Accurate knowledge of the magnetic field distribution is essential for precise momentum reconstruction of charged particles in the SPD experiment at NICA. Traditional interpolation of pre-computed field maps becomes computationally expensive for high-resolution grids and may be inaccurate in regions with sparse measurements.
A systematic comparative study of machine learning methods for...
The CEPC working point at the center of mass energy around 240 GeV corresponds to the largest cross section of the higgsstrahlung process and allows effective usage of the recoil mass technique for measurements of Higgs boson parameters. The interesting task is the search for invisible decays of the Higgs boson: H->ZZ*->4nu. The dijet final states of the process e+e-->ZH, where Z boson decays...
The Joint Institute for Nuclear Research (JINR) is approaching the commissioning and operational readiness of the NICA (Nuclotron-based Ion Collider fAcility) mega-science accelerator complex. One of its future detectors is the SPD (Spin Physics Detector), designed to study the spin structure of nucleons. The Range System (muon system) is an integral part of the detector, used for muon...
Fast calorimeter simulation requires accurate generative models and meaningful
measures of physical fidelity. We integrate physics-supervised distributional evaluation and residual regularization into our conditional diffusion pipeline for the first time. Shower generation is conditioned on particle entry coordinates and momentum, allowing the model to represent variations in detector response...
A Graph Attention Neural Network (GANN) for particle track clustering by events in time slices obtained at the Spin Physics Detector (SPD) of the Nuclotron-based Ion Collider fAcility (NICA) is presented. A novel approach is applied to time slice processing at high particle multiplicity and pile-up. The model is trained using a hierarchical approach and evaluated on time slices generated by...
Holographic QCD models describe strongly interacting matter through a five-dimensional gravitational background. In Einstein–Maxwell–Dilaton models, the warp factor A(z) deforms the AdS metric and affects QCD thermodynamics, the phase structure, and heavy-meson spectra. Such models are commonly calibrated using lattice-QCD data [1, 2] in contrast with employment on the experimental data.
In...
Focusing aerogel ring-imaging Cherenkov (FARICH) detectors require the reconstruction of sparse photon patterns in the presence of background caused by a photosensor dark count rate. We present the Track-Conditioned Ring Assignment Transformer (TCRAT), an attention-based model that simultaneously assigns photon hits, reconstructs the particle velocity, and performs particle identification in...
The talk presents the first experience of applying deep neural networks to simulate the responses of microstrip GEM detectors used in the BM@N experiment at the NICA accelerator complex (JINR). The preparation of the training dataset based on data obtained using classical Monte-Carlo simulation methods implemented in the BMNROOT software is discussed. The advantages of the neural network...
We discuss neural networks for cosmic ray classification at the Carpet-3 experiment. We improve gamma ray sensitivity and obtain new limits on the isotropic diffuse gamma ray flux from Carpet-3 data, improving previously obtained results. We also study the possibility of measuring the cosmic ray mass composition at the knee region at the Carpet-3 experiment, comparing standard reconstruction...
Baikal-GVD is a deep underwater Cherenkov telescope at the bottom of Lake Baikal, built to detect high-energy astrophysical neutrinos. The energy of the neutrino is one of the key parameters of a recorded event, which makes its accurate measurement an important task of the experiment.
We apply transformer neural networks to the reconstruction of the energy of a muon produced by a muon...
The aim of this work is to forecast the onset time of the second maximum of proton enhancements associated with the transport of particles by solar-wind structures. The "zeroth-order" forecast is based on the kinematic parameters of the solar wind measured at the L1 Lagrange point, and is subsequently refined by machine-learning methods. The difficulty of this problem lies in the nonlinear,...
In order to improve the circulation purification efficiency of the PandaX-xT dark matter detector and efficiently achieve the ultra-trace background required by the detector through flow field optimization, this study conducts CFD numerical simulations of the xenon flow field based on the structure of the PandaX-4T detector. The multi-variable simulation results are introduced into an...
Two main types of ground-based experimental facilities are used to register EAS: atmospheric Cherenkov telescopes and arrays of Cherenkov detectors. The unique TAIGA scientific facility includes both types of installations. The joint analysis of the data coming from these installations should improve the quality of the reconstructed EAS parameters compared to traditional methods that analyze...
Baikal-GVD neutrino telescope is located in the Southern basin of Lake Baikal with the instrumented depth ranging from 730 to 1275 m. The measured noise rates of the detector channels range from ten to hundreds of kilohertz, depending on season and the depth of the installed photomultiplier. We present an evolution of the method currently employed for muon track event reconstruction in...
We present an overview of the deep learning pipeline developed for data processing at the Baikal-GVD experiment. The pipeline covers the complete data analysis chain, from noise hit suppression to the reconstruction of neutrino arrival directions and energies. By capturing spatial and temporal correlations across optical modules, Transformer and Graph Neural Network (GNN) architectures...
We review applications of large language models in astrophysics and astroparticle physics, including literature retrieval, scientific programming, observational data analysis, and computational experiments. We compare three complementary approaches: domain adaptation, retrieval-augmented generation, and agents that use external tools. Building on this comparison, we propose an architecture for...
