org/10.5194/hess-27-3221-202
A POP ⋆ is Born: Formal Predictable Out-of-Order Processor Model
No full textModern processors, even at the mid-range level, include multi-level caches, pipelines with branch predictors, or Out-of-Order (OoO) execution. While these are essential for average-case performance, they also increase the complexity of worst-case execution time analysis. OoO execution, for instance, is prone to timing anomalies and, due to the lack of efficient abstractions, quickly leads to state-space explosion. Consequently, it remains highly challenging in the context of critical real-time systems.This work proposes the first generic approach to predictable OoO execution, which is formally modeled and proven in the F* language and experimentally evaluated through simulations in gem5. Performance is evaluated on different processor models for MiBench and Embench programs. The average slowdown for an ARM A710-like processor model amounts to about 18.3% due to an implementation particularity of gem5. Eliminating bias from this issue reduces the slowdown to only 8.8%-10.4%
First measurement of the decay-time-integrated asymmetry in decays
No full textInternational audienceA measurement of the flavour-untagged decay-time-integrated asymmetry in the flavour-specific decay , , is performed using proton-proton collision data collected by the LHCb experiment between 2016 and 2018 at a center-of-mass energy of , corresponding to a total integrated luminosity of . The asymmetry is measured in two meson decay modes, and . The combined result, , is consistent with the Standard Model expectation and provides a direct constraint on new physics in tree-level -hadron decays
First measurement of time-dependent violation in the flavor-changing neutral-current decay
No full textInternational audienceA flavor-tagged time-dependent analysis of decays is performed across the full dimuon mass range excluding the and resonance regions. The analysis uses proton-proton collision data collected by the LHCb experiment in 2011--2018 at center-of-mass energies of 7, 8 and 13TeV, corresponding to an integrated luminosity of 9. The CP violation parameters are determined to be and , where the first uncertainties are statistical and the second are systematic. The results are consistent with the Standard Model prediction. This is the first experimental study of time-dependent CP violation in processes