A5106315352- Particle physics theoretical and experimental studies
- High-Energy Particle Collisions Research
- Quantum Chromodynamics and Particle Interactions
- Particle Detector Development and Performance
- Dark Matter and Cosmic Phenomena
- Computational Physics and Python Applications
- Neutrino Physics Research
- Cosmology and Gravitation Theories
- Medical Imaging Techniques and Applications
- Black Holes and Theoretical Physics
- Particle Accelerators and Free-Electron Lasers
- Astrophysics and Cosmic Phenomena
- Anomaly Detection Techniques and Applications
- Radiation Detection and Scintillator Technologies
- Distributed and Parallel Computing Systems
- Atomic and Subatomic Physics Research
- Superconducting Materials and Applications
- Parallel Computing and Optimization Techniques
- Fractal and DNA sequence analysis
- Big Data Technologies and Applications
- Generative Adversarial Networks and Image Synthesis
- Scientific Computing and Data Management
- Advanced Neural Network Applications
- Gamma-ray bursts and supernovae
- Nuclear physics research studies
European Organization for Nuclear Research
2016-2025
Institute of High Energy Physics
2015-2024
University of Antwerp
2024
A. Alikhanyan National Laboratory
2022-2024
University of California, San Diego
2023
Paul Scherrer Institute
2014-2023
Queen Mary University of London
2011-2022
Istituto Nazionale di Fisica Nucleare, Sezione di Bologna
2008-2022
Istituto Nazionale di Fisica Nucleare, Sezione di Roma I
2001-2022
Yandex (Russia)
2017-2022
Recent results at the Large Hadron Collider (LHC) have pointed to enhanced physics capabilities through improvement of real-time event processing techniques. Machine learning methods are ubiquitous and proven be very powerful in LHC physics, particle as a whole. However, exploration use such techniques low-latency, low-power FPGA hardware has only just begun. FPGA-based trigger data acquisition (DAQ) systems extremely low, sub-microsecond latency requirements that unique physics. We present...
The discovery by the ATLAS and CMS experiments of a new boson with mass around 125 GeV measured properties compatible those Standard-Model Higgs boson, coupled absence discoveries phenomena beyond Standard Model at TeV scale, has triggered interest in ideas for future factories. A circular e+e- collider hosted 80 to 100 km tunnel, TLEP, is among most attractive solutions proposed so far. It clean experimental environment, produces high luminosity top-quark, W Z studies, accommodates multiple...
A bstract Using variational autoencoders trained on known physics processes, we develop a one-sided threshold test to isolate previously unseen processes as outlier events. Since the autoencoder training does not depend any specific new signature, proposed procedure doesn’t make assumptions nature of physics. An event selection based this algorithm would be complementary classic LHC searches, typically model-dependent hypothesis testing. Such an deliver list anomalous events, that...
Using detailed simulations of calorimeter showers as training data, we investigate the use deep learning algorithms for simulation and reconstruction particles produced in high-energy physics collisions. We train neural networks on shower data at calorimeter-cell level, show significant improvements when using these compared to methods which rely currently-used state-of-the-art algorithms. define two models: an end-to-end network performs simultaneous particle identification energy...
Abstract is a flexible open-source tool which, given the Standard Model or any of its extensions, allows to (i) fit model parameters set experimental observables; (ii) obtain predictions for observables. can be used either in Monte Carlo mode, perform Bayesian Markov Chain analysis model, as library, observables point parameter space allowing statistical framework. In present version, around thousand have been implemented and several new physics scenarios. this paper, we describe general...
We describe the outcome of a data challenge conducted as part Dark Machines Initiative and Les Houches 2019 workshop on Physics at TeV colliders. The challenged aims detecting signals new physics LHC using unsupervised machine learning algorithms. First, we propose how an anomaly score could be implemented to define model-independent signal regions in searches. large benchmark dataset, consisting >1 Billion simulated events corresponding $10~\rm{fb}^{-1}$ proton-proton collisions...
We assess the impact of very recent measurement top-quark mass by CMS Collaboration on fit electroweak data in standard model and beyond, with particular emphasis prediction for W boson. then compare this average corresponding experimental measurements including new CDF Collaboration, discuss its compatibility model, physics models oblique corrections, dimension-six effective field theory. Finally, we present updated global to precision these models.
The ongoing quest to discover new phenomena at the LHC necessitates continuous development of algorithms and technologies. Established approaches like machine learning, along with emerging technologies such as quantum computing show promise in enhancement experimental capabilities. In this work, we propose a strategy for anomaly detection tasks based on unsupervised demonstrate its effectiveness identifying phenomena. designed models-an kernel two clustering algorithms-are trained detect...
The present report documents the results of Working Group 2: B, D and K decays, workshop on Flavor in Era LHC, held at CERN from November 2005 through March 2007. With advent we will be able to probe New Physics (NP) up energy scales almost one order magnitude larger than it has been possible with accelerator facilities. While direct detection new particles main avenue establish presence NP indirect searches provide precious complementary information, since most probably not measure full...
The discovery of a Higgs particle is possible in variety search channels at the LHC. However, true identity any putative boson will, first, remain ambiguous until one has experimentally excluded other assignments quantum numbers and couplings. We quantify degree to which can discriminate standard model from ``look-alikes'' at, or close to, moment focus on fully-reconstructible golden decay mode pair $Z$ bosons four-lepton final state. Considering both on-shell off-shell $Z$'s, we show how...
We investigate the performance of a jet identification algorithm based on interaction networks (JEDI-net) to identify all-hadronic decays high-momentum heavy particles produced at LHC and distinguish them from ordinary jets originating hadronization quarks gluons. The dynamics are described as set one-to-one interactions between constituents. Based representation learned these interactions, is associated one considered categories. Unlike other architectures, JEDI-net models achieve their...
We present results from a state-of-the-art fit of electroweak precision observables and Higgs-boson signal-strength measurements performed using 7 8 TeV data the Large Hadron Collider. Based on HEPfit package, our study updates traditional extends it to include measurements. As result we obtain constraints new physics corrections both couplings. projected accuracy taking into account expected sensitivities at future colliders.
Using generative adversarial networks (GANs), we investigate the possibility of creating large amounts analysis-specific simulated LHC events at limited computing cost. This kind model is analysis specific in sense that it directly generates high-level features used last stage a given physics analyses, learning N-dimensional distribution relevant context selection. We apply this idea to generation muon four-momenta $Z \to \mu\mu$ LHC. highlight how use-case issues emerge when distributions...
Exploiting the rapid advances in probabilistic inference, particular variational Bayes and autoencoders (VAEs), for anomaly detection (AD) tasks remains an open research question. Previous works argued that training VAE models only with inliers is insufficient framework should be significantly modified order to discriminate anomalous instances. In this work, we exploit deep conditional autoencoder (CVAE) define original loss function together a metric targets hierarchically structured data...
Abstract We discuss a method that employs multilayer perceptron to detect deviations from reference model in large multivariate datasets. Our data analysis strategy does not rely on any prior assumption the nature of deviation. It is designed be sensitive small discrepancies arise datasets dominated by model. The main conceptual building blocks were introduced D’Agnolo and Wulzer (Phys Rev D 99 (1), 015014. 10.1103/PhysRevD.99.015014 . arXiv:1806.02350 [hep-ph], 2019). Here we make decisive...
We introduce an automated tool for deploying ultra low-latency, low-power deep neural networks with convolutional layers on FPGAs. By extending the hls4ml library, we demonstrate inference latency of $5\,\mu$s using architectures, targeting microsecond applications like those at CERN Large Hadron Collider. Considering benchmark models trained Street View House Numbers Dataset, various methods model compression in order to fit computational constraints a typical FPGA device used trigger and...
Abstract We apply an Adversarially Learned Anomaly Detection (ALAD) algorithm to the problem of detecting new physics processes in proton–proton collisions at Large Hadron Collider. detection based on ALAD matches performances reached by Variational Autoencoders, with a substantial improvement some cases. Training 4.4 fb $$^{-1}$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:msup><mml:mrow/><mml:mrow><mml:mo>-</mml:mo><mml:mn>1</mml:mn></mml:mrow></mml:msup></mml:math> 8 TeV...