The direction of extensive air showers can be used to determine the source of gamma quanta and plays an important role in estimating the energy of the primary particle. The data from an array of non-imaging Cherenkov detector stations HiSCORE in the TAIGA experiment registering the number of photoelectrons and detection time can be used to estimate the shower direction with high accuracy. In this work, we use artificial neural networks trained on Monte Carlo-simulated TAIGA HiSCORE data for gamma quanta to obtain shower direction estimates. The neural networks are multilayer perceptrons with skip connections using partial data from several HiSCORE stations as inputs; composite estimates are derived from multiple individual estimates by the neural networks. We apply a two-stage algorithm in which the direction estimates obtained in the first stage are used to transform the input data and refine the estimates. The mean error of the final estimates is less than 0.25 degrees. The approach will be used for multimodal analysis of the data from several types of detectors used in the TAIGA experiment.

Modern Imaging Atmospheric Cherenkov Telescopes (IACTs) generate a huge amount of data that must be classified automatically, ideally in real time. Currently, machine learning-based solutions are increasingly being used to solve classification problems. However, these classifiers require proper training data sets to work correctly. The problem with training neural networks on real IACT data is that these data need to be pre-labeled, whereas such labeling is difficult and its results are estimates. In addition, the distribution of incoming events is highly imbalanced. Firstly, there is an imbalance in the types of events, since the number of detected gamma quanta is significantly less than the number of protons. Secondly, the energy distribution of particles of the same type is also imbalanced, since high-energy particles are extremely rare. This imbalance results in poorly trained classifiers that, once trained, do not handle rare events correctly. Using only conventional Monte Carlo event simulation methods to solve this problem is possible, but extremely resource-intensive and time-consuming. To address this issue, we propose to perform data augmentation with artificially generated events of the desired type and energy using conditional generative adversarial networks (cGANs), distinguishing classes by energy values. In the paper, we describe a simple algorithm for generating balanced data sets using cGANs. Thus, the proposed neural network model produces both imbalanced data sets for physical analysis as well as balanced data sets suitable for training other neural networks.

In this work, the ability of rare VHE gamma ray selection with neural network methods is investigated in the case when cosmic radiation flux strongly prevails (ratio up to 10^4 over the gamma radiation flux from a point source). This ratio is valid for the Crab Nebula in the TeV energy range, since the Crab is a well-studied source for calibration and test of various methods and installations in gamma astronomy. The part of TAIGA experiment which includes three Imaging Atmospheric Cherenkov Telescopes observes this gamma-source too. Cherenkov telescopes obtain images of Extensive Air Showers. Hillas parameters can be used to analyse images in standard processing method, or images can be processed with convolutional neural networks. In this work we would like to describe the main steps and results obtained in the gamma/hadron separation task from the Crab Nebula with neural network methods. The results obtained are compared with standard processing method applied in the TAIGA collaboration and using Hillas parameter cuts. It is demonstrated that a signal was received at the level of higher than 5.5σ in 21 h of Crab Nebula observations after processing the experimental data with the neural network method.

It has been previously demonstrated by Panov et al. (2021) that the TAIGA-HiSCORE Cherenkov array, originally built for cosmic ray physics and ultrahigh-energy gamma-ray astronomy studies using the extensive air shower method, can be used in conventional optical astronomy for wide-field searches for rare nanosecond optical transients of astrophysical origin. The field of view of the facility is on the scale of 1 ster, and it is capable of detecting very rare transients in the visible light range with fluxes greater than approximately 3000 quanta/m2/10 ns (10 ns is the apparatus integration time) and pulse durations of 10 ns. Among the potential sources of distant nanosecond optical transients are the evaporation of primary black holes, magnetic reconnection in the accretion disks of black holes, and signals from distant lasers of extraterrestrial civilizations. The paper describes the methods and results of the search for optical transients using the TAIGA-HiSCORE Cherenkov array from 2018 to 2022 (four winter seasons of data collection). No reliable astrophysical candidates for optical transients were found. We set an upper bound on the flux of the searched events as ∼ 1pt 1 ×10^ - 3 events/ster/h.

The TAIGA-HiSCORE setup is a wide-angle Cherenkov detector array for recording extensive air showers (EASs). The array comprises over 120 stations located in the Tunka Valley near Lake Baikal. One of the main tasks of data analysis in the TAIGA-HiSCORE experiment is to determine the axis direction of the EASs and their core location. These parameters are used to determine the source of gamma rays and play an important role in estimating the energy of the primary particle. The data collected by HiSCORE stations include signal amplitude and arrival time and allow for estimation of the shower direction of arrival. In this work, we use Monte Carlo simulation data for HiSCORE to demonstrate the feasibility of determining the EAS axis directions with artificial neural networks. Our approach employs multilayer perceptrons with skip connections, which take data from subsets of HiSCORE stations as input. Multiple station subsets are selected to derive more accurate composite estimates. Furthermore, we use a two-stage algorithm, where the initial direction estimates in the first stage are refined in the second stage. The final estimates have an average error of less than 0.25 ^∘ . We plan to use this approach as a part of multimodal analysis of data obtained from several types of detectors used in the TAIGA experiment.

The paper is devoted to the modeling and analysis of data detected by the TAIGA-IACT installation in the stereo mode. Five Imaging Atmospheric Cherenkov Telescopes (IACT) with a viewing angle of 9.6° are expected to be included in the installation. Today there are three telescopes spaced far apart (from 320 to 500 m) in the installation. The effective area of the installation is as large as 0.6 km2; therefore, it is possible to conduct statistically significant measurements of weak γ-ray sources in the energy range above 10 TeV over a reasonable observation time (300–400 h). The Monte Carlo procedure for simulating the hadrons and γ-rays detected by the telescopes is described as is the procedure for reconstructing the parameters of extensive air showers, such as the arrival direction of an event, the axis position, the depth of the maximum of shower development (Xmax), and the primary-particle energy. In order to solve the problem of γ-hadron separation, the criteria for selecting γ-rays detected in the stereo mode have been optimized and the effective area of the installation has been calculated.

The more correct recalculation from the measured Cherenkov light fluxes at distances of 200 (Q200) and 100 (Q100) m from the Extensive Air Shower (EAS) core to the energy of the primary particle has been developed using the results of M-C simulation by the CORSIKA code, assuming a light primary composition of cosmic rays. Using the new conversion expressions, a differential energy spectrum was obtained according to the data of the Tunka-133 array for 7 years of operation and the TAIGA-HiSCORE array for 2 years of operation.

The Tunka Advanced Instrument for Gamma-ray and cosmic ray Astrophysics (TAIGA) is a hybrid experiment for the measurement of Extensive Air Showers (EAS) with good spectral resolution in the TeV to PeV energy range. In this domain, the long-sought Pevatrons can be detected. Currently the hybrid TAIGA detector combines two wide angle shower front Cherenkov light sampling timing arrays (HiSCORE and Tunka-133), two 4m class, 10 degrees aperture Imaging Air Cherenkov Telescopes (IACTs) and 240 m(2) surface and underground charged particle detector stations. Our goal is to introduce a new hybrid reconstruction technique, combining the good angular and shower core resolution of HiSCORE with the gamma-hadron separation power of imaging telescopes. This approach allows to maximize the effective area and simultaneously to reach a good gamma-hadron separation at low energies (few TeV). At higher energies, muon detectors are planned to enhance gamma-hadron separation. During the commissioning phase of the first and second IACT, several sources were observed. First detections of known sources with the first telescope show the functionality of the TAIGA IACTs. Here, the status of the TAIGA experiment will be presented, along with first results from the current configuration.

An analysis is performed of the spectrum of gamma rays from the Crab Nebula in the 4–100 TeV range of energies, obtained using data from two Atmospheric Cherenkov Telescopes that are part of the TAIGA complex. A way of selecting and restoring the energy of gamma rays is described that includes a procedure for restoring the energy spectrum.

The Tunka-Grande scintillation array is described. Scientific results obtained over the first five years of its operation are presented. Prospects for studying cosmic rays in the 1016–1018 eV range of energies are discussed.

This paper presents the results of an analysis of observations of the Crab Nebula gamma-ray source with the first two atmospheric Cherenkov telescopes of the TAIGA (Tunka Advanced Instrument for cosmic ray physics and Gamma Astronomy) astrophysical complex in the stereo mode of observations. The article analyzed observational data from 2020 to 2021. Over 36 hours of observations, a signal was obtained at a statistical significance level of 5 σ and a spectrum of gamma rays was plotted in the energy range from 2 to 70 TeV. The paper describes a technique for gamma–hadron separation and reconstruction of detected gamma-rays energy.

High-energy cosmic-ray research via the detection of Cherenkov radiation from extensive air showers was begun in the Tunka valley (50 km to the west from the southern extremity of Lake Baikal) in the early 1990s. A series of large arrays combined into the TAIGA (Tunka Advanced Instrument for cosmic-ray physics and Gamma Astronomy) astrophysical facility and designed to study gamma rays and charged cosmic rays have been created in the elapsed time. Descriptions of the facility arrays and the main results obtained while investigating high-energy cosmic rays are presented. Plans for a further development of the astrophysical facility are discussed.

A study is made of the astroclimatic conditions for performing nighttime astrophysical observations on the territory of the Greater Altai region. Nighttime data from both the VIIRS radiometer of the Suomi NPP satellite platform and the AIRS hyperspectrometer of the Aqua satellite are used. Topographic and astroclimatic criteria show that the Chuya steppe region (Altai Republic, Russia) and the plateau of Lake Khubsugul (Aimak Khuvsgel, Mongolia) are best suited for the deployment a full-scale gamma astronomy experiment. Infrastructure considerations make the territory in the western part of the Chuya steppe preferable.

The 2nd TAIGA imaging air Cherenkov telescope (IACT) has successfully been put into operation in the Tunka Valley in fall 2020. Currently three more telescopes are under construction and completion. The ability to use the telescopes in the so-called stereo mode of image analysis by taking into account the unusually large distance between them (from 320 m to 500 m), well exceeding the inter-telescope distances in conventional IACT stereo systems, is being explored and discussed. The results of the dedicated Monte Carlo are compared with the experiment data from the 1st and the 2nd TAIGA-IACT.

In this paper we present the first attempt of adaptation the Random Forest (RF) machine learning algorithm to gamma/hadron separation in the TAIGA experiment (Tunka Advanced Instrument for cosmic ray physics and Gamma-ray Astronomy). The TAIGA experiment will include HiSCORE array with 120 wide-angle Cherenkov detectors on the area of 1 \(km^2\) and 5 Imaging Atmospheric Cherenkov Telescopes (IACT) on the same area. At the first stage of the analysis, only images obtained by one IACT were included in consideration. The training process occurs on samples of parameterized images obtained from Monte Carlo (MC) data for gammas and hadrons with a ‘Scaled Hillas Parameters’ standard technique. It was shown that the program effectively separates gamma-like showers, RF method does produce stable results and is robust with respect to input parameters and provides a simple control and setup of the procedure for extracting showers from gamma rays.

The Tunka Advanced Instrument for Gamma-ray and cosmic ray Astrophysics (TAIGA) is a hybrid experiment for the measurement of Extensive Air Showers (EAS) with good spectral resolution in the TeV to PeV energy range. In this domain, the long-sought Pevatrons can be detected. Currently the TAIGA detector complex combines a two wide angle shower front Cherenkov light sampling timing arrays (HiSCORE and Tunka-133), two 4 m class, 10 $${}^{\circ}$$ aperture Imaging Air Cherenkov Telescopes (IACTs) and 240 m $${}^{2}$$ surface and underground charged particle detector stations. Our goal is to introduce a new hybrid reconstruction technique, combining the good angular and shower core resolution of HiSCORE with the gamma-hadron separation power of imaging air Cherenkov telescopes. This approach allows to maximize the effective area and simultaneously to reach a good gamma-hadron separation at low energies (few teraelectronvolts). At higher energies, muon detectors are planned to enhance gamma-hadron separation. During the commissioning phase of the first and second IACT, several sources were observed. First detections of known sources with the first telescope show the functionality of the TAIGA IACTs. Here, the status of the TAIGA experiment will be presented, along with first results from the current configuration.

TAIGA (Tunka Advanced Instrument for cosmic ray physics and Gamma Astronomy) Astrophysical complex is being developed for studies of gamma rays and charged cosmic rays in the energy range of 10 $${}^{13}$$ –10 $${}^{18}$$ eV. The complex is located in the Tunka Valley, about 50 km from Lake Baikal. In this report we present the experiment status and plans for study of high-energy cosmic-ray physics as well as main results reached by wide-angle TAIGA-HiSCORE and Tunka-133 Cherenkov arrays of the Astrophysical complex. Plans to study cosmic rays by means of other arrays of the complex, namely scintillation array Tunka-Grande and new TAIGA-muons array and system of IACT telescopes, are discussed too.

A wide-angle Cerenkov array TAIGA-HiSCORE (FOV $$\sim$$ 0.6 sr), was originally created as a part of TAIGA installation for high-energy gamma-ray astronomy and cosmic ray physics. Array now consist on nearly 100 optical stations on the area of 1 km $${}^{2}$$ . Due to high accuracy and stability ( $$\sim$$ 1 ns) of time synchronization of the optical stations the accuracy of EAS arrival direction reconstruction is reached 0.1 $${}^{\circ}$$ . It was proven that the array can also be used to search for nanosecond events of the optical range. The report discusses the method of searching for optical transients using the HiSCORE array and demonstrates its performance on a real example of detecting signals from an artificial Earth satellite. The search for this short flares in the HiSCORE data of the winter season 2018–2019 is carried out. One candidate for double repeater has been detected, but the estimated probability of random simulation of such a transient by background EAS events is not less than 10 $$\%$$ , which does not allow us to say that the detected candidate corresponds to a real astrophysical transient. An upper bound on the frequency of optical spikes with flux density of more than $$10^{-4}\textrm{ erg/s/cm}^{2}$$ and a duration of more than 5 ns is established as $$\sim 2\times 10^{-3}$$  events/sr/h.