Conveners
Computing and Data handling: Computing and Data handling
- Gavin Davies (University Of Mississippi)
Computing and Data handling: Computing and Data handling
- Thea Aarrestad (ETH Zurich (CH))
- Gavin Davies (University Of Mississippi)
Computing and Data handling: Computing and Data handling
- Thea Aarrestad (ETH Zurich (CH))
Computing and Data handling: Computing and Data handling
- Fabio Catalano (CERN)
- Dagmar Adamova (Czech Academy of Sciences (CZ))
Computing and Data handling: Computing and Data handling
- James Catmore (University of Oslo (NO))
- Gavin Davies (University Of Mississippi)
Computing and Data handling: Computing and Data handling
- James Catmore (University of Oslo (NO))
- Gavin Davies (University Of Mississippi)
Computing and Data handling: Computing and Data handling
- Gavin Davies (University Of Mississippi)
- Fabio Catalano (CERN)
Computing and Data handling: Computing and Data handling
- Dagmar Adamova (Czech Academy of Sciences (CZ))
- Fabio Catalano (CERN)
The CMS experiment has recently established a new Common Analysis Tools (CAT) group. The CAT group implements a forum for the discussion, dissemination, organization and development of analysis tools, broadly bridging the gap between the CMS data and simulation datasets and the publication-grade plots and results. In this talk we discuss some of the recent developments carried out in the...
The ATLAS experiment at CERN comprises almost 6000 members. To develop and maintain a system allowing them to analyze the experiment's data, significant effort is required. Such a system consists of millions of lines of code, hundreds of thousand computer cores, and hundreds of petabytes of data. Even a system of this size, while sufficient for current needs, will need to be significantly...
The Auger Offline Framework is a general-purpose C++-based software that allows the reconstruction of the events detected by the Pierre Auger Observatory. Thanks to its modular structure, the collaborators can contribute to the code development with their algorithms and sequencing instructions required for their analyses. It is also possible to feed the Auger Offline Framework with different...
ALICE records Pb-Pb collisions in Run 3 at an unprecedented rate of 50 kHz, storing all data in continuous readout (triggerless) mode. The main purpose of the ALICE online computing farm is the calibration of the detectors and the compression of the recorded data. The detector with the largest data volume by far is the TPC, and the online farm is thus optimized for fast and efficient...
Recent advances in X-ray beamline technologies, including the advent of very high-brilliance beamlines at next-generation synchrotron sources and advanced detector instrumentation, have led to an exponential increase in the speed of data collection. As a consequence, there is an increasing need for a data analysis platform that can refine and optimize data collection strategies in real-time...
High Energy Photon Source(HEPS) will produce huge amount of data. Efficiently storing, analyzing, and sharing this huge amount of data presents a significant challenge for HEPS.
HEPS Computing and Communication System(HEPSCC), has designed and established a network and computing system. A deliciated machine room and high speed network have been ready for production. A computing architecture...
The Deep Underground Neutrino Experiment (DUNE) is a next-generation long-baseline neutrino experiment currently under construction in the US. The experiment consists of a broadband neutrino beam from Fermilab to the Sanford Underground Research Facility (SURF) in Lead, South Dakota, a high-precision near detector, and a large liquid argon time-projection chamber (LArTPC) far detector. The...
The HL-LHC will open an unprecedented window on the weak-scale nature of the universe, providing high-precision measurements of the standard model (SM) as well as searches for new physics beyond the SM. Collecting the information-rich datasets required by such measurements and searches will be a challenging task, given the harsh environment of 200 proton-proton interactions per bunch crossing....
This R&D project, initiated by the DOE Nuclear Physics AI-Machine Learning initiative in 2022, leverages AI to address data processing challenges in high-energy nuclear experiments (RHIC, LHC, and future EIC). Our focus is on developing a demonstrator for real-time processing of high-rate data streams from sPHENIX experiment tracking detectors. Integrating streaming readout and intelligent...
The KOTO experiment's main aim is to measure the branching ratio of the CP-violating $K_L\rightarrow\pi^0\nu\bar{\nu}$ decay. However, data targeting other physics studies can also be recorded at KOTO. Events are rejected or tagged at the L1 stage of KOTO's DAQ based on total energy deposition in different detectors, and trigger modes with high rate are prescaled. The L2 has been recently...
New readout electronics for the ATLAS LAr Calorimeters are being developed, within the framework of the experimental upgrades for the HL-LHC, to be able to operate with a pile-up of up to 200 simultaneous pp interactions. Moreover, the calorimeter signals of up to 25 subsequent collisions are overlapping, which increases the difficulty of energy reconstruction. The energy computation will be...
We present the preparation, deployment, and testing of an autoencoder trained for unbiased detection of new physics signatures in the CMS experiment Global Trigger (GT) test crate FPGAs during LHC Run 3. The GT makes the final decision whether to readout or discard the data from each LHC collision, which occur at a rate of 40 MHz, within a 50 ns latency. The Neural Network makes a prediction...
The Run 3 data-taking conditions pose unprecedented challenges for the DAQ systems of the LHCb experiment at the LHC. Consequently, the LHCb collaboration is pioneering a fully software trigger to cope with the expected increase in event rate. The upgraded trigger has required advances in hardware architectures, expert systems and machine learning solutions. Among the latter, LHCb has explored...
Tree Tensor Networks (TTNs) are hierarchical tensor structures commonly used for representing many-body quantum systems, but can also be applied to ML tasks such as classification or optimization. We study the implementation of TTNs in high-frequency real-time applications such as the online trigger systems of HEP experiments. The algorithmic nature of TTNs makes them easily deployable on...
The LHCb detector generates vast amounts of data (5 TB/s), necessitating efficient algorithms to select data of interest and reduce the bandwidth to acceptable levels in real time. Deploying machine learning (ML) models for inference at all trigger stages is challenging, as the models need to fulfill strict throughput requirements.
To achieve the throughput aims, optimized batched...
Identification of hadronic jets originating from heavy-flavor quarks is extremely important to several physics analyses in High Energy Physics, such as studies of the properties of the top quark and the Higgs boson, and searches for new physics. Recent algorithms used in the CMS experiment were developed using state-of-the-art machine-learning techniques to distinguish jets emerging from the...
Flavour-tagging is a critical component of the ATLAS experiment physics programme. Existing flavour tagging algorithms rely on several low-level taggers, which are a combination of physically informed algorithms and machine learning models. A novel approach presented here instead uses a single machine learning model based on reconstructed tracks, avoiding the need for low-level taggers based...
We propose a differentiable vertex fitting algorithm that can be used for secondary vertex fitting, and that can be seamlessly integrated into neural networks for jet flavour tagging. Vertex fitting is formulated as an optimization problem where gradients of the optimized solution vertex are defined through implicit differentiation and can be passed to upstream or downstream neural network...
Many extensions of the standard model can give rise to tau leptons produced in non-conventional signatures in the detector. For example, certain long-lived particles can decay to produce taus that are displaced from the primary proton-proton interaction vertex. The standard tau reconstruction and identification techniques are suboptimal for displaced tau leptons, which require specialized...
The upcoming HL-LHC represents a steep increase in the average number of pp interactions and hence in the computing resources required for offline track reconstruction of the ATLAS Inner Tracker (ITk). Track pattern recognition algorithms based on Graph Neural Networks (GNNs) have been demonstrated as a promising approach to these challenges. We present in this contribution a novel algorithm...
In this presentation we describe the performance obtained running machine learning models studied for the ATLAS Muon High Level Trigger. These models are designed for hit position reconstruction and track pattern recognition with a tracking detector, on different models of commercially available Xilinx FPGA cards: Alveo U50, Alveo U250, and Versal VCK5000. We compare the inference times...
Increases in instantaneous luminosity and detector granularity will increase the amount of data that has to be analyzed by high-energy physics experiments, whether in real time or offline, by an order of magnitude. In this context, Graph Neural Networks have received a great deal of attention in the community for the reconstruction of charged particles, because their computational complexity...
Tracking charged particles in high-energy physics experiments is a computationally intensive task. With the advent of the High Luminosity LHC era, which is expected to significantly increase the number of proton-proton interactions per beam collision, the amount of data to be analysed will increase dramatically.
Traditional algorithms suffer from scaling problems. We are investigating the...
Tracking is one of the most crucial components of reconstruction in collider experiments. It is known for high consumption of computing resources, and various investigations are ongoing to cope with this challenge. The track reconstruction can be considered as a quadratic unconstrained binary optimization (QUBO) problem. Recent progress with two complementary approaches will be presented: (1)...
Deep learning methods are becoming indispensable in the data analysis of particle physics experiments, with current neutrino studies demonstrating their superiority over traditional tools in various domains, particularly in identifying particles produced by neutrino interactions and fitting their trajectories. This talk will showcase a comprehensive reconstruction strategy of the neutrino...
The Deep Underground Neutrino Experiment (DUNE) is a next-generation long-baseline neutrino oscillation experiment with a broad researchย programย including measuring CP-violation in the neutrino sector, determining neutrino mass ordering and studying neutrinos from space.ย DUNEย will employ massive, high-precision Liquid-Argon Time-Projection Chambers at the far siteย (70 kt total mass)ย to produce...
Deep learning can give a significant impact on physics performance of electron-positron Higgs factories such as ILC and FCCee. We are working on two topics on event reconstruction to apply deep learning; one is jet flavor tagging. We apply particle transformer to ILD full simulation to obtain jet flavor, including strange tagging. The other one is particle flow, which clusters calorimeter hits...
In this work, we would like to present a novel approach to the reconstruction of multiple calorimetric clusters within the Large Hadron Collider forward (LHCf) experiment using Machine Learning techniques. The LHCf experiment is dedicated to understand the hadronic components of cosmic rays by measuring particle production in the forward region of LHC collisions. One of the significant...
Hyper-Kamiokande (Hyper-K) is a next generation water-Cherenkov neutrino experiment, currently under construction to build on the success of its predecessor Super-Kamiokande (Super-K). With 8 times greater fiducial volume and enhanced detection capabilities, it will have significantly reduced statistical uncertainties as compared to Super-K. For corresponding suppression of backgrounds and...
The fidelity of detector simulation is crucial for precision experiments, such as DUNE which uses liquid argon time projection chambers (LArTPCs). We can improve the detector simulation by performing dedicated calibration measurements. Using conventional calibration approaches, typically we are only able to tackle individual detector processes per measurement. However, the detector effects are...
In preparation for Run 3 at the LHC, the MC Simulation performed with Geant4 within ATLAS has undergone significant improvements to enhance its computational performance and overall efficiency. This talk offers a comprehensive overview of the optimizations implemented in the ATLAS simulation for Run 3. Notable developments include the application of EM range cuts, the implementation of Neutron...
Simulating detector and reconstruction effects on physics quantities is of paramount importance for data analysis, but unsustainably costly for the upcoming HEP experiments.
The most radical approach to speed-up detector simulation is a Flash Simulation, as proposed by the LHCb collaboration in Lamarr, a software package implementing a novel simulation paradigm relying on deep generative...
The simulation of MC events is a crucial task and an indispensable ingredient for every physics analysis. To reduce the CPU needs of the GEANT simulation, ATLAS has developed a strong program to replace parts of the simulation chain by fast simulation tools. Among those tools is AtlFast3, which utilizes a combination of Generative Adversarial Networks and sophisticated parametrizations for the...
The Circular Electron Positron Collider (CEPC) is a future Higgs factory to measure the Higgs boson properties. Like the other future experiments, the simulation software plays a crucial role in CEPC for detector designs, algorithm optimization and physics studies. Due to similar requirements, the software stack from the Key4hep project has been adopted by CEPC. As the initial application of...
Micropattern Gaseous Detectors (MPGDs) rely heavily on the simulation of the particle passage as conducting these studies allows scientists to cut huge costs and development for prototyping. Even though Garfield++ is a very important part of the simulation of MPGDs, it is very comprehensively intensive particularly when large detector volumes and high gas gains are required. In order to mimic...
Particle physics relies on Monte Carlo (MC) event generators for theory-data comparison, necessitating several samples to address theoretical systematic uncertainties at a high computational cost. The MC statistic becomes a limiting factor and the significant computational cost a bottleneck in most physics analyses. In this talk, the Deep neural network using Classification for Tuning and...
The CMS Collaboration has recently approved the publication of full statistical models of physics analyses. This includes the publication of the CMS data, which facilitates the construction of the full likelihood. The statistical inference tool "Combine" needed for this purpose is now available under an open source licence. This talk highlights some features of Combine and discusses the use of...
With the growing datasets of HEP experiments, statistical analysis becomes more computationally demanding, requiring improvements in existing statistical analysis software. One way forward is to use Automatic Differentiation (AD) in likelihood fitting, which is often done with RooFit (a toolkit that is part of ROOT.) As of recently, RooFit can generate the gradient code for a given likelihood...
A flexible and dynamic environment capable of accessing distributed data and resources efficiently, is a key aspect for HEP data analysis, especially for the HL-LHC era. A quasi-interactive declarative solution, like ROOT RDataFrame, with scale-up capabilities via open-source standards like Dask, can profit from the "HPC, Big Data and Quantum Computing" Italian Center DataLake model under...
In recent years, the data published by the Particle Data Group (PDG) in the Review of Particle Physics has primarily been consulted on the PDG web pages and in pdgLive, or downloaded in the form of PDF files. A new set of tools (PDG API) makes PDG data easily accessible in machine-readable format and includes a REST API, downloadable database files containing the PDG data, and an associated...
The Key4hep project aims at providing a complete software stack to enable complete and detailed detector studies for future experiments. It was first envisaged five years ago by members of the CEPC, CLIC, ILC and FCC communities and has since managed to attract contributions also from others, such as the EIC or the MuonCollider. Leveraging established community tools, as well as developing new...
"Data deluge" refers to the situation where the sheer volume of new data generated overwhelms the capacity of institutions to manage it and researchers to use it. This is becoming a common problem in industry and big science facilities like the MAX IV laboratory and the LHC.
As a solution to this problem, a small collaboration of researchers has developed a machine learning-based data...
Jiangmen Underground Neutrino Observatory (JUNO), located in the southern part of China, will be the worldโs largest liquid scintillator (LS) detector upon completion. Equipped with 20 kton LS, about 17612 20-inch PMTs and 25600 3-inch PMTs in the central detector (CD), JUNO will provide a unique apparatus to probe the mysteries of neutrinos, particularly the neutrino mass ordering puzzle. In...
Advanced machine-learning (ML) based methods are being increasingly used to tackle the analyses of large and complex datasets. At CMS we explore the unique opportunity to exploit these new ML methods to extract information and address scientific questions to search for physics beyond the Standard Model, with the overarching aim to discern possible signatures for new physics. In this talk we...
Designing the next generation colliders and detectors involves solving optimization problems in high-dimensional spaces where the optimal solutions may nest in regions that even a team of expert humans would not explore.
Resorting to Artificial Intelligence to assist the experimental design introduces however significant computational challenges in terms of generation and processing of the...
Recent advances in AI have been significant, with large language models demonstrating astonishing capabilities that hold the promise of driving new scientific discoveries in high-energy physics. In this report, we will discuss two potential approaches to large models. The first is a specialized intelligent agent for BESIII experiment based on large language models, encompassing its brain,...
The CMS Tracker in Run 3 is made up of thousands of silicon modules (Pixel:1856 modules, Strip: 15148 modules). Because of the aging of the detector, and all other possible accidents that may happen during the operations, there is the need for constant monitoring of the detector components, in order to guarantee the best data quality. The procedures and tools adopted by the CMS Tracker group...
This talk will summarise a method based on machine learning to play the devil's advocate and investigate the impact of unknown systematic effects in a quantitative way. This method proceeds by reversing the measurement process and using the physics results to interpret systematic effects under the Standard Model hypothesis. We explore this idea in arXiv:2303.15956 by considering the...
