-
$ψ$(2S) production in Pb-Pb collisions measured by ALICE at the LHC
Authors:
Victor Feuillard
Abstract:
Charmonium production is a probe sensitive to deconfinement in nucleus-nucleus collisions. The production of J/$ψ$ via regeneration within the QGP or at the phase boundary has been identified as an important ingredient for the description of the observed J/$ψ$ nuclear modification factor at the LHC. The $ψ(2S)$ production relative to J/$ψ$ is a possible discriminator between the two regeneration s…
▽ More
Charmonium production is a probe sensitive to deconfinement in nucleus-nucleus collisions. The production of J/$ψ$ via regeneration within the QGP or at the phase boundary has been identified as an important ingredient for the description of the observed J/$ψ$ nuclear modification factor at the LHC. The $ψ(2S)$ production relative to J/$ψ$ is a possible discriminator between the two regeneration scenarios. Studies of $ψ(2S)$ production in central nucleus-nucleus collisions at low transverse momentum ($p_{\rm T}$) are crucial, particularly at the LHC, where regeneration appears to be dominant. A significant $ψ$(2S) signal is extracted at low $p_{\rm T}$ and forward rapidity in the dimuon decay channel for the first time. This measurement relies on the $ψ(2S)$ cross section measured recently in pp collisions at $\sqrt{s_{\rm }}=5.02$ TeV with an unprecedented precision compared to previous ALICE results. In this contribution, we present newly published results on the $ψ(2S)$-to-J/$ψ$ ratio and the $ψ(2S)$ nuclear modification factor in Pb-Pb collisions at $\sqrt{s_{\rm NN}}=5.02$ TeV. Results are reported as a function of centrality and $p_{\rm T}$ for $p_{\rm T}<12$ GeV/c and are compared to available NA50 and CMS measurements. Comparisons to transport and statistical hadronization model predictions are also provided to shed light on the charmonium states recombination mechanism
△ Less
Submitted 15 September, 2023;
originally announced September 2023.
-
Overview of quarkonium production in ALICE
Authors:
Victor Feuillard
Abstract:
ALICE is an experiment mainly dedicated to the study of the quark-gluon plasma (QGP), a state of nuclear matter where quarks and gluons are deconfined, due to high temperature and density. The production of quarkonia, bound states of either a charm and anti-charm quark pair for charmonia or a bottom and anti-bottom quark pair for bottomonia, is a very useful probe of this state of matter. Indeed,…
▽ More
ALICE is an experiment mainly dedicated to the study of the quark-gluon plasma (QGP), a state of nuclear matter where quarks and gluons are deconfined, due to high temperature and density. The production of quarkonia, bound states of either a charm and anti-charm quark pair for charmonia or a bottom and anti-bottom quark pair for bottomonia, is a very useful probe of this state of matter. Indeed, the presence of the QGP modifies their production yields, due to a competition between medium-induced suppression and, at least for charmonia, a recombination mechanism in medium or at hadronization. The measurement of quarkonium production in proton-proton (pp) collisions is also essential to investigate its production mechanism as well as to test quantum chromodynamics (QCD) models. It also provides a reference for the investigation of the properties of the QGP in nucleus--nucleus collisions. Measurements in proton--nucleus collisions allow one to study the cold nuclear matter effects. In addition, high multiplicity pp collisions are useful to investigate multiparton interactions and to search for collectivity in small systems. In this contribution, we will report on the latest quarkonium production results, obtained with the ALICE detector in pp collisions, for different center-of-mass energies, at midrapidity and forward rapidity. We will also present the measurements of the quarkonium nuclear modification factor in p--Pb collisions at $\sqrt{s_{NN}}=8.16$~TeV. Finally, in Pb--Pb collisions at $\sqrt{s_{NN}}=5.02$~TeV, we will report on the new $ψ(2S)$ measurement, the nuclear modification factor of the prompt and non-prompt J/$ψ$, as well as the first measurement of quarkonium polarization, the latter carried out also as a function of the event plane. All the measurements will be compared to theoretical predictions.
△ Less
Submitted 22 December, 2022;
originally announced December 2022.
-
Measurement of quarkonium production in ALICE
Authors:
Victor Feuillard
Abstract:
ALICE is designated to study the quark-gluon plasma (QGP), a state of matter where, due to high temperature and density, quarks and gluons are deconfined. One of the probes used to investigate this state of matter is quarkonium states, bound states of either a charm and anti-charm quark pair (charmonia) or a bottom and anti-bottom quark pair (bottomonia). The presence of the QGP is expected to mod…
▽ More
ALICE is designated to study the quark-gluon plasma (QGP), a state of matter where, due to high temperature and density, quarks and gluons are deconfined. One of the probes used to investigate this state of matter is quarkonium states, bound states of either a charm and anti-charm quark pair (charmonia) or a bottom and anti-bottom quark pair (bottomonia). The presence of the QGP is expected to modify the quarkonium production yields in a very specific way due to a balance between medium-induced suppression, and a recombination mechanism or a hadronization mechanism. To understand the the properties of the QGP in nucleus-nucleus collisions, it is essential to measure the quarkonium differential yields in proton-proton collisions, as it provides a reference and allows the investigation of quarkonium production mechanisms, as well as in proton-nucleus collisions to understand the cold nuclear matter effects that appear. In this contribution, the latest results for quarkonium production measured with the ALICE detector in pp collisions at different collision energies are reported. The measurements of the nuclear modification factor and anisotropic flow in Pb-Pb collisions at $\sqrt{ s_{NN}} = 5.02$ TeV and in p-Pb at $\sqrt{ s_{NN}}=8 .16$ TeV at mid- and forward rapidity are also reported. All measurements are compared to various theoretical predictions.
△ Less
Submitted 14 January, 2021; v1 submitted 11 January, 2021;
originally announced January 2021.
-
A Multivariate Extreme Value Theory Approach to Anomaly Clustering and Visualization
Authors:
Maël Chiapino,
Stéphan Clémençon,
Vincent Feuillard,
Anne Sabourin
Abstract:
In a wide variety of situations, anomalies in the behaviour of a complex system, whose health is monitored through the observation of a random vector X = (X1,. .. , X d) valued in R d , correspond to the simultaneous occurrence of extreme values for certain subgroups $α$ $\subset$ {1,. .. , d} of variables Xj. Under the heavy-tail assumption, which is precisely appropriate for modeling these pheno…
▽ More
In a wide variety of situations, anomalies in the behaviour of a complex system, whose health is monitored through the observation of a random vector X = (X1,. .. , X d) valued in R d , correspond to the simultaneous occurrence of extreme values for certain subgroups $α$ $\subset$ {1,. .. , d} of variables Xj. Under the heavy-tail assumption, which is precisely appropriate for modeling these phenomena, statistical methods relying on multivariate extreme value theory have been developed in the past few years for identifying such events/subgroups. This paper exploits this approach much further by means of a novel mixture model that permits to describe the distribution of extremal observations and where the anomaly type $α$ is viewed as a latent variable. One may then take advantage of the model by assigning to any extreme point a posterior probability for each anomaly type $α$, defining implicitly a similarity measure between anomalies. It is explained at length how the latter permits to cluster extreme observations and obtain an informative planar representation of anomalies using standard graph-mining tools. The relevance and usefulness of the clustering and 2-d visual display thus designed is illustrated on simulated datasets and on real observations as well, in the aeronautics application domain.
△ Less
Submitted 17 July, 2019;
originally announced July 2019.
-
Anomaly detection with Wasserstein GAN
Authors:
Ilyass Haloui,
Jayant Sen Gupta,
Vincent Feuillard
Abstract:
Generative adversarial networks are a class of generative algorithms that have been widely used to produce state-of-the-art samples. In this paper, we investigate GAN to perform anomaly detection on time series dataset. In order to achieve this goal, a bibliography is made focusing on theoretical properties of GAN and GAN used for anomaly detection. A Wasserstein GAN has been chosen to learn the r…
▽ More
Generative adversarial networks are a class of generative algorithms that have been widely used to produce state-of-the-art samples. In this paper, we investigate GAN to perform anomaly detection on time series dataset. In order to achieve this goal, a bibliography is made focusing on theoretical properties of GAN and GAN used for anomaly detection. A Wasserstein GAN has been chosen to learn the representation of normal data distribution and a stacked encoder with the generator performs the anomaly detection. W-GAN with encoder seems to produce state of the art anomaly detection scores on MNIST dataset and we investigate its usage on multi-variate time series.
△ Less
Submitted 11 December, 2018; v1 submitted 6 December, 2018;
originally announced December 2018.
-
Charmonium production in Pb-Pb collisions measured by ALICE at the LHC
Authors:
Victor Feuillard
Abstract:
The ALICE experiment is dedicated to the study of the Quark-Gluon Plasma (QGP), a state of matter where, due to high temperature and density, quarks and gluons are deconfined. One of the probes studied to investigate this state of matter is the production of charmonium states, such as the J/$ψ$ and the $ψ$(2S). Indeed, the presence of the QGP is expected to modify the charmonium production yields,…
▽ More
The ALICE experiment is dedicated to the study of the Quark-Gluon Plasma (QGP), a state of matter where, due to high temperature and density, quarks and gluons are deconfined. One of the probes studied to investigate this state of matter is the production of charmonium states, such as the J/$ψ$ and the $ψ$(2S). Indeed, the presence of the QGP is expected to modify the charmonium production yields, due to a balance between the color screening of the charm quark potential and a recombination mechanism. A suppression of the production yields in heavy ion collisions with respect to pp collisions scaled by the mean number of binary nucleon-nucleon collisions was observed by ALICE in Pb-Pb collisions at a centre-of-mass energy per nucleon pair $\sqrt{ s_{NN}} = 2.76$ TeV. The observed suppression is smaller than the one found by PHENIX at an order of magnitude lower collision energy. This behavior can be explained by a stronger contribution from recombination processes at LHC than at lower energies. In this presentation, we report on new results for the charmonium production in Pb-Pb collisions measured at forward rapidity with the the muon spectrometer of ALICE at $\sqrt{ s_{NN}} =5.02$ TeV and their comparison with previous results and model predictions.
△ Less
Submitted 27 March, 2017; v1 submitted 24 March, 2017;
originally announced March 2017.
-
Calibration of One-Class SVM for MV set estimation
Authors:
Albert Thomas,
Vincent Feuillard,
Alexandre Gramfort
Abstract:
A general approach for anomaly detection or novelty detection consists in estimating high density regions or Minimum Volume (MV) sets. The One-Class Support Vector Machine (OCSVM) is a state-of-the-art algorithm for estimating such regions from high dimensional data. Yet it suffers from practical limitations. When applied to a limited number of samples it can lead to poor performance even when pic…
▽ More
A general approach for anomaly detection or novelty detection consists in estimating high density regions or Minimum Volume (MV) sets. The One-Class Support Vector Machine (OCSVM) is a state-of-the-art algorithm for estimating such regions from high dimensional data. Yet it suffers from practical limitations. When applied to a limited number of samples it can lead to poor performance even when picking the best hyperparameters. Moreover the solution of OCSVM is very sensitive to the selection of hyperparameters which makes it hard to optimize in an unsupervised setting. We present a new approach to estimate MV sets using the OCSVM with a different choice of the parameter controlling the proportion of outliers. The solution function of the OCSVM is learnt on a training set and the desired probability mass is obtained by adjusting the offset on a test set to prevent overfitting. Models learnt on different train/test splits are then aggregated to reduce the variance induced by such random splits. Our approach makes it possible to tune the hyperparameters automatically and obtain nested set estimates. Experimental results show that our approach outperforms the standard OCSVM formulation while suffering less from the curse of dimensionality than kernel density estimates. Results on actual data sets are also presented.
△ Less
Submitted 30 August, 2015;
originally announced August 2015.
-
Numerical studies of the metamodel fitting and validation processes
Authors:
Bertrand Iooss,
Loïc Boussouf,
Vincent Feuillard,
Amandine Marrel
Abstract:
Complex computer codes, for instance simulating physical phenomena, are often too time expensive to be directly used to perform uncertainty, sensitivity, optimization and robustness analyses. A widely accepted method to circumvent this problem consists in replacing cpu time expensive computer models by cpu inexpensive mathematical functions, called metamodels. In this paper, we focus on the Gaussi…
▽ More
Complex computer codes, for instance simulating physical phenomena, are often too time expensive to be directly used to perform uncertainty, sensitivity, optimization and robustness analyses. A widely accepted method to circumvent this problem consists in replacing cpu time expensive computer models by cpu inexpensive mathematical functions, called metamodels. In this paper, we focus on the Gaussian process metamodel and two essential steps of its definition phase. First, the initial design of the computer code input variables (which allows to fit the metamodel) has to honor adequate space filling properties. We adopt a numerical approach to compare the performance of different types of space filling designs, in the class of the optimal Latin hypercube samples, in terms of the predictivity of the subsequent fitted metamodel. We conclude that such samples with minimal wrap-around discrepancy are particularly well-suited for the Gaussian process metamodel fitting. Second, the metamodel validation process consists in evaluating the metamodel predictivity with respect to the initial computer code. We propose and test an algorithm which optimizes the distance between the validation points and the metamodel learning points in order to estimate the true metamodel predictivity with a minimum number of validation points. Comparisons with classical validation algorithms and application to a nuclear safety computer code show the relevance of this new sequential validation design.
△ Less
Submitted 23 September, 2010; v1 submitted 7 January, 2010;
originally announced January 2010.
