Search

In this paper, we present the results of charged particle track reconstruction in CLAS12 using artificial intelligence. In our approach, we use neural networks working together to identify tracks based on the raw signals in the Drift Chambers. A Convolutional Auto-Encoder is used to de-noise raw data by removing the hits that do not satisfy the patterns for tracks, and second Multi-Layer Perceptron is used to identify tracks from combinations of clusters in the drift chambers. Our method increases the tracking efficiency by 50% for multi-particle final states already conducted experiments. The de-noising results indicate that future experiments can run at higher luminosity without degradation of the data quality. This in turn will lead to significant benefits for the CLAS12 physics program. [ABSTRACT FROM AUTHOR]

The Jiangmen Underground Neutrino Observatory (JUNO) experiment is set to begin data taking in 2024 with the aim of determining the neutrino mass ordering (NMO) to a significance of 3 σ after 6 years of data taking. Achieving this goal requires effective background suppression, with the background induced by cosmic-ray muons being one of the most significant sources of interference in the NMO study. Accurately simulating the cosmic-ray muon background is crucial for the success of the experiment, but the sheer number of optical photons produced by the muon makes this detector simulation process extremely time-consuming using traditional methods such as Geant4. This paper presents a fast muon simulation method that employs neural networks to expedite the simulation process. Our approach achieves an order-of-magnitude speed-up in simulation time compared to Geant4, while still producing accurate results. [ABSTRACT FROM AUTHOR]

The findable, accessible, interoperable, and reusable (FAIR) data principles serve as a framework for examining, evaluating, and improving data sharing to advance scientific endeavors. There is an emerging trend to adapt these principles for machine learning models—algorithms that learn from data without specific coding—and, more generally, AI models, due to AI's swiftly growing impact on scientific and engineering sectors. In this paper, we propose a practical definition of the FAIR principles for AI models and provide a template program for their adoption. We exemplify this strategy with an implementation from high-energy physics, where a graph neural network is employed to detect Higgs bosons decaying into two bottom quarks. [ABSTRACT FROM AUTHOR]

The operating conditions defining the current data taking campaign at the Large Hadron Collider, known as Run 3, present unparalleled challenges for the real-time data acquisition workflow of the LHCb experiment at CERN. To address the anticipated surge in luminosity and consequent event rate, the LHCb experiment is transitioning to a fully software-based trigger system. This evolution necessitated innovations in hardware configurations, software paradigms, and algorithmic design. A significant advancement is the integration of monotonic Lipschitz Neural Networks into the LHCb trigger system. These deep learning models offer certified robustness against detector instabilities, and the ability to encode domain-specific inductive biases. Such properties are crucial for the inclusive heavy-flavour triggers and, most notably, for the topological triggers designed to inclusively select b-hadron candidates by exploiting the unique kinematic and decay topologies of beauty decays. This paper describes the recent progress in integrating Lipschitz Neural Networks into the topological triggers, highlighting the resulting enhanced sensitivity to highly displaced multi-body candidates produced within the LHCb acceptance. [ABSTRACT FROM AUTHOR]

There are ongoing efforts to investigate theories that aim to explain the current shortcomings of the Standard Model of particle physics. One such effort is the Long-Lived Particle Jet Tagging Algorithm, based on a DNN (Deep Neural Network), which is used to search for exotic new particles. This paper describes two novel optimisations in the design of this DNN, suitable for implementation on an FPGA-based accelerator. The first involves the adoption of cyclic random access memories and the reuse of multiply-accumulate operations. The second involves storing matrices distributed over many RAM memories with elements grouped by index. An evaluation of the proposed methods and hardware architectures is also included. The proposed optimisations can yield performance enhancements by more than an order of magnitude compared to software implementations. The innovations can also lead to smaller FPGA footprints and accordingly reduce power consumption, allowing for instance duplication of compute units to achieve increases in effective throughput. [ABSTRACT FROM AUTHOR]