Roy Cruz Candelaria, Wouter Deconinck, Aman Desai
et al.
We present the material and resources developed for training physicists on containerization technologies enabled by Apptainer. In the context of analysis preservation using Apptainer's capabilities, we have developed examples that execute common tools in High Energy Physics (HEP) and Nuclear Physics within containers. Training physicists on containerization technologies is of utmost importance in today's research landscape. By embracing these technologies, users can achieve enhanced reproducibility, portability, collaboration, and resource efficiency, assuring the conditions and integrity of the scientific analysis process. This training module,``Introduction to Apptainer/Singularity'', is part of the HEP Software Foundation Training Center, which aims to equip newcomers to the field of High Energy Physics with the necessary software skills and best practices.
Ibrahim Elsharkawy, Vinicius Mikuni, Wahid Bhimji
et al.
We present OmniMol, a state-of-the-art transformer-based small molecule machine-learned interatomic potential (MLIP). OmniMol is built by adapting Omnilearned, a foundation model for particle jets found in high-energy physics (HEP) experiments such as at the Large Hadron Collider (LHC). Omnilearned is built with a Point-Edge-Transformer (PET) and pre-trained using a diverse set of one billion particle jets. It includes an interaction-matrix attention bias that injects pairwise sub-nuclear (HEP) or atomic (molecular-dynamics) physics directly into the transformer's attention logits, steering the network toward physically meaningful neighborhoods without sacrificing expressivity. We demonstrate OmniMol using the oMol dataset and find excellent performance even with relatively few examples for fine-tuning. This study lays the foundation for building interdisciplinary connections, given datasets represented as collections of point clouds.
Detector and event visualization are crucial components of high-energy physics~(HEP) experimental software. Virtual Reality~(VR) technologies and multimedia development platforms such as Unity offer enhanced display effects and flexible extensibility for visualization in HEP experiments. In this study, we present a VR-based method for detector and event displays in the Jiangmen Underground Neutrino Observatory~(JUNO) experiment. This method shares the same detector geometry descriptions and event data model as those in offline software and provides necessary data conversion interfaces. The VR methodology facilitates an immersive exploration of the virtual environment in JUNO, enabling users to investigate detector geometry, visualize event data, and tune the detector simulation and event reconstruction algorithms. Additionally, this approach supports applications in data monitoring, physics data analysis, and public outreach initiatives.
Florian Rehm, Sofia Vallecorsa, Michele Grossi
et al.
The prospect of quantum computing with a potential exponential speed-up compared to classical computing identifies it as a promising method in the search for alternative future High Energy Physics (HEP) simulation approaches. HEP simulations, such as employed at the Large Hadron Collider at CERN, are extraordinarily complex and require an immense amount of computing resources in hardware and time. For some HEP simulations, classical machine learning models have already been successfully developed and tested, resulting in several orders of magnitude speed-up. In this research, we proceed to the next step and explore whether quantum computing can provide sufficient accuracy, and further improvements, suggesting it as an exciting direction of future investigations. With a small prototype model, we demonstrate a full quantum Generative Adversarial Network (GAN) model for generating downsized eight-pixel calorimeter shower images. The advantage over previous quantum models is that the model generates real individual images containing pixel energy values instead of simple probability distributions averaged over a test sample. To complete the picture, the results of the full quantum GAN model are compared to hybrid quantum-classical models using a classical discriminator neural network.
Pavel Fileviez Perez, Andrea Pocar, K. S. Babu
et al.
In this report we discuss the main theories to understand the origin of baryon and lepton number violation in physics beyond the Standard Model. We present the theoretical predictions for rare processes such as neutrinoless double beta decay, proton decay, and neutron-antineutron oscillation, and overview the prospects to discover these rare processes in the near future. The possibility to observe baryon and lepton violating signatures at current and future colliders and through precision studies of other rare processes, and the testability of different baryogenesis mechanisms is discussed in detail. A healthy and broad experimental program looking for proton decay, neutrinoless double beta decay and neutron-antineutron oscillations is essential to make new discoveries in this field. These searches are carried out at various experimental facilities in the US and abroad, and use instrumentation arching across traditional HEP/NP boundaries. In addition, experiments such as those at the Large Hadron Collider could discover exotic baryon and/or lepton number violating signatures connected to low energy scale theories for neutrino masses, supersymmetric models with R-parity violation, new gauge theories or other mechanisms for physics beyond the Standard Model. The landscape presented in this report could be crucial to discover the underlying mechanism for neutrino masses and the matter-antimatter asymmetry in the universe.
For many decades High Energy Physics (HEP) instrumentation has been concentrated on detectors of ionizing radiation -- where the energy of incident particles or photons is sufficient to create mobile charge in gas, liquid, or solid material, which can be processed by front end electronics (FEE) to provide information about the position, energy, and timing of the incident radiation. However, recently-proposed HEP experiments need to sense or control EM radiation in the radiofrequency (RF) range, where ionization detectors are unavailable. These experiments can take advantage of emerging microelectronics developments fostered by the explosive growth of wireless data communications in the commercial sector. Moore's Law advances in semiconductor technology have brought about the recent development of advanced microelectronic components with groundbreaking levels of analog-digital integration and processing speed. In particular, RF "System-on-Chip" (RFSoC) platforms offer multiple data converter interfaces to the analog world (ADCs and DACs) having bandwidths approaching 10GHz and abundant digital signal processing resources on the same silicon die. Such devices eliminate the complex PC board interfaces that have long been used to couple discrete ADCs and DACs to FPGA processors, thus radically reducing power consumption, impedance mismatch, and footprint area, while allowing analog preconditioning circuits to be eliminated in favor of digital processing. Costed for wide deployment, these devices are helping to accelerate the trend towards "software defined radio" in several high-volume commercial markets. In this whitepaper we highlight some HEP applications where leading-edge RF microelectronics can be a key enabler.
With the advancement of the accelerator systems and the requirements of high luminosity particle beams to reach different physics goals, detectors with good position resolution and high rate handling capability have become essential for designing any High Energy Physics (HEP) experiments. The Gas Electron Multiplier (GEM) detectors are widely used in many HEP experiments as a tracking device because of their good spatial resolution and rate handling capability. The presence of the dielectric medium inside the active volume of the GEM detector changes its behaviour when exposed to external radiation. This mechanism is commonly referred as the charging-up effect. In this article, the effect of the charging-up phenomenon and the initial polarisation effect of the dielectric on the gain of the chamber are reported for a single mask triple GEM chamber with Ar/CO2 gas mixture.
In high energy physics (HEP), jets are collections of correlated particles produced ubiquitously in particle collisions such as those at the CERN Large Hadron Collider (LHC). Machine learning (ML)-based generative models, such as generative adversarial networks (GANs), have the potential to significantly accelerate LHC jet simulations. However, despite jets having a natural representation as a set of particles in momentum-space, a.k.a. a particle cloud, there exist no generative models applied to such a dataset. In this work, we introduce a new particle cloud dataset (JetNet), and apply to it existing point cloud GANs. Results are evaluated using (1) 1-Wasserstein distances between high- and low-level feature distributions, (2) a newly developed Fréchet ParticleNet Distance, and (3) the coverage and (4) minimum matching distance metrics. Existing GANs are found to be inadequate for physics applications, hence we develop a new message passing GAN (MPGAN), which outperforms existing point cloud GANs on virtually every metric and shows promise for use in HEP. We propose JetNet as a novel point-cloud-style dataset for the ML community to experiment with, and set MPGAN as a benchmark to improve upon for future generative models. Additionally, to facilitate research and improve accessibility and reproducibility in this area, we release the open-source JetNet Python package with interfaces for particle cloud datasets, implementations for evaluation and loss metrics, and more tools for ML in HEP development.
We investigate unification prospects of two heavy scalar triplet extension of the standard model where, in the absence of any right-handed neutrino (RHN), type-II seesaw accounts for current oscillation data with hierarchical neutrino masses consistent with cosmological bounds and the lighter triplet decay explains baryon asymmetry of the Universe via leptogenesis. We note that the absence of RHNs in the fundamental fermion representations of SU(5) delineates its outstanding position compared to SO(10) (or $E_6$). In addition, SU(5) needs smaller scalar representations ${15}_{H1}\oplus {15}_{H2}$ compared to much larger representations ${126}_{H1}\oplus {126}_{H2} \subset $ SO(10) (or ${351}^{\prime}_{H1}\oplus {351}^{\prime}_{H2} \subset E_6$). We show how precision gauge coupling unification is achieved through SU(5) with the predictions of different sets of two heavy triplet masses which, besides being compatible with type-II seesaw, are also consistent with unflavoured or $τ-$ flavoured leptogenesis. In addition to an intermediate mass colour octet fermion, completion of precision gauge coupling unification is found to require essentially the presence of the well known weak triplet fermion $Σ(3,0,1)$ in its mass range $M_Σ\simeq {\cal O}(500-3000)$ GeV out of which the dominant dark matter (DM) resonance mass $M_Σ\ge 2.4$ TeV is known to account for the observed cosmological relic density. The deficiency in relic density for lighter $Σ$ solutions is compensated by an additional scalar singlet DM. A GUT ansatz is noted to ensure vacuum stability of the SM scalar potential for all types of unification solutions realised in this work. We discuss proton lifetime estimations for $p\to e^+π^0$ compatible with the present Hyper-Kamiokande bound as a function of an unknown mixing parameter in the model.
Matteo Migliorini, Riccardo Castellotti, Luca Canali
et al.
The effective utilization at scale of complex machine learning (ML) techniques for HEP use cases poses several technological challenges, most importantly on the actual implementation of dedicated end-to-end data pipelines. A solution to these challenges is presented, which allows training neural network classifiers using solutions from the Big Data and data science ecosystems, integrated with tools, software, and platforms common in the HEP environment. In particular, Apache Spark is exploited for data preparation and feature engineering, running the corresponding (Python) code interactively on Jupyter notebooks. Key integrations and libraries that make Spark capable of ingesting data stored using ROOT format and accessed via the XRootD protocol, are described and discussed. Training of the neural network models, defined using the Keras API, is performed in a distributed fashion on Spark clusters by using BigDL with Analytics Zoo and also by using TensorFlow, notably for distributed training on CPU and GPU resourcess. The implementation and the results of the distributed training are described in detail in this work.