HAL-Polytechnique
Not a member yet
51406 research outputs found
Sort by
Oracle-Efficient Combinatorial Semi-Bandits
No full textInternational audienceWe study the combinatorial semi-bandit problem where an agent selects a subset of base arms and receives individual feedback. While this generalizes the classical multi-armed bandit and has broad applicability, its scalability is limited by the high cost of combinatorial optimization, requiring oracle queries at every round. To tackle this, we propose oracle-efficient frameworks that significantly reduce oracle calls while maintaining tight regret guarantees. For the worst-case linear reward setting, our algorithms achieve regret using only oracle queries. We also propose covariance-adaptive algorithms that leverage noise structure for improved regret, and extend our approach to general (non-linear) rewards. Overall, our methods reduce oracle usage from linear to (doubly) logarithmic in time, with strong theoretical guarantees
Nature of momentum- and orbital-dependent magnetic fluctuations in Sr2RuO4
No full textInternational audienc
Modeling Jovian Plasma‐Europa Interactions: Innovative Atmosphere and Ionosphere Depiction for JUICE Mission Insights
No full textInternational audienceThe JUpiter ICy moons Explorer (JUICE) mission, launched by the European Space Agency (ESA) in April 2023, aims to explore Jupiter and its icy moons, particularly focusing on Europa, Ganymede, and Callisto. This study uses the Latmos Hybrid Simulation (LatHyS) model to simulate Europa's plasma and field environment, emphasizing the upcoming JUICE flybys in July 2032. The LatHyS model, incorporating a detailed 3D exospheric model and a self‐consistent ionosphere, allows for comprehensive analysis of the moon‐plasma interactions at ion scales. Our simulations, validated against Galileo's E4 flyby data, demonstrate the model's accuracy in reproducing key features of the plasma environment and ionospheric dynamics. To characterize the system's response to neutral/ionospheric environment assumptions, we compare three simulations, with different neutral and ionosphere impacts, revealing the ionosphere's influence on the magnetic field intensity. Using an atmosphere derived from a planetary exosphere simulation model like EGM allows for the consideration of various asymmetries and multiple major neutral species, supporting further studies on ionospheric ion dynamics. Results highlight the complex interactions influenced by Europa's neutral and ionospheric conditions, providing insights for the anticipated JUICE observations. The measured signatures primarily depend on the interactions with the ionosphere along the spacecraft's trajectory and simulations show that a dense ionosphere deduced from radio occultation observations cannot reproduce the observed in situ signatures. The exosphere being a source for planetary ions, it was shown that it is significant to consider the spatial asymmetries in global interaction models in order to better account for its impact on the system
Why is the volatility of single stocks so much rougher than that of the S&P500?
No full textSoumis à "Quantitative Finance"The Nested factor model was introduced by Chicheportiche et al. in [24] to represent nonlinear correlations between stocks. Stock returns are explained by a standard factor model, but the (log)-volatilities of factors and residuals are themselves decomposed into factor modes, with a common dominant volatility mode affecting both market and sector factors but also residuals. Here, we consider the case of a single factor where the only dominant log-volatility mode is rough, with a Hurst exponent H ≃ 0.11 and the log-volatility residuals are "super-rough", with H ≃ 0. We demonstrate that such a construction naturally accounts for the somewhat surprising stylized fact reported by Wu et al. in [23], where it has been observed that the Hurst exponents of stock indexes are large compared to those of individual stocks. We propose a statistical procedure to estimate the Hurst factor exponent from the stock returns dynamics together with theoretical guarantees of its consistency. We demonstrate the effectiveness of our approach through numerical experiments and apply it to daily stock data from the S&P500 index. The estimated roughness exponents for both the factor and idiosyncratic components validate the assumptions underlying our mode
Improving satellite remote sensing estimates of the global terrestrial water cycle via neural network modeling
No full textInternational audienceSatellite remote sensing provides important observations of Earth’s water cycle, but combining different satellite datasets often fails to produce a balanced water budget, highlighting the errors and uncertainties in these observations. This study introduces a novel approach combining optimal interpolation with neural network modeling to improve global water cycle estimates. We first balance water budget components (precipitation, evapotranspiration, runoff, and water storage change) across 1,358 river basins using optimal interpolation. We then train neural networks to reproduce these results and extend them to ungaged basins. After validating the approach on 340 independent basins, we apply it globally to create calibrated water cycle estimates at 0.5°resolution. Our method significantly reduces water budget imbalances in validation basins, decreasing the mean imbalance from 11 to 0.03 mm/month and reducing its variance from 44 to 24 mm/month. The calibrated datasets perform particularly well when applied to estimating evapotranspiration via the water budget method, achieving accuracy comparable to state-of-the-art methods. This is particularly useful in regions without ground-based measurements, and has broad applications in water resources planning and management. This study helps identify where satellite datasets need correction and demonstrates the benefits of machine learning for studying the water cycle at the global scale
MIRRORED-Anims: Motion Inversion for Rig-space Retargeting to Obtain a Reliable Enlarged Dataset of Character Animations
No full textInternational audienceWe propose MIRRORED-Anims, a novel retargeting procedure for transferring motion between skinned humanoid characters of different morphologies. It is designed so as to mimic the strengths of the closed-source Mixamo's retargeting method, currently used as a standard to create motion databases and train all state-of-the-art learning-based retargeting methods, despite severe shortcomings (namely, a lack of character diversity and notable penetration artifacts).Taking inspiration from the toolsets of 3D animators, our retargeting algorithm relies on the control rigs used to manipulate skinned characters, by identifying and transferring controller values on predefined bone mechanisms. While producing motions which are closer to Mixamo's ground truth than any state-of-the-art learning-based technique, MIRRORED-Anims creates fewer penetration artifacts than observed in the Mixamo dataset, improving the perceived quality of the output. Moreover, motion can be retargeted in real-time to and from the SMPL body model, making it possible to leverage the large motion databases available in SMPL format for the retargeting task. Because it relies solely on transparent, explainable rig operations, MIRRORED-Anims can be used to generate ground-truth motions for any humanoid character, providing a reliable baseline for the future training of learning-based methods
Inclusive and differential measurements of the cross section and the cross section ratio in proton-proton collisions at = 13 TeV
No full textInternational audienceInclusive and differential cross section measurements of top quark pair () production in association with a photon () are performed as a function of lepton, photon, top quark, and kinematic observables, using data from proton-proton collisions at = 13 TeV, corresponding to an integrated luminosity of 138 fb. Events containing two leptons (electrons or muons) and a photon in the final state are considered. The fiducial cross section of is measured to be 137 8 fb, in a phase space including events with a high momentum, isolated photon. The fiducial cross section of is also measured to be 56 5 fb when considering only events where the photon is emitted in the production part of the process. Both measurements are in agreement with the theoretical predictions, of 126 19 fb and 57 5 fb, respectively. Differential measurements are performed at the particle and parton levels. Additionally, inclusive and differential ratios between the cross sections of and production are measured. The inclusive ratio is found to be 0.0133 0.0005, in agreement with the standard model prediction of 0.0127 0.0008. The top quark charge asymmetry in production is also measured to be 0.012 0.042, compatible with both the standard model prediction and with no asymmetry
A Framework to Attribute Tropical Multiscale Precipitation Extremes to Rain Event Morphology in Deep Convective Systems
No full textInternational audienceAbstract The different spatiotemporal scales used to calculate extreme precipitation intensities can lead to diverging interpretation when investigating their physical origin, impacts, and sensitivity to climate. Besides, the contribution of mesoscale convective systems (MCSs) to tropical precipitation extremes remains loosely quantified on various scales, in particular on kilometer scales. Here, we construct a framework to analyze the cooccurrence of extreme precipitation at km‐scale and 1° × 1 day scale to compare their properties in terms of precipitation morphology and regional predominance. Using a storm‐tracking algorithm, we contrast the occurrence and precipitation statistics for two types of convective systems across 10 global storm‐resolving models and one geostationary satellite product. We do not find a large statistical dependence between rain extremes on these two scales, and they occur in distinct regions. Heavy km‐scale events occur mostly over continents and 40% of them are produced by MCSs in observations. Their intensity is independent from the area of rain features. Conversely, heavy 1° × 1 day rain intensities are dependent on the area of rain features, and occur more frequently over oceans, and a third of these events are produced by MCSs. Overall, the transition from deep to MCSs connect extremes across both scales. Compared to observations, models consistently underestimate the precipitating surface and show large discrepancies in the contribution of convective systems to precipitation extremes at each scale. This diagnostic is a key criterion for evaluating the ability of global storm‐resolving models to represent how individual convective systems produce realistic heavy rain distributions
1D nanoporous membrane boosts the ionic conductivity of electrolytes.
No full textInternational audienc
IgG detection in human serum employing non-functionalized chromium doped zinc gallate nanoparticles
No full textInternational audienceChromium-doped zinc gallate (ZnGa2O4:Cr3 +) nanoparticles (ZGO) show promising potential for antigen immunodetection using persistent luminescence, thereby reducing autofluorescence interference. Recently, we have shown that ZGO prepared by hydrothermal treatment at 120°C for 24 h can be used for in vitro biodetection in simple media such as phosphate-buffered saline. In this study, we investigated the effect of the protocol used to synthesize these ZGO nanoparticles, using a hydrothermal treatment at 220°C for different durations (6 h, 12 h, and 24 h), followed by calcination at 500°C. The nanoparticle size determined by transmission electron microscopy after grinding and centrifugation was found to be around 15 nm. The persistent luminescence signal of the ZGO nanoparticles varied with the hydrothermal synthesis conditions. Moreover, in the presence of H2O2, these nanoparticles show a signal enhancement dependent on the hydrothermal duration, with a 12 h treatment showing the highest 8-fold luminescence increase in the presence of H2O2 produced by glucose oxidase mediated glucose degradation. Based on these results, these non-functionalized nanoparticles were successfully used to develop a persistent luminescence-based sandwich immunoassay for autofluorescence-free detection of antigens in undiluted human serum samples, using rabbit IgG as a model antigen. This study highlights the promising potential for biosensing applications of persistent ZGO nanophosphors for IgG detection in a complex medium (undiluted human serum), with a linear range from 1 ng mL−1 to 104 ng mL−1 and a limit of detection of 0.01 ng mL−1. The present optimization of ZGO nanophosphor synthesis offers promising prospects for medical diagnostics due to their increased sensitivity and ability to eliminate autofluorescence interference, as well as their ease of use, since no functionalization of the ZGO NPs is required before use