HAL-Polytechnique
Not a member yet
51406 research outputs found
Sort by
Probing early parton emissions in heavy ion collisions using the Lund jet plane
No full textInternational audienceIn scattering experiments, high-virtuality partons, i.e., quarks and gluons, initiate a series of additional parton emissions to create collimated sprays of particles known as jets. This paper presents a measurement of the Lund jet plane (LJP) of high-energy jets produced in lead-lead (PbPb) collisions and compares the results to data for proton-proton (pp) collisions. The LJP is formed by iteratively declustering the constituents of a jet into consecutive emissions and recording the relative transverse momentum () and angle of the resulting emission with respect to its emitter. The angular distributions of two different slices of the LJP are investigated for jets with radius parameter of 0.4 and transverse momentum in the range 2001000 GeV. The PbPb (pp) data were recorded by the CMS experiment in 2018 (2017) and correspond to an integrated luminosity of 1.7 nb (301 pb) at a nucleon-nucleon center-of-mass energy of 5.02 TeV. The measurement was designed to test whether the earliest jet emissions are produced before the formation of the quark-gluon plasma (QGP) in PbPb collisions. Within the experimental uncertainties, no significant difference is observed between the angular distribution of high- emissions in \pp and PbPb collisions, which is consistent with these emissions occurring early in the jet evolution, before substantial interaction with the QGP
Bipartite Independent Set Reconfiguration: General and RNA-Inspired Parameterized Algorithms
No full textInternational audienceThe computation of energy barriers between RNA structures is a classic NP-hard problem in Bioinformatics, of which the Independent Set (IS) reconfiguration in bipartite graphs represents a natural generalization. Parameterized algorithms, based on parameters taking limited or bounded values on biological instances, are thus crucial towards practical solutions. In this work, we show that bipartite IS reconfiguration is slice-wise Polynomial (XP) solvable for both the {range} of IS sizes allowed along the reconfiguration, and the {arboricity} when the input is restricted to a circle graph. Such a setting is relevant to Bioinformatics as it provides a solution to the RNA energy barrier problem.We propose algorithms based on a {divide-and-conquer} approach, yielding a -space, -time algorithm for the range , and a -space, -time algorithm for the arboricity .Then, we demonstrate the practicality of these algorithms on a benchmark consisting of random RNA instances
Linear Bandits beyond Inner Product Spaces, the case of Bandit Optimal Transport
No full textLinear bandits have long been a central topic in online learning, with applications ranging from recommendation systems to adaptive clinical trials. Their general learnability has been established when the objective is to minimise the inner product between a cost parameter and the decision variable. While this is highly general, this reliance on an inner product structure belies the name of \emph{linear} bandits, and fails to account for problems such as Optimal Transport. Using the Kantorovich formulation of Optimal Transport as an example, we show that an inner product structure is \emph{not} necessary to achieve efficient learning in linear bandits. We propose a refinement of the classical OFUL algorithm that operates by embedding the action set into a Hilbertian subspace, where confidence sets can be built via least-squares estimation. Actions are then constrained to this subspace by penalising optimism. The analysis is completed by leveraging convergence results from penalised (entropic) transport to the Kantorovich problem. Up to this approximation term, the resulting algorithm achieves the same trajectorial regret upper bounds as the OFUL algorithm, which we turn into worst-case regret using functional regression techniques. Its regret interpolates between and , depending on the regularity of the cost function, and recovers the parametric rate in finite-dimensional settings
Unbiased Approximate Vector-Jacobian Products for Efficient Backpropagation
No full textIn this work we introduce methods to reduce the computational and memory costs of training deep neural networks. Our approach consists in replacing exact vector-jacobian products by randomized, unbiased approximations thereof during backpropagation. We provide a theoretical analysis of the trade-off between the number of epochs needed to achieve a target precision and the cost reduction for each epoch. We then identify specific unbiased estimates of vector-jacobian products for which we establish desirable optimality properties of minimal variance under sparsity constraints. Finally we provide in-depth experiments on multi-layer perceptrons, BagNets and Visual Transfomers architectures. These validate our theoretical results, and confirm the potential of our proposed unbiased randomized backpropagation approach for reducing the cost of deep learning
The Path to the Goal : Feature-augmented Hawkes Processes for Event-levelAttribution
No full textAttribution, the problem of assigning proportional responsibility for an outcome to each event in a temporal sequence, is central to diverse applications ranging from marketing and seismology to sports analytics. While incorporating exogenous features substantially enhances the expressiveness of attribution models, existing approaches lack the scalability required to integrate modern machine learning. We introduce FeatHawkes, a feature-augmented Hawkes process framework for event-level attribution in continuous time. Our core contribution is a novel first-order optimization routine for Hawkes processes that leverages stochastic gradient methods, scaling favorably with both dataset size and feature dimensionality. This gradient-based formulation enables compatibility with automatic differentiation and end-to-end machine learning pipelines. We release FeatHawkes as an open-source Python library and demonstrate its effectiveness through synthetic experiments and a case study on professional football data, where the framework supports scenario-based analyses, such as evaluating the modelled effect of player substitutions in a lineup
M-theory on as Type 0A
No full textInternational audienceWe propose an exotic geometric M-theory dual for the weak coupling Type 0A string: compactification on a sub-Planckian (two circles connected at a point), where strong quantum effects lead to fields living on distinct resolutions of that space. Moreover we argue that tachyon condensation of the 0A theory corresponds to shrinking of one of the two circles leading to the IIA supersymmetric string. We use this and other dualities to provide an F-theoretic description of the axio-dilaton and the tachyonic field of Type 0B and argue for the existence of a strong coupling critical point of the potential using the resulting duality symmetry . The existence of this critical point can also be argued using conventional M-theory dualities. If this critical point is unique it is an unstable dS vacuum. Using this we propose a strong coupling conformal fixed point for a non-supersymmetric gauge theory in four dimensions living on coincident branes of 0B
Measurement of the -channel single top quark cross section in proton-proton collisions at = 5.02 TeV
No full textInternational audienceThe single top quark -channel production cross section is measured in proton-proton collisions at the CERN LHC at = 5.02 TeV, using data recorded with the CMS detector in 2017, corresponding to an integrated luminosity of 302 pb, and resulting in the first CMS measurement of the process at that energy. Events with one electron or muon and two or more jets, among which at least one is identified as originating from a b quark fragmentation, are analyzed. The combined cross section of single top quark (tq) and single top antiquark () production is = 25.4 (stat) (syst) 0.5 (lumi) pb. The individual cross sections are measured to be = 17.6 (stat) (syst) 0.3 (lumi) pb and = 6.6 (stat) (syst) 0.1 (lumi) pb. Their ratio is measured to be = 2.7 (stat) (syst). The absolute value of the CabibboKobayashiMaskawa matrix element is found to be = 0.92 0.09 (exp) 0.01 (thy). The measurements are in good agreement with the standard model predictions at next-to-next-to-leading order accuracy in quantum chromodynamics
Rapid estimation of microstructure using infrared imaging and solidification modeling in wire-laser directed energy deposition
No full textInternational audienceThe widespread deployment of Directed Energy Deposition Additive Manufacturing is limited by the lack of control over the produced material depending on process parameters: in particular, the microstructure resulting from rapid solidification. While cost-efficient numerical simulations have been developed to predict temperature evolution and microstructure, their reliability hinges on high-quality experimental validation. This study first addresses this challenge by introducing a simple and cost-effective infrared measurement procedure that combines a single-band camera and a dual-band pyrometer to quantitatively measure temperature fields during wire-laser DED. To do so, the apparent emissivity field was identified and found to be highly heterogeneous due to localized cover gas and oxidation. In addition, to enable rapid microstructure estimation, fast computational procedures are proposed to (i) calculate the thermal gradient field using Fast Fourier Transform, and (ii) simulate solidification in the melt pool, including competitive growth between columnar dendritic grains, using a recent Voronoi tessellation-based model. The computation time is compatible with the development of an online monitoring procedure. The resulting microstructure predictions were validated against Electron Backscatter Diffraction measurements, demonstrating excellent qualitative agreement. This work validates the proposed approach as a promising tool for closed-loop control of microstructure during DED
A stationarity principle generating effective boundary conditions for second-order homogenization
No full textInternational audienceWe derive an effective model for a periodic chain of linearly-elastic springs,achieving second-order accuracy in the scale separation parameter . The chain has finite length and is made up of springs connecting bothnearest- and next-nearest-neighbors: it serves as a one-dimensional prototypefor higher-order periodic homogenization problems with boundaries. This typeof problem has been approached by inserting two-scale expansions into theequations of equilibrium in the bulk and by matching them with boundary-layersolutions. We explore an alternative method operating at the energy level,bypassing the cumbersome matching procedure. We start from an ansatz of themicroscopic displacement accounting for both boundary layers and forsmall-scale fluctuations in the bulk, and insert it into the discrete energy.This yields a continuous energy functional depending on the macroscopicdisplacement , in the form of a series expansion in powers of. We call it a {\tmem{pseudo-energy}} asit is not positive when truncated at order~. The boundary termsin the pseudo-energy account for boundary layers in an effective way. Bymaking the pseudo-energy stationary order by order in , we derivethe homogenized equations of equilibrium along with effective boundaryconditions. We provide quantitative validation showing that the effectivemodel is correct to second order. We point out the special form of theeffective higher-order tractions, which has been overlooked in strain-gradienttheories proposed so far
Measurements of the inclusive W and Z boson production cross sections and their ratios in proton-proton collisions at = 13.6 TeV
No full textInternational audienceMeasurements are presented of the W and Z boson production cross sections in proton-proton collisions at a center-of-mass energy of 13.6 TeV. Data collected in 2022 and corresponding to an integrated luminosity of 5.01 fb with one or two identified muons in the final state are analyzed. The results for the products of total inclusive cross sections and branching fractions for muonic decays of W and Z bosons are 11.93 0.08 (syst) 0.17 (lumi) (acc) nb for W boson production, 8.86 0.06 (syst) 0.12 (lumi) (acc) nb for W boson production, and 2.021 0.009 (syst) 0.028 (lumi) (acc) nb for the Z boson production in the dimuon mass range of 60-120 GeV, all with negligible statistical uncertainties. Furthermore, the corresponding fiducial cross sections, as well as cross section ratios for both fiducial and total phase space, are provided. The ratios include charge-separated results for W boson production (W and W) and the sum of the two contributions (W), each relative to the measured Z boson production cross section. Additionally, the ratio of the measured cross sections for W and W boson production is reported. All measurements are in agreement with theoretical predictions, calculated at next-to-next-to-leading order accuracy in quantum chromodynamics